diff --git a/.claude/skills/write-sglang-test/SKILL.md b/.claude/skills/write-sglang-test/SKILL.md index 9271f3f11..a35ddd6e1 100644 --- a/.claude/skills/write-sglang-test/SKILL.md +++ b/.claude/skills/write-sglang-test/SKILL.md @@ -18,10 +18,10 @@ description: Guide for writing SGLang CI/UT tests following project conventions. | Scenario | Model | CI Registration | Suite | |----------|-------|-----------------|-------| -| **Unit tests** (no server / engine launch) | None | `register_cpu_ci` (prefer) or `register_cuda_ci` | `stage-a-cpu-only` or `stage-b-test-small-1-gpu` | -| **Common / backend-independent** (middleware, abort, routing, config, arg parsing) | `DEFAULT_SMALL_MODEL_NAME_FOR_TEST` (1B) | `register_cuda_ci` only | `stage-b-test-small-1-gpu` | -| **Model-agnostic functionality** (sampling, session, OpenAI API features) | `DEFAULT_SMALL_MODEL_NAME_FOR_TEST` (1B) | `register_cuda_ci` (+ AMD if relevant) | `stage-b-test-small-1-gpu` | -| **General performance** (single node, no spec/DP/parallelism) | `DEFAULT_MODEL_NAME_FOR_TEST` (8B) | `register_cuda_ci` | `stage-b-test-large-1-gpu` | +| **Unit tests** (no server / engine launch) | None | `register_cpu_ci` (prefer) or `register_cuda_ci` | `stage-a-test-cpu` or `stage-b-test-1-gpu-small` | +| **Common / backend-independent** (middleware, abort, routing, config, arg parsing) | `DEFAULT_SMALL_MODEL_NAME_FOR_TEST` (1B) | `register_cuda_ci` only | `stage-b-test-1-gpu-small` | +| **Model-agnostic functionality** (sampling, session, OpenAI API features) | `DEFAULT_SMALL_MODEL_NAME_FOR_TEST` (1B) | `register_cuda_ci` (+ AMD if relevant) | `stage-b-test-1-gpu-small` | +| **General performance** (single node, no spec/DP/parallelism) | `DEFAULT_MODEL_NAME_FOR_TEST` (8B) | `register_cuda_ci` | `stage-b-test-1-gpu-large` | | **Bigger features** (spec, DP, TP, disaggregation) | Case by case | Case by case | See suite table below | **Key principle for E2E tests**: Do NOT add `register_amd_ci` unless the test specifically exercises AMD/ROCm code paths. Common E2E tests just need any GPU to run — duplicating across backends wastes CI time with no extra coverage. @@ -43,10 +43,10 @@ Defined in `python/sglang/test/test_utils.py`: | Suite | Runner | Scenario | |-------|--------|----------| -| `stage-a-cpu-only` | CPU | CPU unit tests | -| `stage-b-test-small-1-gpu` | 1× 5090 (32GB) | Small model tests | -| `stage-b-test-large-1-gpu` | 1× H100 (80GB) | 8B model tests | -| `stage-b-test-large-2-gpu` | 2× H100 | TP=2 tests | +| `stage-a-test-cpu` | CPU | CPU unit tests | +| `stage-b-test-1-gpu-small` | 1× 5090 (32GB) | Small model tests | +| `stage-b-test-1-gpu-large` | 1× H100 (80GB) | 8B model tests | +| `stage-b-test-2-gpu-large` | 2× H100 | TP=2 tests | | `stage-c-test-4-gpu-h100` | 4× H100 | TP=4 / EP tests | | `stage-c-test-8-gpu-h200` | 8× H200 | Large-scale multi-GPU | | `nightly-1-gpu` | 1 GPU | Nightly-only | @@ -70,7 +70,7 @@ from sglang.srt. import TargetClass from sglang.test.ci.ci_register import register_cpu_ci from sglang.test.test_utils import CustomTestCase -register_cpu_ci(est_time=5, suite="stage-a-cpu-only") +register_cpu_ci(est_time=5, suite="stage-a-test-cpu") # Prefer CPU. Only use register_cuda_ci when the test truly needs a GPU. class TestTargetClass(CustomTestCase): @@ -110,7 +110,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=60, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=60, suite="stage-b-test-1-gpu-small") class TestMyFeature(CustomTestCase): @@ -159,7 +159,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=300, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=300, suite="stage-b-test-1-gpu-large") class TestMyFeaturePerf(CustomTestCase): @@ -227,7 +227,7 @@ Every test file in `test/registered/` **must** call a registration function at m ```python from sglang.test.ci.ci_register import register_cuda_ci -register_cuda_ci(est_time=60, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=60, suite="stage-b-test-1-gpu-small") ``` Parameters: diff --git a/.github/actions/wait-for-jobs/action.yml b/.github/actions/wait-for-jobs/action.yml index 36169fd20..c41bed474 100644 --- a/.github/actions/wait-for-jobs/action.yml +++ b/.github/actions/wait-for-jobs/action.yml @@ -8,7 +8,7 @@ inputs: jobs: description: | JSON array of job specs to wait for. Each element is either: - - a string: exact job name (e.g. "stage-a-test-small-1-gpu") + - a string: exact job name (e.g. "stage-a-test-1-gpu-small") - an object { "prefix": "...", "expected_count": N }: for matrix jobs required: true max-wait-minutes: diff --git a/.github/workflows/pr-test-amd-rocm720.yml b/.github/workflows/pr-test-amd-rocm720.yml index 5e21e505b..dcc6b4925 100644 --- a/.github/workflows/pr-test-amd-rocm720.yml +++ b/.github/workflows/pr-test-amd-rocm720.yml @@ -35,13 +35,13 @@ on: - '' - sgl-kernel-unit-test-amd - sgl-kernel-unit-test-2-gpu-amd - - stage-a-test-small-1-gpu-amd + - stage-a-test-1-gpu-small-amd - jit-kernel-unit-test-amd - - stage-b-test-small-1-gpu-amd - - stage-b-test-small-1-gpu-amd-nondeterministic - - stage-b-test-small-1-gpu-amd-mi35x - - stage-b-test-large-1-gpu-amd - - stage-b-test-large-2-gpu-amd + - stage-b-test-1-gpu-small-amd + - stage-b-test-1-gpu-small-amd-nondeterministic + - stage-b-test-1-gpu-small-amd-mi35x + - stage-b-test-1-gpu-large-amd + - stage-b-test-2-gpu-large-amd - multimodal-gen-test-1-gpu-amd - multimodal-gen-test-2-gpu-amd - stage-c-test-large-8-gpu-amd @@ -246,12 +246,12 @@ jobs: # =============================================== primary ==================================================== - stage-a-test-small-1-gpu-amd: + stage-a-test-1-gpu-small-amd: needs: [check-changes] if: | always() && ( - (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-a-test-small-1-gpu-amd,')) || + (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-a-test-1-gpu-small-amd,')) || ( !(inputs.target_stage || inputs.target_stage_select) && (!failure() && !cancelled()) && @@ -283,7 +283,7 @@ jobs: - name: Run test timeout-minutes: 10 run: | - bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-a-test-small-1-gpu-amd ${{ inputs.continue_on_error && '--continue-on-error' || '' }} + bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-a-test-1-gpu-small-amd ${{ inputs.continue_on_error && '--continue-on-error' || '' }} jit-kernel-unit-test-amd: needs: [check-changes] @@ -323,12 +323,12 @@ jobs: run: | bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout" python3 -m pytest -q python/sglang/jit_kernel/tests/test_store_cache.py - stage-b-test-small-1-gpu-amd: + stage-b-test-1-gpu-small-amd: needs: [check-changes] if: | always() && ( - (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-small-1-gpu-amd,')) || + (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-1-gpu-small-amd,')) || ( !(inputs.target_stage || inputs.target_stage_select) && (!failure() && !cancelled()) && @@ -360,14 +360,14 @@ jobs: - name: Run test timeout-minutes: 30 run: | - bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-small-1-gpu-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 14 --timeout-per-file 1800 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} + bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-1-gpu-small-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 14 --timeout-per-file 1800 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} - stage-b-test-small-1-gpu-amd-nondeterministic: + stage-b-test-1-gpu-small-amd-nondeterministic: needs: [check-changes] if: | always() && ( - (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-small-1-gpu-amd-nondeterministic,')) || + (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-1-gpu-small-amd-nondeterministic,')) || ( !(inputs.target_stage || inputs.target_stage_select) && (!failure() && !cancelled()) && @@ -398,14 +398,14 @@ jobs: - name: Run test timeout-minutes: 30 run: | - bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-small-1-gpu-amd-nondeterministic --timeout-per-file 1800 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} + bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-1-gpu-small-amd-nondeterministic --timeout-per-file 1800 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} - stage-b-test-small-1-gpu-amd-mi35x: + stage-b-test-1-gpu-small-amd-mi35x: needs: [check-changes] if: | always() && ( - (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-small-1-gpu-amd-mi35x,')) || + (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-1-gpu-small-amd-mi35x,')) || ( !(inputs.target_stage || inputs.target_stage_select) && (!failure() && !cancelled()) && @@ -436,14 +436,14 @@ jobs: - name: Run test timeout-minutes: 30 run: | - bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-small-1-gpu-amd-mi35x ${{ inputs.continue_on_error && '--continue-on-error' || '' }} + bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-1-gpu-small-amd-mi35x ${{ inputs.continue_on_error && '--continue-on-error' || '' }} - stage-b-test-large-1-gpu-amd: + stage-b-test-1-gpu-large-amd: needs: [check-changes] if: | always() && ( - (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-large-1-gpu-amd,')) || + (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-1-gpu-large-amd,')) || ( !(inputs.target_stage || inputs.target_stage_select) && (!failure() && !cancelled()) && @@ -475,14 +475,14 @@ jobs: - name: Run test timeout-minutes: 30 run: | - bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-large-1-gpu-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 2 --timeout-per-file 1800 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} + bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-1-gpu-large-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 2 --timeout-per-file 1800 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} - stage-b-test-large-2-gpu-amd: + stage-b-test-2-gpu-large-amd: needs: [check-changes] if: | always() && ( - (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-large-2-gpu-amd,')) || + (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-2-gpu-large-amd,')) || ( !(inputs.target_stage || inputs.target_stage_select) && (!failure() && !cancelled()) && @@ -514,7 +514,7 @@ jobs: - name: Run test timeout-minutes: 30 run: | - bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-large-2-gpu-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 2 --timeout-per-file 1800 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} + bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-2-gpu-large-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 2 --timeout-per-file 1800 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} multimodal-gen-test-1-gpu-amd: needs: [check-changes] @@ -984,13 +984,13 @@ jobs: multimodal-gen-test-1-gpu-amd, multimodal-gen-test-2-gpu-amd, - stage-a-test-small-1-gpu-amd, + stage-a-test-1-gpu-small-amd, jit-kernel-unit-test-amd, - stage-b-test-small-1-gpu-amd, - stage-b-test-small-1-gpu-amd-nondeterministic, - stage-b-test-small-1-gpu-amd-mi35x, - stage-b-test-large-1-gpu-amd, - stage-b-test-large-2-gpu-amd, + stage-b-test-1-gpu-small-amd, + stage-b-test-1-gpu-small-amd-nondeterministic, + stage-b-test-1-gpu-small-amd-mi35x, + stage-b-test-1-gpu-large-amd, + stage-b-test-2-gpu-large-amd, stage-b-test-large-8-gpu-35x-disaggregation-amd, stage-c-test-large-8-gpu-amd, stage-c-test-large-8-gpu-amd-mi35x, diff --git a/.github/workflows/pr-test-amd.yml b/.github/workflows/pr-test-amd.yml index d231527ae..561d1075f 100644 --- a/.github/workflows/pr-test-amd.yml +++ b/.github/workflows/pr-test-amd.yml @@ -33,13 +33,13 @@ on: - '' - sgl-kernel-unit-test-amd - sgl-kernel-unit-test-2-gpu-amd - - stage-a-test-small-1-gpu-amd + - stage-a-test-1-gpu-small-amd - jit-kernel-unit-test-amd - - stage-b-test-small-1-gpu-amd - - stage-b-test-small-1-gpu-amd-nondeterministic - - stage-b-test-small-1-gpu-amd-mi35x - - stage-b-test-large-1-gpu-amd - - stage-b-test-large-2-gpu-amd + - stage-b-test-1-gpu-small-amd + - stage-b-test-1-gpu-small-amd-nondeterministic + - stage-b-test-1-gpu-small-amd-mi35x + - stage-b-test-1-gpu-large-amd + - stage-b-test-2-gpu-large-amd - multimodal-gen-test-1-gpu-amd - multimodal-gen-test-2-gpu-amd - stage-c-test-large-8-gpu-amd @@ -246,12 +246,12 @@ jobs: # =============================================== primary ==================================================== - stage-a-test-small-1-gpu-amd: + stage-a-test-1-gpu-small-amd: needs: [check-changes] if: | always() && ( - (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-a-test-small-1-gpu-amd,')) || + (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-a-test-1-gpu-small-amd,')) || ( !(inputs.target_stage || inputs.target_stage_select) && (!failure() && !cancelled()) && @@ -284,7 +284,7 @@ jobs: - name: Run test timeout-minutes: 10 run: | - bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-a-test-small-1-gpu-amd ${{ inputs.continue_on_error && '--continue-on-error' || '' }} + bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-a-test-1-gpu-small-amd ${{ inputs.continue_on_error && '--continue-on-error' || '' }} jit-kernel-unit-test-amd: needs: [check-changes] @@ -325,12 +325,12 @@ jobs: run: | bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout" python3 -m pytest -q python/sglang/jit_kernel/tests/test_store_cache.py - stage-b-test-small-1-gpu-amd: - needs: [check-changes, stage-a-test-small-1-gpu-amd] + stage-b-test-1-gpu-small-amd: + needs: [check-changes, stage-a-test-1-gpu-small-amd] if: | always() && ( - (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-small-1-gpu-amd,')) || + (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-1-gpu-small-amd,')) || ( !(inputs.target_stage || inputs.target_stage_select) && (!failure() && !cancelled()) && @@ -363,14 +363,14 @@ jobs: - name: Run test timeout-minutes: 30 run: | - bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-small-1-gpu-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 14 --timeout-per-file 1800 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} + bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-1-gpu-small-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 14 --timeout-per-file 1800 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} - stage-b-test-small-1-gpu-amd-nondeterministic: - needs: [check-changes, stage-a-test-small-1-gpu-amd] + stage-b-test-1-gpu-small-amd-nondeterministic: + needs: [check-changes, stage-a-test-1-gpu-small-amd] if: | always() && ( - (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-small-1-gpu-amd-nondeterministic,')) || + (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-1-gpu-small-amd-nondeterministic,')) || ( !(inputs.target_stage || inputs.target_stage_select) && (!failure() && !cancelled()) && @@ -402,14 +402,14 @@ jobs: - name: Run test timeout-minutes: 30 run: | - bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-small-1-gpu-amd-nondeterministic --timeout-per-file 1800 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} + bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-1-gpu-small-amd-nondeterministic --timeout-per-file 1800 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} - stage-b-test-small-1-gpu-amd-mi35x: - needs: [check-changes, stage-a-test-small-1-gpu-amd] + stage-b-test-1-gpu-small-amd-mi35x: + needs: [check-changes, stage-a-test-1-gpu-small-amd] if: | always() && ( - (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-small-1-gpu-amd-mi35x,')) || + (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-1-gpu-small-amd-mi35x,')) || ( !(inputs.target_stage || inputs.target_stage_select) && (!failure() && !cancelled()) && @@ -441,14 +441,14 @@ jobs: - name: Run test timeout-minutes: 30 run: | - bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-small-1-gpu-amd-mi35x ${{ inputs.continue_on_error && '--continue-on-error' || '' }} + bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-1-gpu-small-amd-mi35x ${{ inputs.continue_on_error && '--continue-on-error' || '' }} - stage-b-test-large-1-gpu-amd: - needs: [check-changes, stage-a-test-small-1-gpu-amd] + stage-b-test-1-gpu-large-amd: + needs: [check-changes, stage-a-test-1-gpu-small-amd] if: | always() && ( - (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-large-1-gpu-amd,')) || + (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-1-gpu-large-amd,')) || ( !(inputs.target_stage || inputs.target_stage_select) && (!failure() && !cancelled()) && @@ -481,14 +481,14 @@ jobs: - name: Run test timeout-minutes: 30 run: | - bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-large-1-gpu-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 2 --timeout-per-file 1800 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} + bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-1-gpu-large-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 2 --timeout-per-file 1800 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} - stage-b-test-large-2-gpu-amd: - needs: [check-changes, stage-a-test-small-1-gpu-amd] + stage-b-test-2-gpu-large-amd: + needs: [check-changes, stage-a-test-1-gpu-small-amd] if: | always() && ( - (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-large-2-gpu-amd,')) || + (contains(format(',{0},', inputs.target_stage || inputs.target_stage_select), ',stage-b-test-2-gpu-large-amd,')) || ( !(inputs.target_stage || inputs.target_stage_select) && (!failure() && !cancelled()) && @@ -521,7 +521,7 @@ jobs: - name: Run test timeout-minutes: 30 run: | - bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-large-2-gpu-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 2 --timeout-per-file 1800 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} + bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-2-gpu-large-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 2 --timeout-per-file 1800 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} multimodal-gen-test-1-gpu-amd: needs: [check-changes] @@ -779,7 +779,7 @@ jobs: stage-c-test-large-8-gpu-amd: - needs: [check-changes, call-gate, stage-b-test-small-1-gpu-amd, stage-b-test-large-2-gpu-amd] + needs: [check-changes, call-gate, stage-b-test-1-gpu-small-amd, stage-b-test-2-gpu-large-amd] if: | always() && ( @@ -827,7 +827,7 @@ jobs: bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-c-test-large-8-gpu-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 3 --timeout-per-file 3600 ${{ inputs.continue_on_error && '--continue-on-error' || '' }} stage-c-test-large-8-gpu-amd-mi35x: - needs: [check-changes, call-gate, stage-b-test-small-1-gpu-amd, stage-b-test-large-2-gpu-amd] + needs: [check-changes, call-gate, stage-b-test-1-gpu-small-amd, stage-b-test-2-gpu-large-amd] if: | always() && ( @@ -868,7 +868,7 @@ jobs: # =============================================== Disaggregation ==================================================== stage-b-test-large-8-gpu-35x-disaggregation-amd: - needs: [check-changes, stage-a-test-small-1-gpu-amd] + needs: [check-changes, stage-a-test-1-gpu-small-amd] if: | always() && ( @@ -989,13 +989,13 @@ jobs: multimodal-gen-test-1-gpu-amd, multimodal-gen-test-2-gpu-amd, - stage-a-test-small-1-gpu-amd, + stage-a-test-1-gpu-small-amd, jit-kernel-unit-test-amd, - stage-b-test-small-1-gpu-amd, - stage-b-test-small-1-gpu-amd-nondeterministic, - stage-b-test-small-1-gpu-amd-mi35x, - stage-b-test-large-1-gpu-amd, - stage-b-test-large-2-gpu-amd, + stage-b-test-1-gpu-small-amd, + stage-b-test-1-gpu-small-amd-nondeterministic, + stage-b-test-1-gpu-small-amd-mi35x, + stage-b-test-1-gpu-large-amd, + stage-b-test-2-gpu-large-amd, stage-b-test-large-8-gpu-35x-disaggregation-amd, stage-c-test-large-8-gpu-amd, stage-c-test-large-8-gpu-amd-mi35x, diff --git a/.github/workflows/pr-test.yml b/.github/workflows/pr-test.yml index 4de2fc135..f037cb234 100644 --- a/.github/workflows/pr-test.yml +++ b/.github/workflows/pr-test.yml @@ -345,7 +345,7 @@ jobs: id: wait with: stage-name: stage-a - jobs: '["stage-a-test-small-1-gpu", "stage-a-cpu-only"]' + jobs: '["stage-a-test-1-gpu-small", "stage-a-test-cpu"]' max-wait-minutes: '240' wait-for-stage-b: @@ -377,9 +377,9 @@ jobs: stage-name: stage-b jobs: | [ - {"prefix": "stage-b-test-small-1-gpu", "expected_count": 8}, - {"prefix": "stage-b-test-large-1-gpu", "expected_count": 14}, - {"prefix": "stage-b-test-large-2-gpu", "expected_count": 4}, + {"prefix": "stage-b-test-1-gpu-small", "expected_count": 8}, + {"prefix": "stage-b-test-1-gpu-large", "expected_count": 14}, + {"prefix": "stage-b-test-2-gpu-large", "expected_count": 4}, {"prefix": "stage-b-test-4-gpu-b200", "expected_count": 1} ] max-wait-minutes: '480' @@ -833,12 +833,12 @@ jobs: # =============================================== primary ==================================================== # Runs on 5090 (32GB, SM120) - stage-a-test-small-1-gpu: + stage-a-test-1-gpu-small: needs: [check-changes, call-gate, sgl-kernel-build-wheels] if: | always() && ( - (inputs.target_stage == 'stage-a-test-small-1-gpu') || + (inputs.target_stage == 'stage-a-test-1-gpu-small') || ( !inputs.target_stage && ((github.event_name == 'schedule' || inputs.test_parallel_dispatch == true) || (!failure() && !cancelled())) && @@ -876,17 +876,17 @@ jobs: CONTINUE_ON_ERROR_FLAG: ${{ needs.check-changes.outputs.continue_on_error == 'true' && '--continue-on-error' || '' }} run: | cd test/ - python3 run_suite.py --hw cuda --suite stage-a-test-small-1-gpu $CONTINUE_ON_ERROR_FLAG + python3 run_suite.py --hw cuda --suite stage-a-test-1-gpu-small $CONTINUE_ON_ERROR_FLAG - uses: ./.github/actions/upload-cuda-coredumps if: always() - stage-a-cpu-only: + stage-a-test-cpu: needs: [check-changes, call-gate] if: | always() && ( - (inputs.target_stage == 'stage-a-cpu-only') || + (inputs.target_stage == 'stage-a-test-cpu') || ( !inputs.target_stage && ((github.event_name == 'schedule' || inputs.test_parallel_dispatch == true) || (!failure() && !cancelled())) && @@ -932,15 +932,15 @@ jobs: CONTINUE_ON_ERROR_FLAG: ${{ needs.check-changes.outputs.continue_on_error == 'true' && '--continue-on-error' || '' }} run: | cd test/ - python3 run_suite.py --hw cpu --suite stage-a-cpu-only $CONTINUE_ON_ERROR_FLAG + python3 run_suite.py --hw cpu --suite stage-a-test-cpu $CONTINUE_ON_ERROR_FLAG # Runs on 5090 (32GB, SM120) - stage-b-test-small-1-gpu: + stage-b-test-1-gpu-small: needs: [check-changes, call-gate, wait-for-stage-a, sgl-kernel-build-wheels] if: | always() && ( - (inputs.target_stage == 'stage-b-test-small-1-gpu') || + (inputs.target_stage == 'stage-b-test-1-gpu-small') || ( !inputs.target_stage && ((github.event_name == 'schedule' || inputs.test_parallel_dispatch == true) || (!failure() && !cancelled())) && @@ -988,7 +988,7 @@ jobs: run: | source /etc/profile.d/sglang-ci.sh cd test/ - python3 run_suite.py --hw cuda --suite stage-b-test-small-1-gpu --auto-partition-id ${{ matrix.partition }} --auto-partition-size 8 $CONTINUE_ON_ERROR_FLAG + python3 run_suite.py --hw cuda --suite stage-b-test-1-gpu-small --auto-partition-id ${{ matrix.partition }} --auto-partition-size 8 $CONTINUE_ON_ERROR_FLAG - uses: ./.github/actions/upload-cuda-coredumps if: always() @@ -996,12 +996,12 @@ jobs: artifact-suffix: ${{ matrix.partition }} # Runs on H100 (80GB, SM90) - tests that don't pass on 5090 (FA3, FP8, high VRAM, etc.) - stage-b-test-large-1-gpu: + stage-b-test-1-gpu-large: needs: [check-changes, call-gate, wait-for-stage-a, sgl-kernel-build-wheels] if: | always() && ( - (inputs.target_stage == 'stage-b-test-large-1-gpu') || + (inputs.target_stage == 'stage-b-test-1-gpu-large') || ( !inputs.target_stage && ((github.event_name == 'schedule' || inputs.test_parallel_dispatch == true) || (!failure() && !cancelled())) && @@ -1044,19 +1044,19 @@ jobs: CONTINUE_ON_ERROR_FLAG: ${{ needs.check-changes.outputs.continue_on_error == 'true' && '--continue-on-error' || '' }} run: | cd test/ - python3 run_suite.py --hw cuda --suite stage-b-test-large-1-gpu --auto-partition-id ${{ matrix.partition }} --auto-partition-size 14 --timeout-per-file 1800 $CONTINUE_ON_ERROR_FLAG + python3 run_suite.py --hw cuda --suite stage-b-test-1-gpu-large --auto-partition-id ${{ matrix.partition }} --auto-partition-size 14 --timeout-per-file 1800 $CONTINUE_ON_ERROR_FLAG - uses: ./.github/actions/upload-cuda-coredumps if: always() with: artifact-suffix: ${{ matrix.partition }} - stage-b-test-large-2-gpu: + stage-b-test-2-gpu-large: needs: [check-changes, call-gate, wait-for-stage-a, sgl-kernel-build-wheels] if: | always() && ( - (inputs.target_stage == 'stage-b-test-large-2-gpu') || + (inputs.target_stage == 'stage-b-test-2-gpu-large') || ( !inputs.target_stage && ((github.event_name == 'schedule' || inputs.test_parallel_dispatch == true) || (!failure() && !cancelled())) && @@ -1101,7 +1101,7 @@ jobs: CONTINUE_ON_ERROR_FLAG: ${{ needs.check-changes.outputs.continue_on_error == 'true' && '--continue-on-error' || '' }} run: | cd test/ - python3 run_suite.py --hw cuda --suite stage-b-test-large-2-gpu --auto-partition-id ${{ matrix.partition }} --auto-partition-size 4 $CONTINUE_ON_ERROR_FLAG + python3 run_suite.py --hw cuda --suite stage-b-test-2-gpu-large --auto-partition-id ${{ matrix.partition }} --auto-partition-size 4 $CONTINUE_ON_ERROR_FLAG - uses: ./.github/actions/upload-cuda-coredumps if: always() @@ -1502,12 +1502,12 @@ jobs: with: artifact-suffix: ${{ matrix.part }} - stage-c-test-deepep-4-gpu: + stage-c-test-deepep-4-gpu-h100: needs: [check-changes, call-gate, wait-for-stage-b] if: | always() && ( - (inputs.target_stage == 'stage-c-test-deepep-4-gpu') || + (inputs.target_stage == 'stage-c-test-deepep-4-gpu-h100') || ( !inputs.target_stage && ((github.event_name == 'schedule' || inputs.test_parallel_dispatch == true) || (!failure() && !cancelled())) && @@ -1557,7 +1557,7 @@ jobs: CONTINUE_ON_ERROR_FLAG: ${{ needs.check-changes.outputs.continue_on_error == 'true' && '--continue-on-error' || '' }} run: | cd test - python3 run_suite.py --hw cuda --suite stage-c-test-deepep-4-gpu $CONTINUE_ON_ERROR_FLAG + python3 run_suite.py --hw cuda --suite stage-c-test-deepep-4-gpu-h100 $CONTINUE_ON_ERROR_FLAG - uses: ./.github/actions/upload-cuda-coredumps if: always() @@ -1753,16 +1753,16 @@ jobs: multimodal-gen-test-1-gpu, multimodal-gen-test-2-gpu, - stage-a-test-small-1-gpu, - stage-a-cpu-only, - stage-b-test-small-1-gpu, - stage-b-test-large-1-gpu, - stage-b-test-large-2-gpu, + stage-a-test-1-gpu-small, + stage-a-test-cpu, + stage-b-test-1-gpu-small, + stage-b-test-1-gpu-large, + stage-b-test-2-gpu-large, stage-b-test-4-gpu-b200, stage-c-test-4-gpu-h100, stage-c-test-8-gpu-h20, stage-c-test-8-gpu-h200, - stage-c-test-deepep-4-gpu, + stage-c-test-deepep-4-gpu-h100, stage-c-test-deepep-8-gpu-h200, stage-c-test-4-gpu-b200, # stage-c-test-4-gpu-gb200, # Temporarily disabled — no GB200 runner diff --git a/python/sglang/jit_kernel/tests/test_moe_lora_align_block_size.py b/python/sglang/jit_kernel/tests/test_moe_lora_align_block_size.py index 89efb6d08..8a19baefa 100644 --- a/python/sglang/jit_kernel/tests/test_moe_lora_align_block_size.py +++ b/python/sglang/jit_kernel/tests/test_moe_lora_align_block_size.py @@ -10,7 +10,7 @@ import torch from sglang.jit_kernel.moe_lora_align import moe_lora_align_block_size from sglang.test.ci.ci_register import register_cuda_ci -register_cuda_ci(est_time=80, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=80, suite="stage-b-test-1-gpu-large") def round_up(x, base): diff --git a/scripts/ci/utils/slash_command_handler.py b/scripts/ci/utils/slash_command_handler.py index ae44f0905..0b2e4535b 100644 --- a/scripts/ci/utils/slash_command_handler.py +++ b/scripts/ci/utils/slash_command_handler.py @@ -253,18 +253,18 @@ def handle_rerun_stage( # Valid NVIDIA stage names that support target_stage nvidia_stages = [ - "stage-a-test-small-1-gpu", - "stage-a-cpu-only", - "stage-b-test-small-1-gpu", - "stage-b-test-large-1-gpu", - "stage-b-test-large-2-gpu", + "stage-a-test-1-gpu-small", + "stage-a-test-cpu", + "stage-b-test-1-gpu-small", + "stage-b-test-1-gpu-large", + "stage-b-test-2-gpu-large", "stage-b-test-4-gpu-b200", "stage-c-test-4-gpu-h100", "stage-c-test-8-gpu-h200", "stage-c-test-8-gpu-h20", "stage-c-test-4-gpu-b200", "stage-c-test-4-gpu-gb200", - "stage-c-test-deepep-4-gpu", + "stage-c-test-deepep-4-gpu-h100", "stage-c-test-deepep-8-gpu-h200", "multimodal-gen-test-1-gpu", "multimodal-gen-test-2-gpu", @@ -274,12 +274,12 @@ def handle_rerun_stage( amd_stages = [ "sgl-kernel-unit-test-amd", "sgl-kernel-unit-test-2-gpu-amd", - "stage-a-test-small-1-gpu-amd", - "stage-b-test-small-1-gpu-amd", - "stage-b-test-small-1-gpu-amd-nondeterministic", - "stage-b-test-small-1-gpu-amd-mi35x", - "stage-b-test-large-1-gpu-amd", - "stage-b-test-large-2-gpu-amd", + "stage-a-test-1-gpu-small-amd", + "stage-b-test-1-gpu-small-amd", + "stage-b-test-1-gpu-small-amd-nondeterministic", + "stage-b-test-1-gpu-small-amd-mi35x", + "stage-b-test-1-gpu-large-amd", + "stage-b-test-2-gpu-large-amd", "multimodal-gen-test-1-gpu-amd", "multimodal-gen-test-2-gpu-amd", "stage-c-test-large-8-gpu-amd", @@ -412,23 +412,23 @@ def handle_rerun_stage( CUDA_SUITE_TO_RUNNER = { - "stage-a-test-small-1-gpu": "1-gpu-5090", - "stage-a-cpu-only": "ubuntu-latest", - "stage-b-test-small-1-gpu": "1-gpu-5090", - "stage-b-test-large-1-gpu": "1-gpu-h100", - "stage-b-test-large-2-gpu": "2-gpu-h100", + "stage-a-test-1-gpu-small": "1-gpu-5090", + "stage-a-test-cpu": "ubuntu-latest", + "stage-b-test-1-gpu-small": "1-gpu-5090", + "stage-b-test-1-gpu-large": "1-gpu-h100", + "stage-b-test-2-gpu-large": "2-gpu-h100", "stage-b-test-4-gpu-b200": "4-gpu-b200", "stage-c-test-4-gpu-h100": "4-gpu-h100", "stage-c-test-8-gpu-h200": "8-gpu-h200", "stage-c-test-8-gpu-h20": "8-gpu-h20", "stage-c-test-4-gpu-b200": "4-gpu-b200", - "stage-c-test-deepep-4-gpu": "4-gpu-h100", + "stage-c-test-deepep-4-gpu-h100": "4-gpu-h100", "stage-c-test-deepep-8-gpu-h200": "8-gpu-h200", } DEEPEP_SUITES = { "stage-c-test-8-gpu-h20", - "stage-c-test-deepep-4-gpu", + "stage-c-test-deepep-4-gpu-h100", "stage-c-test-deepep-8-gpu-h200", } diff --git a/test/README.md b/test/README.md index 09805d3c2..090aaa048 100644 --- a/test/README.md +++ b/test/README.md @@ -47,23 +47,23 @@ from sglang.test.ci.ci_register import ( ) # Per-commit test (small 1-gpu, runs on 5090) -register_cuda_ci(est_time=80, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=80, suite="stage-b-test-1-gpu-small") # Per-commit test (large 1-gpu, runs on H100) -register_cuda_ci(est_time=120, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=120, suite="stage-b-test-1-gpu-large") # Per-commit test (2-gpu) -register_cuda_ci(est_time=200, suite="stage-b-test-large-2-gpu") +register_cuda_ci(est_time=200, suite="stage-b-test-2-gpu-large") # Nightly-only test register_cuda_ci(est_time=200, suite="nightly-1-gpu", nightly=True) # Multi-backend test -register_cuda_ci(est_time=80, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=120, suite="stage-a-test-small-1-gpu-amd") +register_cuda_ci(est_time=80, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=120, suite="stage-a-test-1-gpu-small-amd") # Temporarily disabled test -register_cuda_ci(est_time=80, suite="stage-b-test-small-1-gpu", disabled="flaky - see #12345") +register_cuda_ci(est_time=80, suite="stage-b-test-1-gpu-small", disabled="flaky - see #12345") ``` ### Choosing Between 1-GPU Suites (5090 vs H100) @@ -72,14 +72,14 @@ When adding 1-GPU tests, choose the appropriate suite based on hardware compatib | Suite | Runner | GPU | When to Use | |-------|--------|-----|-------------| -| `stage-a-test-small-1-gpu` | `1-gpu-5090` | RTX 5090 (32GB, SM120) | Stage A per-commit smoke on 5090 (CUDA) | -| `stage-a-test-small-1-gpu-amd` | AMD CI runners | ROCm | Stage A per-commit smoke (AMD) | -| `stage-b-test-small-1-gpu` | `1-gpu-5090` | RTX 5090 (32GB, SM120) | 5090-compatible tests (preferred) | -| `stage-b-test-large-1-gpu` | `1-gpu-h100` | H100 (80GB, SM90) | Large models or 5090-incompatible tests | +| `stage-a-test-1-gpu-small` | `1-gpu-5090` | RTX 5090 (32GB, SM120) | Stage A per-commit smoke on 5090 (CUDA) | +| `stage-a-test-1-gpu-small-amd` | AMD CI runners | ROCm | Stage A per-commit smoke (AMD) | +| `stage-b-test-1-gpu-small` | `1-gpu-5090` | RTX 5090 (32GB, SM120) | 5090-compatible tests (preferred) | +| `stage-b-test-1-gpu-large` | `1-gpu-h100` | H100 (80GB, SM90) | Large models or 5090-incompatible tests | -**Use `stage-b-test-small-1-gpu` (5090) whenever possible** - this is the preferred suite for most 1-GPU tests. +**Use `stage-b-test-1-gpu-small` (5090) whenever possible** - this is the preferred suite for most 1-GPU tests. -**Use `stage-b-test-large-1-gpu` (H100) if ANY of these apply:** +**Use `stage-b-test-1-gpu-large` (H100) if ANY of these apply:** 1. **Architecture incompatibility (SM120/Blackwell)**: - FA3 attention backend (requires SM≤90) @@ -95,18 +95,18 @@ When adding 1-GPU tests, choose the appropriate suite based on hardware compatib - Weight update/sync tests - Certain spec decoding tests -If a test cannot run on 5090 due to any of the above, use `stage-b-test-large-1-gpu` which runs on H100. +If a test cannot run on 5090 due to any of the above, use `stage-b-test-1-gpu-large` which runs on H100. ### Available Suites **Per-Commit (CUDA)**: -- Stage A: `stage-a-test-small-1-gpu` (5090), `stage-a-test-2`, `stage-a-test-cpu` -- Stage B: `stage-b-test-small-1-gpu` (5090), `stage-b-test-large-1-gpu` (H100), `stage-b-test-large-2-gpu` -- Stage C (4-GPU): `stage-c-test-4-gpu-h100`, `stage-c-test-4-gpu-b200`, `stage-c-test-4-gpu-gb200`, `stage-c-test-deepep-4-gpu` +- Stage A: `stage-a-test-1-gpu-small` (5090), `stage-a-test-2`, `stage-a-test-cpu` +- Stage B: `stage-b-test-1-gpu-small` (5090), `stage-b-test-1-gpu-large` (H100), `stage-b-test-2-gpu-large` +- Stage C (4-GPU): `stage-c-test-4-gpu-h100`, `stage-c-test-4-gpu-b200`, `stage-c-test-4-gpu-gb200`, `stage-c-test-deepep-4-gpu-h100` - Stage C (8-GPU): `stage-c-test-8-gpu-h20`, `stage-c-test-8-gpu-h200`, `stage-c-test-8-gpu-b200`, `stage-c-test-deepep-8-gpu-h200` **Per-Commit (AMD)**: -- `stage-a-test-small-1-gpu-amd`, `stage-b-test-small-1-gpu-amd`, `stage-b-test-large-2-gpu-amd` +- `stage-a-test-1-gpu-small-amd`, `stage-b-test-1-gpu-small-amd`, `stage-b-test-2-gpu-large-amd` **Nightly**: - `nightly-1-gpu`, `nightly-2-gpu`, `nightly-4-gpu`, `nightly-8-gpu`, etc. @@ -115,13 +115,13 @@ If a test cannot run on 5090 due to any of the above, use `stage-b-test-large-1- ```bash # Run per-commit tests -python test/run_suite.py --hw cuda --suite stage-b-test-small-1-gpu +python test/run_suite.py --hw cuda --suite stage-b-test-1-gpu-small # Run nightly tests python test/run_suite.py --hw cuda --suite nightly-1-gpu --nightly # With auto-partitioning (for parallel CI jobs) -python test/run_suite.py --hw cuda --suite stage-b-test-small-1-gpu \ +python test/run_suite.py --hw cuda --suite stage-b-test-1-gpu-small \ --auto-partition-id 0 --auto-partition-size 4 ``` diff --git a/test/registered/amd/disaggregation/test_disaggregation_basic.py b/test/registered/amd/disaggregation/test_disaggregation_basic.py index 86b3051e2..c88e2aca4 100644 --- a/test/registered/amd/disaggregation/test_disaggregation_basic.py +++ b/test/registered/amd/disaggregation/test_disaggregation_basic.py @@ -212,7 +212,7 @@ class TestDisaggregationAccuracy(PDDisaggregationServerBase): ) -# register_amd_ci(est_time=300, suite="stage-b-test-large-2-gpu-amd") +# register_amd_ci(est_time=300, suite="stage-b-test-2-gpu-large-amd") class TestDisaggregationMooncakeFailure(PDDisaggregationServerBase): @classmethod def setUpClass(cls): @@ -322,7 +322,7 @@ class TestDisaggregationMooncakeFailure(PDDisaggregationServerBase): raise e from health_check_error -# register_amd_ci(est_time=300, suite="stage-b-test-large-2-gpu-amd") +# register_amd_ci(est_time=300, suite="stage-b-test-2-gpu-large-amd") class TestDisaggregationSimulatedRetract(PDDisaggregationServerBase): @classmethod def setUpClass(cls): diff --git a/test/registered/attention/test_chunk_gated_delta_rule.py b/test/registered/attention/test_chunk_gated_delta_rule.py index 8fb71e2c8..e5d53c32a 100644 --- a/test/registered/attention/test_chunk_gated_delta_rule.py +++ b/test/registered/attention/test_chunk_gated_delta_rule.py @@ -8,7 +8,7 @@ from sglang.srt.layers.attention.fla.fused_recurrent import ( ) from sglang.test.ci.ci_register import register_cuda_ci -register_cuda_ci(est_time=60, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=60, suite="stage-b-test-1-gpu-large") @unittest.skipIf(not torch.cuda.is_available(), "Test requires CUDA") diff --git a/test/registered/attention/test_create_kvindices.py b/test/registered/attention/test_create_kvindices.py index 3117abe96..0b06e8222 100644 --- a/test/registered/attention/test_create_kvindices.py +++ b/test/registered/attention/test_create_kvindices.py @@ -9,8 +9,8 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.test_utils import CustomTestCase # Triton kernel unit test for KV indices creation -register_cuda_ci(est_time=10, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=10, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=10, suite="stage-b-test-1-gpu-small-amd") class TestCreateKvIndices(CustomTestCase): diff --git a/test/registered/attention/test_fa3.py b/test/registered/attention/test_fa3.py index 3ecf91ad4..56969b9c8 100644 --- a/test/registered/attention/test_fa3.py +++ b/test/registered/attention/test_fa3.py @@ -20,7 +20,7 @@ from sglang.test.test_utils import ( # FlashAttention3 integration tests (requires SM 90+ / H100) # Multiple test classes: FA3, FA3+MLA, FA3+SpecDecode variants -register_cuda_ci(est_time=300, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=300, suite="stage-b-test-1-gpu-large") GSM_DATASET_PATH = None diff --git a/test/registered/attention/test_hybrid_attn_backend.py b/test/registered/attention/test_hybrid_attn_backend.py index 27f6ade20..1c70ee032 100644 --- a/test/registered/attention/test_hybrid_attn_backend.py +++ b/test/registered/attention/test_hybrid_attn_backend.py @@ -20,7 +20,7 @@ from sglang.test.test_utils import ( # Hybrid attention backend tests (FA3 prefill + FlashInfer decode, requires SM 90+ / H100) # Multiple test classes: base, MLA, TorchCompile, SpecDecode variants -register_cuda_ci(est_time=200, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=200, suite="stage-b-test-1-gpu-large") GSM_DATASET_PATH = None diff --git a/test/registered/attention/test_kda_kernels.py b/test/registered/attention/test_kda_kernels.py index 57d3962eb..d7a971934 100644 --- a/test/registered/attention/test_kda_kernels.py +++ b/test/registered/attention/test_kda_kernels.py @@ -8,7 +8,7 @@ from sglang.srt.layers.attention.fla.fused_sigmoid_gating_recurrent import ( from sglang.srt.layers.attention.fla.kda import fused_kda_gate, fused_recurrent_kda from sglang.test.ci.ci_register import register_cuda_ci -register_cuda_ci(est_time=30, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=30, suite="stage-b-test-1-gpu-large") @unittest.skipIf(not torch.cuda.is_available(), "Test requires CUDA") diff --git a/test/registered/attention/test_normal_decode_set_metadata.py b/test/registered/attention/test_normal_decode_set_metadata.py index d4101c67e..21e701521 100644 --- a/test/registered/attention/test_normal_decode_set_metadata.py +++ b/test/registered/attention/test_normal_decode_set_metadata.py @@ -21,7 +21,7 @@ from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.test_utils import CustomTestCase # Register this test for CUDA CI in stage-b (fast attention/kernel tests) -register_cuda_ci(est_time=25, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=25, suite="stage-b-test-1-gpu-large") def reference_normal_decode_set_metadata( diff --git a/test/registered/attention/test_torch_native_attention_backend.py b/test/registered/attention/test_torch_native_attention_backend.py index 2e6bc256d..0d2c9f330 100644 --- a/test/registered/attention/test_torch_native_attention_backend.py +++ b/test/registered/attention/test_torch_native_attention_backend.py @@ -18,8 +18,8 @@ from sglang.test.test_utils import ( ) # Torch native attention backend integration test with MMLU eval -register_cuda_ci(est_time=169, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=150, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=169, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=150, suite="stage-b-test-1-gpu-small-amd") class TestTorchNativeAttnBackend(CustomTestCase): diff --git a/test/registered/attention/test_triton_attention_backend.py b/test/registered/attention/test_triton_attention_backend.py index 52757d383..47900a20a 100644 --- a/test/registered/attention/test_triton_attention_backend.py +++ b/test/registered/attention/test_triton_attention_backend.py @@ -20,8 +20,8 @@ from sglang.test.test_utils import ( ) # Triton attention backend integration test with latency benchmark and MMLU eval -register_cuda_ci(est_time=200, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=1400, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=200, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=1400, suite="stage-b-test-1-gpu-small-amd") class TestTritonAttnBackend(CustomTestCase): diff --git a/test/registered/attention/test_triton_attention_kernels.py b/test/registered/attention/test_triton_attention_kernels.py index d9520172b..89bc81ee0 100644 --- a/test/registered/attention/test_triton_attention_kernels.py +++ b/test/registered/attention/test_triton_attention_kernels.py @@ -23,8 +23,8 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.test_utils import CustomTestCase, is_in_amd_ci # Triton attention kernel unit tests (decode, extend, prefill) -register_cuda_ci(est_time=30, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=30, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=30, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=30, suite="stage-b-test-1-gpu-small-amd") def extend_attention_fwd_torch( diff --git a/test/registered/attention/test_triton_sliding_window.py b/test/registered/attention/test_triton_sliding_window.py index 6eeb1e04a..dbc4ba3d5 100644 --- a/test/registered/attention/test_triton_sliding_window.py +++ b/test/registered/attention/test_triton_sliding_window.py @@ -16,8 +16,8 @@ from sglang.test.test_utils import ( ) # Sliding window attention with Triton backend (Gemma-3 model) -register_cuda_ci(est_time=100, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=200, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=100, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=200, suite="stage-b-test-1-gpu-small-amd") class TestSlidingWindowAttentionTriton(CustomTestCase): diff --git a/test/registered/attention/test_wave_attention_kernels.py b/test/registered/attention/test_wave_attention_kernels.py index 43051bd5a..9ffaaaf12 100644 --- a/test/registered/attention/test_wave_attention_kernels.py +++ b/test/registered/attention/test_wave_attention_kernels.py @@ -25,7 +25,7 @@ from sglang.srt.utils import get_device from sglang.test.ci.ci_register import register_amd_ci # Wave attention kernel unit tests (AMD only - requires wave_lang) -register_amd_ci(est_time=60, suite="stage-a-test-small-1-gpu-amd") +register_amd_ci(est_time=60, suite="stage-a-test-1-gpu-small-amd") class TestWaveAttention(unittest.TestCase): diff --git a/test/registered/backends/test_torch_compile.py b/test/registered/backends/test_torch_compile.py index 3e4454313..a426b79ce 100644 --- a/test/registered/backends/test_torch_compile.py +++ b/test/registered/backends/test_torch_compile.py @@ -16,8 +16,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=144, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=1100, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=144, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=1100, suite="stage-b-test-1-gpu-small-amd") class TestTorchCompile(CustomTestCase): diff --git a/test/registered/bench_fn/test_benchmark_datasets_api.py b/test/registered/bench_fn/test_benchmark_datasets_api.py index 96f781a22..82c8573d7 100644 --- a/test/registered/bench_fn/test_benchmark_datasets_api.py +++ b/test/registered/bench_fn/test_benchmark_datasets_api.py @@ -26,7 +26,7 @@ from sglang.benchmark.datasets.random import sample_random_requests from sglang.benchmark.datasets.sharegpt import sample_sharegpt_requests from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=5, suite="stage-a-cpu-only") +register_cpu_ci(est_time=5, suite="stage-a-test-cpu") class _DummyTokenTensor: diff --git a/test/registered/constrained_decoding/test_constrained_decoding.py b/test/registered/constrained_decoding/test_constrained_decoding.py index 66280138d..b705b6f99 100644 --- a/test/registered/constrained_decoding/test_constrained_decoding.py +++ b/test/registered/constrained_decoding/test_constrained_decoding.py @@ -13,8 +13,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=111, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=179, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=111, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=179, suite="stage-b-test-1-gpu-small-amd") class ServerWithGrammar(CustomTestCase): diff --git a/test/registered/core/test_deterministic.py b/test/registered/core/test_deterministic.py index aafbf101c..6d205a369 100644 --- a/test/registered/core/test_deterministic.py +++ b/test/registered/core/test_deterministic.py @@ -16,8 +16,8 @@ from sglang.test.test_deterministic_utils import ( ) from sglang.test.test_utils import is_in_amd_ci -register_cuda_ci(est_time=278, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=278, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=278, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=278, suite="stage-b-test-1-gpu-small-amd") @unittest.skipIf(is_in_amd_ci(), "Skip for AMD CI.") diff --git a/test/registered/core/test_gpt_oss_1gpu.py b/test/registered/core/test_gpt_oss_1gpu.py index 05b708541..179c5e8f3 100644 --- a/test/registered/core/test_gpt_oss_1gpu.py +++ b/test/registered/core/test_gpt_oss_1gpu.py @@ -3,8 +3,8 @@ import unittest from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.gpt_oss_common import BaseTestGptOss -register_cuda_ci(est_time=519, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=750, suite="stage-b-test-small-1-gpu-amd-mi35x") +register_cuda_ci(est_time=519, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=750, suite="stage-b-test-1-gpu-small-amd-mi35x") class TestGptOss1Gpu(BaseTestGptOss): diff --git a/test/registered/core/test_gpt_oss_sm120.py b/test/registered/core/test_gpt_oss_sm120.py index ee2f05716..8fc74a9d7 100644 --- a/test/registered/core/test_gpt_oss_sm120.py +++ b/test/registered/core/test_gpt_oss_sm120.py @@ -5,7 +5,7 @@ import torch from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.gpt_oss_common import BaseTestGptOss -register_cuda_ci(est_time=500, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=500, suite="stage-b-test-1-gpu-small") @unittest.skipIf(not torch.cuda.is_available(), "CUDA is not available") diff --git a/test/registered/core/test_hidden_states.py b/test/registered/core/test_hidden_states.py index 4bbdf828a..20c4db95e 100644 --- a/test/registered/core/test_hidden_states.py +++ b/test/registered/core/test_hidden_states.py @@ -8,8 +8,8 @@ from sglang.srt.utils import get_device, is_hip from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.test_utils import DEFAULT_SMALL_MODEL_NAME_FOR_TEST, CustomTestCase -register_cuda_ci(est_time=55, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=55, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=55, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=55, suite="stage-b-test-1-gpu-small-amd") _is_hip = is_hip() if _is_hip: diff --git a/test/registered/core/test_page_size.py b/test/registered/core/test_page_size.py index b28bd3c7d..b3e1b333e 100644 --- a/test/registered/core/test_page_size.py +++ b/test/registered/core/test_page_size.py @@ -13,8 +13,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=60, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=60, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=60, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=60, suite="stage-b-test-1-gpu-small-amd") class TestPageSize(CustomTestCase): diff --git a/test/registered/core/test_request_queue_validation.py b/test/registered/core/test_request_queue_validation.py index c76e006ca..36727c4de 100644 --- a/test/registered/core/test_request_queue_validation.py +++ b/test/registered/core/test_request_queue_validation.py @@ -17,8 +17,8 @@ from sglang.test.test_utils import ( send_generate_requests, ) -register_cuda_ci(est_time=47, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=70, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=47, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=70, suite="stage-b-test-1-gpu-small-amd") class TestMaxQueuedRequests(CustomTestCase): diff --git a/test/registered/core/test_score_api.py b/test/registered/core/test_score_api.py index 5d15d564f..9f7199c39 100644 --- a/test/registered/core/test_score_api.py +++ b/test/registered/core/test_score_api.py @@ -7,7 +7,7 @@ from sglang.srt.entrypoints.engine import Engine from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.test_utils import DEFAULT_SMALL_MODEL_NAME_FOR_TEST, CustomTestCase -register_cuda_ci(est_time=260, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=260, suite="stage-b-test-1-gpu-large") TEST_MODEL_NAME = DEFAULT_SMALL_MODEL_NAME_FOR_TEST diff --git a/test/registered/core/test_srt_endpoint.py b/test/registered/core/test_srt_endpoint.py index 92aeff84c..46c5853a6 100644 --- a/test/registered/core/test_srt_endpoint.py +++ b/test/registered/core/test_srt_endpoint.py @@ -27,8 +27,8 @@ from sglang.test.test_utils import ( run_logprob_check, ) -register_cuda_ci(est_time=127, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=130, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=127, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=130, suite="stage-b-test-1-gpu-small-amd") class TestSRTEndpoint(CustomTestCase): diff --git a/test/registered/core/test_srt_engine.py b/test/registered/core/test_srt_engine.py index eb988a7ea..ab70eaaf1 100644 --- a/test/registered/core/test_srt_engine.py +++ b/test/registered/core/test_srt_engine.py @@ -22,8 +22,8 @@ from sglang.test.test_utils import ( CustomTestCase, ) -register_cuda_ci(est_time=252, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=261, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=252, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=261, suite="stage-b-test-1-gpu-small-amd") class TestSRTEngine(CustomTestCase): diff --git a/test/registered/debug_utils/comparator/aligner/entrypoint/test_executor.py b/test/registered/debug_utils/comparator/aligner/entrypoint/test_executor.py index fe10e02d5..0c807c1e6 100644 --- a/test/registered/debug_utils/comparator/aligner/entrypoint/test_executor.py +++ b/test/registered/debug_utils/comparator/aligner/entrypoint/test_executor.py @@ -28,7 +28,7 @@ from sglang.srt.debug_utils.comparator.dims_spec import ParallelAxis, TokenLayou from sglang.srt.debug_utils.comparator.utils import Pair from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=15, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=15, suite="stage-a-test-cpu", nightly=True) class TestExecuteSubPlans: diff --git a/test/registered/debug_utils/comparator/aligner/entrypoint/test_planner.py b/test/registered/debug_utils/comparator/aligner/entrypoint/test_planner.py index 16a5b135d..b7b9f89d7 100644 --- a/test/registered/debug_utils/comparator/aligner/entrypoint/test_planner.py +++ b/test/registered/debug_utils/comparator/aligner/entrypoint/test_planner.py @@ -25,7 +25,7 @@ from sglang.srt.debug_utils.comparator.dims_spec import TokenLayout from sglang.srt.debug_utils.comparator.utils import Pair from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=15, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=15, suite="stage-a-test-cpu", nightly=True) def _make_meta( diff --git a/test/registered/debug_utils/comparator/aligner/reorderer/test_executor.py b/test/registered/debug_utils/comparator/aligner/reorderer/test_executor.py index 47a5a5f40..e5b166232 100644 --- a/test/registered/debug_utils/comparator/aligner/reorderer/test_executor.py +++ b/test/registered/debug_utils/comparator/aligner/reorderer/test_executor.py @@ -22,7 +22,7 @@ from sglang.srt.debug_utils.comparator.aligner.unsharder.types import ( from sglang.srt.debug_utils.comparator.dims_spec import ParallelAxis from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) def _zigzag_order(cp_size: int) -> list[int]: diff --git a/test/registered/debug_utils/comparator/aligner/reorderer/test_planner.py b/test/registered/debug_utils/comparator/aligner/reorderer/test_planner.py index f429f2ba5..728df373d 100644 --- a/test/registered/debug_utils/comparator/aligner/reorderer/test_planner.py +++ b/test/registered/debug_utils/comparator/aligner/reorderer/test_planner.py @@ -24,7 +24,7 @@ from sglang.srt.debug_utils.comparator.dims_spec import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) class TestComputeReordererPlans: diff --git a/test/registered/debug_utils/comparator/aligner/test_axis_aligner.py b/test/registered/debug_utils/comparator/aligner/test_axis_aligner.py index 80964e170..df7591b81 100644 --- a/test/registered/debug_utils/comparator/aligner/test_axis_aligner.py +++ b/test/registered/debug_utils/comparator/aligner/test_axis_aligner.py @@ -13,7 +13,7 @@ from sglang.srt.debug_utils.comparator.log_sink import log_sink from sglang.srt.debug_utils.comparator.utils import Pair from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=15, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=15, suite="stage-a-test-cpu", nightly=True) class TestComputeAxisAlignerPlan: diff --git a/test/registered/debug_utils/comparator/aligner/token_aligner/test_aux_loader.py b/test/registered/debug_utils/comparator/aligner/token_aligner/test_aux_loader.py index 50ce2fb3b..9e33d0262 100644 --- a/test/registered/debug_utils/comparator/aligner/token_aligner/test_aux_loader.py +++ b/test/registered/debug_utils/comparator/aligner/token_aligner/test_aux_loader.py @@ -19,7 +19,7 @@ from sglang.srt.debug_utils.comparator.log_sink import LogSink from sglang.srt.debug_utils.comparator.output_types import ErrorLog, InfoLog from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=15, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=15, suite="stage-a-test-cpu", nightly=True) _sglang_plugin = _SGLangPlugin() _megatron_plugin = _MegatronPlugin() diff --git a/test/registered/debug_utils/comparator/aligner/token_aligner/test_aux_plugins.py b/test/registered/debug_utils/comparator/aligner/token_aligner/test_aux_plugins.py index 76e34cdf4..9b8cac698 100644 --- a/test/registered/debug_utils/comparator/aligner/token_aligner/test_aux_plugins.py +++ b/test/registered/debug_utils/comparator/aligner/token_aligner/test_aux_plugins.py @@ -16,7 +16,7 @@ from sglang.srt.debug_utils.comparator.aligner.token_aligner.smart.types import from sglang.srt.debug_utils.comparator.dims_spec import TokenLayout from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=15, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=15, suite="stage-a-test-cpu", nightly=True) _sglang_plugin = _SGLangPlugin() _megatron_plugin = _MegatronPlugin() diff --git a/test/registered/debug_utils/comparator/aligner/token_aligner/test_concat_steps.py b/test/registered/debug_utils/comparator/aligner/token_aligner/test_concat_steps.py index 2d89dd914..26b914e20 100644 --- a/test/registered/debug_utils/comparator/aligner/token_aligner/test_concat_steps.py +++ b/test/registered/debug_utils/comparator/aligner/token_aligner/test_concat_steps.py @@ -9,7 +9,7 @@ from sglang.srt.debug_utils.comparator.aligner.token_aligner.concat_steps import from sglang.srt.debug_utils.comparator.utils import Pair from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=15, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=15, suite="stage-a-test-cpu", nightly=True) class TestExecuteConcat: diff --git a/test/registered/debug_utils/comparator/aligner/token_aligner/test_executor.py b/test/registered/debug_utils/comparator/aligner/token_aligner/test_executor.py index de8b9aa01..41a2f6685 100644 --- a/test/registered/debug_utils/comparator/aligner/token_aligner/test_executor.py +++ b/test/registered/debug_utils/comparator/aligner/token_aligner/test_executor.py @@ -25,7 +25,7 @@ from sglang.srt.debug_utils.comparator.dims_spec import TokenLayout from sglang.srt.debug_utils.comparator.utils import Pair from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=15, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=15, suite="stage-a-test-cpu", nightly=True) def _named(tensor: torch.Tensor, names: list[str]) -> torch.Tensor: diff --git a/test/registered/debug_utils/comparator/aligner/token_aligner/test_planner.py b/test/registered/debug_utils/comparator/aligner/token_aligner/test_planner.py index c55673e0c..ce02794bf 100644 --- a/test/registered/debug_utils/comparator/aligner/token_aligner/test_planner.py +++ b/test/registered/debug_utils/comparator/aligner/token_aligner/test_planner.py @@ -23,7 +23,7 @@ from sglang.srt.debug_utils.comparator.dims_spec import TokenLayout from sglang.srt.debug_utils.comparator.utils import Pair from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=30, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=30, suite="stage-a-test-cpu", nightly=True) class TestBuildTokenIndexSGLangThd: diff --git a/test/registered/debug_utils/comparator/aligner/token_aligner/test_thd_seq_lens_loader.py b/test/registered/debug_utils/comparator/aligner/token_aligner/test_thd_seq_lens_loader.py index e72171315..56dc7c03d 100644 --- a/test/registered/debug_utils/comparator/aligner/token_aligner/test_thd_seq_lens_loader.py +++ b/test/registered/debug_utils/comparator/aligner/token_aligner/test_thd_seq_lens_loader.py @@ -14,7 +14,7 @@ from sglang.srt.debug_utils.comparator.aligner.token_aligner.smart.aux_plugins i ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=15, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=15, suite="stage-a-test-cpu", nightly=True) def _save_pt( diff --git a/test/registered/debug_utils/comparator/aligner/unsharder/test_executor.py b/test/registered/debug_utils/comparator/aligner/unsharder/test_executor.py index 1e8e9d0e1..4657b8f09 100644 --- a/test/registered/debug_utils/comparator/aligner/unsharder/test_executor.py +++ b/test/registered/debug_utils/comparator/aligner/unsharder/test_executor.py @@ -27,7 +27,7 @@ from sglang.srt.debug_utils.comparator.dims_spec import ( from sglang.srt.debug_utils.comparator.output_types import ReplicatedCheckResult from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) def _name_tensors( diff --git a/test/registered/debug_utils/comparator/aligner/unsharder/test_parallel_info.py b/test/registered/debug_utils/comparator/aligner/unsharder/test_parallel_info.py index 5c82d6d88..f7553ef1b 100644 --- a/test/registered/debug_utils/comparator/aligner/unsharder/test_parallel_info.py +++ b/test/registered/debug_utils/comparator/aligner/unsharder/test_parallel_info.py @@ -9,7 +9,7 @@ from sglang.srt.debug_utils.comparator.aligner.unsharder.types import AxisInfo from sglang.srt.debug_utils.comparator.dims_spec import ParallelAxis from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) class TestNormalizeParallelInfo: diff --git a/test/registered/debug_utils/comparator/aligner/unsharder/test_planner.py b/test/registered/debug_utils/comparator/aligner/unsharder/test_planner.py index 0887e7aff..cdf55f980 100644 --- a/test/registered/debug_utils/comparator/aligner/unsharder/test_planner.py +++ b/test/registered/debug_utils/comparator/aligner/unsharder/test_planner.py @@ -19,7 +19,7 @@ from sglang.srt.debug_utils.comparator.aligner.unsharder.types import ( from sglang.srt.debug_utils.comparator.dims_spec import ParallelAxis, parse_dims from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) class TestComputeUnsharderPlan: diff --git a/test/registered/debug_utils/comparator/dims_spec/test_dim_parser.py b/test/registered/debug_utils/comparator/dims_spec/test_dim_parser.py index adffcf9fb..caca5205e 100644 --- a/test/registered/debug_utils/comparator/dims_spec/test_dim_parser.py +++ b/test/registered/debug_utils/comparator/dims_spec/test_dim_parser.py @@ -12,7 +12,7 @@ from sglang.srt.debug_utils.comparator.dims_spec import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=5, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=5, suite="stage-a-test-cpu", nightly=True) class TestParseDim: diff --git a/test/registered/debug_utils/comparator/dims_spec/test_dims_parser.py b/test/registered/debug_utils/comparator/dims_spec/test_dims_parser.py index 22f5d3156..b4436f4e1 100644 --- a/test/registered/debug_utils/comparator/dims_spec/test_dims_parser.py +++ b/test/registered/debug_utils/comparator/dims_spec/test_dims_parser.py @@ -15,7 +15,7 @@ from sglang.srt.debug_utils.comparator.dims_spec import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=5, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=5, suite="stage-a-test-cpu", nightly=True) class TestSingletonDimUtilFilterOut: diff --git a/test/registered/debug_utils/comparator/dims_spec/test_tensor_naming.py b/test/registered/debug_utils/comparator/dims_spec/test_tensor_naming.py index b5156a344..761172505 100644 --- a/test/registered/debug_utils/comparator/dims_spec/test_tensor_naming.py +++ b/test/registered/debug_utils/comparator/dims_spec/test_tensor_naming.py @@ -13,7 +13,7 @@ from sglang.srt.debug_utils.comparator.dims_spec import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=5, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=5, suite="stage-a-test-cpu", nightly=True) class TestFindDimIndex: diff --git a/test/registered/debug_utils/comparator/dims_spec/test_types.py b/test/registered/debug_utils/comparator/dims_spec/test_types.py index 1792a9e55..5e799ae9d 100644 --- a/test/registered/debug_utils/comparator/dims_spec/test_types.py +++ b/test/registered/debug_utils/comparator/dims_spec/test_types.py @@ -9,7 +9,7 @@ from sglang.srt.debug_utils.comparator.dims_spec import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=5, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=5, suite="stage-a-test-cpu", nightly=True) class TestDimConstants: diff --git a/test/registered/debug_utils/comparator/tensor_comparator/test_comparator.py b/test/registered/debug_utils/comparator/tensor_comparator/test_comparator.py index 5bea61bf0..6e7bc6e68 100644 --- a/test/registered/debug_utils/comparator/tensor_comparator/test_comparator.py +++ b/test/registered/debug_utils/comparator/tensor_comparator/test_comparator.py @@ -14,7 +14,7 @@ from sglang.srt.debug_utils.comparator.tensor_comparator.comparator import ( from sglang.srt.debug_utils.comparator.tensor_comparator.types import DiffInfo from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=20, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=20, suite="stage-a-test-cpu", nightly=True) class TestComputeTensorInfo: diff --git a/test/registered/debug_utils/comparator/tensor_comparator/test_formatter.py b/test/registered/debug_utils/comparator/tensor_comparator/test_formatter.py index a75e0f67b..83f95852d 100644 --- a/test/registered/debug_utils/comparator/tensor_comparator/test_formatter.py +++ b/test/registered/debug_utils/comparator/tensor_comparator/test_formatter.py @@ -58,7 +58,7 @@ from sglang.srt.debug_utils.comparator.tensor_comparator.types import ( from sglang.srt.debug_utils.comparator.utils import Pair from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) _DEFAULT_PERCENTILE_LINES: list[str] = [ " [blue]p1 [/] -1.8000 -1.8000 [dim]+0.00e+00[/]", diff --git a/test/registered/debug_utils/comparator/tensor_comparator/test_types.py b/test/registered/debug_utils/comparator/tensor_comparator/test_types.py index 593e3b110..6b43456e2 100644 --- a/test/registered/debug_utils/comparator/tensor_comparator/test_types.py +++ b/test/registered/debug_utils/comparator/tensor_comparator/test_types.py @@ -21,7 +21,7 @@ from sglang.srt.debug_utils.comparator.tensor_comparator.types import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) def _make_stats(**overrides) -> TensorStats: diff --git a/test/registered/debug_utils/comparator/test_bundle_comparator.py b/test/registered/debug_utils/comparator/test_bundle_comparator.py index 17cdbd81c..901f885a8 100644 --- a/test/registered/debug_utils/comparator/test_bundle_comparator.py +++ b/test/registered/debug_utils/comparator/test_bundle_comparator.py @@ -15,7 +15,7 @@ from sglang.srt.debug_utils.comparator.utils import Pair from sglang.srt.debug_utils.dump_loader import ValueWithMeta from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=15, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=15, suite="stage-a-test-cpu", nightly=True) def _save_tensor( diff --git a/test/registered/debug_utils/comparator/test_bundle_matcher.py b/test/registered/debug_utils/comparator/test_bundle_matcher.py index 65a56f31c..248d52d23 100644 --- a/test/registered/debug_utils/comparator/test_bundle_matcher.py +++ b/test/registered/debug_utils/comparator/test_bundle_matcher.py @@ -13,7 +13,7 @@ from sglang.srt.debug_utils.comparator.bundle_matcher import ( from sglang.srt.debug_utils.comparator.utils import Pair from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=15, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=15, suite="stage-a-test-cpu", nightly=True) def _make_row( diff --git a/test/registered/debug_utils/comparator/test_display.py b/test/registered/debug_utils/comparator/test_display.py index 0c7bfa44f..c3789bb9c 100644 --- a/test/registered/debug_utils/comparator/test_display.py +++ b/test/registered/debug_utils/comparator/test_display.py @@ -21,7 +21,7 @@ from sglang.srt.debug_utils.comparator.output_types import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) def _render_rich(renderable: object) -> str: diff --git a/test/registered/debug_utils/comparator/test_dp_utils.py b/test/registered/debug_utils/comparator/test_dp_utils.py index 8b74ed13c..db32246ba 100644 --- a/test/registered/debug_utils/comparator/test_dp_utils.py +++ b/test/registered/debug_utils/comparator/test_dp_utils.py @@ -12,7 +12,7 @@ from sglang.srt.debug_utils.comparator.dp_utils import ( from sglang.srt.debug_utils.dump_loader import ValueWithMeta from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=15, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=15, suite="stage-a-test-cpu", nightly=True) def _make_sglang_meta( diff --git a/test/registered/debug_utils/comparator/test_dump_loader.py b/test/registered/debug_utils/comparator/test_dump_loader.py index 7e88fc6ac..26a68d066 100644 --- a/test/registered/debug_utils/comparator/test_dump_loader.py +++ b/test/registered/debug_utils/comparator/test_dump_loader.py @@ -7,7 +7,7 @@ import torch from sglang.srt.debug_utils.dump_loader import read_tokenizer_path from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) def _save_pt( diff --git a/test/registered/debug_utils/comparator/test_entrypoint.py b/test/registered/debug_utils/comparator/test_entrypoint.py index f82a8de9b..a98ca6be1 100644 --- a/test/registered/debug_utils/comparator/test_entrypoint.py +++ b/test/registered/debug_utils/comparator/test_entrypoint.py @@ -30,7 +30,7 @@ from sglang.srt.debug_utils.comparator.output_types import ( from sglang.srt.debug_utils.dumper import DumperConfig, _Dumper, _RecomputeStatus from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=30, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=30, suite="stage-a-test-cpu", nightly=True) _FIXED_EXP_NAME = "my_exp_name" diff --git a/test/registered/debug_utils/comparator/test_log_sink.py b/test/registered/debug_utils/comparator/test_log_sink.py index 66c054f52..90a960d93 100644 --- a/test/registered/debug_utils/comparator/test_log_sink.py +++ b/test/registered/debug_utils/comparator/test_log_sink.py @@ -11,7 +11,7 @@ from sglang.srt.debug_utils.comparator.output_types import ( from sglang.srt.debug_utils.comparator.report_sink import report_sink from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) def _make_error_log(**overrides) -> ErrorLog: diff --git a/test/registered/debug_utils/comparator/test_manually_verify.py b/test/registered/debug_utils/comparator/test_manually_verify.py index dd237cd00..d69c939d7 100644 --- a/test/registered/debug_utils/comparator/test_manually_verify.py +++ b/test/registered/debug_utils/comparator/test_manually_verify.py @@ -23,7 +23,7 @@ import torch from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=60, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=60, suite="stage-a-test-cpu", nightly=True) _PUBLISH_DIR: Path = Path("/tmp/comparator_manual_verify") _PNG_MAGIC: bytes = b"\x89PNG" diff --git a/test/registered/debug_utils/comparator/test_meta_overrider.py b/test/registered/debug_utils/comparator/test_meta_overrider.py index afbacea32..fe2746f8c 100644 --- a/test/registered/debug_utils/comparator/test_meta_overrider.py +++ b/test/registered/debug_utils/comparator/test_meta_overrider.py @@ -16,7 +16,7 @@ from sglang.srt.debug_utils.comparator.meta_overrider import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) # ───────────────────── Unit: MetaOverrideRule ───────────────────── diff --git a/test/registered/debug_utils/comparator/test_model_validation.py b/test/registered/debug_utils/comparator/test_model_validation.py index 1fc372bed..cb88f72f9 100644 --- a/test/registered/debug_utils/comparator/test_model_validation.py +++ b/test/registered/debug_utils/comparator/test_model_validation.py @@ -44,7 +44,7 @@ from sglang.srt.debug_utils.comparator.tensor_comparator.types import ( from sglang.srt.debug_utils.comparator.utils import Pair, _check_equal_lengths from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) class TestCheckEqualLengths: diff --git a/test/registered/debug_utils/comparator/test_output_types.py b/test/registered/debug_utils/comparator/test_output_types.py index 100e03463..2ddbca1ed 100644 --- a/test/registered/debug_utils/comparator/test_output_types.py +++ b/test/registered/debug_utils/comparator/test_output_types.py @@ -55,7 +55,7 @@ from sglang.srt.debug_utils.comparator.output_types import ( from sglang.srt.debug_utils.comparator.utils import Pair from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) def _render_rich(renderable: object) -> str: diff --git a/test/registered/debug_utils/comparator/test_per_token_visualizer.py b/test/registered/debug_utils/comparator/test_per_token_visualizer.py index 31b55d137..a12ef1211 100644 --- a/test/registered/debug_utils/comparator/test_per_token_visualizer.py +++ b/test/registered/debug_utils/comparator/test_per_token_visualizer.py @@ -15,7 +15,7 @@ from sglang.srt.debug_utils.comparator.tensor_comparator.comparator import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=30, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=30, suite="stage-a-test-cpu", nightly=True) _PNG_MAGIC: bytes = b"\x89PNG" diff --git a/test/registered/debug_utils/comparator/test_preset.py b/test/registered/debug_utils/comparator/test_preset.py index 5a06497c5..e6b99df65 100644 --- a/test/registered/debug_utils/comparator/test_preset.py +++ b/test/registered/debug_utils/comparator/test_preset.py @@ -3,7 +3,7 @@ import pytest from sglang.srt.debug_utils.comparator.preset import PRESETS, expand_preset from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=5, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=5, suite="stage-a-test-cpu", nightly=True) class TestExpandPreset: diff --git a/test/registered/debug_utils/comparator/test_utils.py b/test/registered/debug_utils/comparator/test_utils.py index 74adc6081..df2677cab 100644 --- a/test/registered/debug_utils/comparator/test_utils.py +++ b/test/registered/debug_utils/comparator/test_utils.py @@ -17,7 +17,7 @@ from sglang.srt.debug_utils.comparator.utils import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) class TestCalcRelDiff: diff --git a/test/registered/debug_utils/comparator/test_visualizer.py b/test/registered/debug_utils/comparator/test_visualizer.py index d498b5dd8..c6aba96a9 100644 --- a/test/registered/debug_utils/comparator/test_visualizer.py +++ b/test/registered/debug_utils/comparator/test_visualizer.py @@ -10,7 +10,7 @@ from sglang.srt.debug_utils.comparator.visualizer.preprocessing import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=30, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=30, suite="stage-a-test-cpu", nightly=True) class TestPreprocessTensor: diff --git a/test/registered/debug_utils/comparator/testing_helpers.py b/test/registered/debug_utils/comparator/testing_helpers.py index ffe282a19..c2e20c679 100644 --- a/test/registered/debug_utils/comparator/testing_helpers.py +++ b/test/registered/debug_utils/comparator/testing_helpers.py @@ -12,7 +12,7 @@ from sglang.test.ci.ci_register import register_cpu_ci register_cpu_ci( est_time=0, - suite="stage-a-cpu-only", + suite="stage-a-test-cpu", nightly=True, disabled="helper module, no tests", ) diff --git a/test/registered/debug_utils/source_patcher/test_code_patcher.py b/test/registered/debug_utils/source_patcher/test_code_patcher.py index d647cef75..7d791788d 100644 --- a/test/registered/debug_utils/source_patcher/test_code_patcher.py +++ b/test/registered/debug_utils/source_patcher/test_code_patcher.py @@ -10,7 +10,7 @@ from sglang.srt.debug_utils.source_patcher.code_patcher import ( from sglang.srt.debug_utils.source_patcher.types import EditSpec, PatchSpec from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) SAMPLE_MODULE_NAME = "_source_patcher_test_fixtures.sample_module" diff --git a/test/registered/debug_utils/source_patcher/test_dumper_integration.py b/test/registered/debug_utils/source_patcher/test_dumper_integration.py index cd15b3d69..af9d8d077 100644 --- a/test/registered/debug_utils/source_patcher/test_dumper_integration.py +++ b/test/registered/debug_utils/source_patcher/test_dumper_integration.py @@ -8,7 +8,7 @@ import yaml from sglang.srt.debug_utils.dumper import DumperConfig, _Dumper from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) SAMPLE_MODULE_NAME = "_source_patcher_test_fixtures.sample_module" diff --git a/test/registered/debug_utils/source_patcher/test_source_editor.py b/test/registered/debug_utils/source_patcher/test_source_editor.py index 616e289e0..560563dde 100644 --- a/test/registered/debug_utils/source_patcher/test_source_editor.py +++ b/test/registered/debug_utils/source_patcher/test_source_editor.py @@ -5,7 +5,7 @@ from sglang.srt.debug_utils.source_patcher.source_editor import apply_edits from sglang.srt.debug_utils.source_patcher.types import EditSpec, PatchApplicationError from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=10, suite="stage-a-test-cpu", nightly=True) class TestApplyEdits: diff --git a/test/registered/debug_utils/test_cuda_coredump_smoke.py b/test/registered/debug_utils/test_cuda_coredump_smoke.py index c90ecd38e..99ebb7f02 100644 --- a/test/registered/debug_utils/test_cuda_coredump_smoke.py +++ b/test/registered/debug_utils/test_cuda_coredump_smoke.py @@ -12,7 +12,7 @@ from sglang.test.ci.ci_register import register_cuda_ci register_cuda_ci( est_time=10, - suite="stage-a-test-small-1-gpu", + suite="stage-a-test-1-gpu-small", disabled="Manual only: triggers intentional CUDA crash for coredump verification", ) diff --git a/test/registered/debug_utils/test_dump_comparator.py b/test/registered/debug_utils/test_dump_comparator.py index 82115dcdb..8b2d66d54 100644 --- a/test/registered/debug_utils/test_dump_comparator.py +++ b/test/registered/debug_utils/test_dump_comparator.py @@ -14,7 +14,7 @@ from sglang.srt.debug_utils.dump_comparator import ( from sglang.srt.debug_utils.dumper import DumperConfig, _Dumper from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=30, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=30, suite="stage-a-test-cpu", nightly=True) # ----------------------------- Unit tests ----------------------------- diff --git a/test/registered/debug_utils/test_dump_loader.py b/test/registered/debug_utils/test_dump_loader.py index dccd15fee..fed3a5e8a 100644 --- a/test/registered/debug_utils/test_dump_loader.py +++ b/test/registered/debug_utils/test_dump_loader.py @@ -15,7 +15,7 @@ from sglang.srt.debug_utils.dump_loader import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=30, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=30, suite="stage-a-test-cpu", nightly=True) class TestReadMeta: diff --git a/test/registered/debug_utils/test_schedule_simulator.py b/test/registered/debug_utils/test_schedule_simulator.py index b4c5d7dfa..dc7dc35c9 100644 --- a/test/registered/debug_utils/test_schedule_simulator.py +++ b/test/registered/debug_utils/test_schedule_simulator.py @@ -25,7 +25,7 @@ from sglang.srt.debug_utils.schedule_simulator import ( from sglang.test.ci.ci_register import register_cpu_ci from sglang.test.test_utils import CustomTestCase -register_cpu_ci(est_time=120, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=120, suite="stage-a-test-cpu", nightly=True) # ==================== Non-E2E Tests ==================== diff --git a/test/registered/debug_utils/test_tensor_dump_forward_hook.py b/test/registered/debug_utils/test_tensor_dump_forward_hook.py index 8ed8de03b..a4176eecd 100644 --- a/test/registered/debug_utils/test_tensor_dump_forward_hook.py +++ b/test/registered/debug_utils/test_tensor_dump_forward_hook.py @@ -19,12 +19,12 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci register_cuda_ci( est_time=9, - suite="stage-b-test-small-1-gpu", + suite="stage-b-test-1-gpu-small", disabled="Test uses pytest-style function without TestCase class - see #17145", ) register_amd_ci( est_time=15, - suite="stage-b-test-small-1-gpu-amd", + suite="stage-b-test-1-gpu-small-amd", disabled="Test uses pytest-style function without TestCase class - see #17145", ) diff --git a/test/registered/disaggregation/test_disaggregation_basic.py b/test/registered/disaggregation/test_disaggregation_basic.py index 568196aec..ed06bc476 100644 --- a/test/registered/disaggregation/test_disaggregation_basic.py +++ b/test/registered/disaggregation/test_disaggregation_basic.py @@ -20,7 +20,7 @@ from sglang.test.test_utils import ( popen_launch_pd_server, ) -register_cuda_ci(est_time=400, suite="stage-b-test-large-2-gpu") +register_cuda_ci(est_time=400, suite="stage-b-test-2-gpu-large") class TestDisaggregationAccuracy(PDDisaggregationServerBase): diff --git a/test/registered/disaggregation/test_disaggregation_decode_offload.py b/test/registered/disaggregation/test_disaggregation_decode_offload.py index cc22601b8..c7e9235ce 100644 --- a/test/registered/disaggregation/test_disaggregation_decode_offload.py +++ b/test/registered/disaggregation/test_disaggregation_decode_offload.py @@ -18,7 +18,7 @@ from sglang.test.test_utils import ( # Registering the test for CUDA CI with appropriate parameters # Increasing estimated time since we run evaluation twice -register_cuda_ci(est_time=600, suite="stage-b-test-large-2-gpu") +register_cuda_ci(est_time=600, suite="stage-b-test-2-gpu-large") @unittest.skipIf(is_in_ci(), "Temporarily disable the flaky test.") diff --git a/test/registered/disaggregation/test_specv2_kvcache_offloading.py b/test/registered/disaggregation/test_specv2_kvcache_offloading.py index 64afa7b4a..5c403ba2e 100644 --- a/test/registered/disaggregation/test_specv2_kvcache_offloading.py +++ b/test/registered/disaggregation/test_specv2_kvcache_offloading.py @@ -17,7 +17,7 @@ from sglang.srt.disaggregation.decode_kvcache_offload_manager import ( ) from sglang.test.ci.ci_register import register_cuda_ci -register_cuda_ci(est_time=10, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=10, suite="stage-b-test-1-gpu-small") def _make_mock_req( diff --git a/test/registered/distributed/test_data_parallelism.py b/test/registered/distributed/test_data_parallelism.py index 25eba4a36..b93af4584 100644 --- a/test/registered/distributed/test_data_parallelism.py +++ b/test/registered/distributed/test_data_parallelism.py @@ -15,8 +15,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=73, suite="stage-b-test-large-2-gpu") -register_amd_ci(est_time=73, suite="stage-b-test-large-2-gpu-amd") +register_cuda_ci(est_time=73, suite="stage-b-test-2-gpu-large") +register_amd_ci(est_time=73, suite="stage-b-test-2-gpu-large-amd") class TestDataParallelism(CustomTestCase): diff --git a/test/registered/distributed/test_dp_attention.py b/test/registered/distributed/test_dp_attention.py index 4d9d1fb1f..0d7f1838c 100644 --- a/test/registered/distributed/test_dp_attention.py +++ b/test/registered/distributed/test_dp_attention.py @@ -25,7 +25,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=350, suite="stage-b-test-large-2-gpu") +register_cuda_ci(est_time=350, suite="stage-b-test-2-gpu-large") class TestDPAttentionDP2TP2( diff --git a/test/registered/distributed/test_load_weights_from_remote_instance.py b/test/registered/distributed/test_load_weights_from_remote_instance.py index 128b9a86f..f1080caeb 100644 --- a/test/registered/distributed/test_load_weights_from_remote_instance.py +++ b/test/registered/distributed/test_load_weights_from_remote_instance.py @@ -38,8 +38,8 @@ from sglang.utils import terminate_process mp.set_start_method("spawn", force=True) -register_cuda_ci(est_time=72, suite="stage-b-test-large-2-gpu") -register_amd_ci(est_time=72, suite="stage-b-test-large-2-gpu-amd") +register_cuda_ci(est_time=72, suite="stage-b-test-2-gpu-large") +register_amd_ci(est_time=72, suite="stage-b-test-2-gpu-large-amd") def verify_params_close(params1, params2, error_msg): diff --git a/test/registered/distributed/test_parallel_state.py b/test/registered/distributed/test_parallel_state.py index 6bfbb1073..900568139 100644 --- a/test/registered/distributed/test_parallel_state.py +++ b/test/registered/distributed/test_parallel_state.py @@ -43,7 +43,7 @@ import pytest from sglang.test.ci.ci_register import register_cuda_ci -register_cuda_ci(est_time=5, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=5, suite="stage-b-test-1-gpu-small") # Import the actual parallel_state module parallel_state = pytest.importorskip("sglang.srt.distributed.parallel_state") diff --git a/test/registered/dllm/test_llada2_mini.py b/test/registered/dllm/test_llada2_mini.py index a853e0a9b..3bec211ea 100644 --- a/test/registered/dllm/test_llada2_mini.py +++ b/test/registered/dllm/test_llada2_mini.py @@ -1,7 +1,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=181, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=330, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=181, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=330, suite="stage-b-test-1-gpu-small-amd") import unittest from types import SimpleNamespace diff --git a/test/registered/dllm/test_llada2_mini_amd.py b/test/registered/dllm/test_llada2_mini_amd.py index ad7aeae40..68e0cfec9 100644 --- a/test/registered/dllm/test_llada2_mini_amd.py +++ b/test/registered/dllm/test_llada2_mini_amd.py @@ -20,7 +20,7 @@ from sglang.test.test_utils import ( write_github_step_summary, ) -register_amd_ci(est_time=1000, suite="stage-b-test-small-1-gpu-amd") +register_amd_ci(est_time=1000, suite="stage-b-test-1-gpu-small-amd") class TestLLaDA2MiniAMD(CustomTestCase): diff --git a/test/registered/embedding/test_embedding_models.py b/test/registered/embedding/test_embedding_models.py index a496ab55d..4a9c43a95 100644 --- a/test/registered/embedding/test_embedding_models.py +++ b/test/registered/embedding/test_embedding_models.py @@ -32,10 +32,10 @@ from sglang.test.test_utils import ( # Embedding model tests register_amd_ci( est_time=73, - suite="stage-b-test-small-1-gpu-amd", + suite="stage-b-test-1-gpu-small-amd", disabled="see https://github.com/sgl-project/sglang/issues/11127", ) -register_cuda_ci(est_time=73, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=73, suite="stage-b-test-1-gpu-small") MODEL_TO_CONFIG = { "Alibaba-NLP/gte-Qwen2-1.5B-instruct": (1, 1e-5), diff --git a/test/registered/embedding/test_encoder_embedding_models.py b/test/registered/embedding/test_encoder_embedding_models.py index 721ec7932..115b26b2c 100644 --- a/test/registered/embedding/test_encoder_embedding_models.py +++ b/test/registered/embedding/test_encoder_embedding_models.py @@ -29,7 +29,7 @@ from sglang.test.test_utils import CustomTestCase, get_similarities, is_in_ci # python -m unittest test_encoder_embedding_models.TestEncoderEmbeddingModels.test_prefill_logits -register_cuda_ci(est_time=270, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=270, suite="stage-b-test-1-gpu-small") MODELS = [("BAAI/bge-small-en", 1, 1e-5), ("BAAI/bge-m3", 1, 1e-5)] diff --git a/test/registered/embedding/test_input_embeddings.py b/test/registered/embedding/test_input_embeddings.py index bf80328f8..cbfc3dc19 100644 --- a/test/registered/embedding/test_input_embeddings.py +++ b/test/registered/embedding/test_input_embeddings.py @@ -16,8 +16,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=38, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=38, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=38, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=38, suite="stage-b-test-1-gpu-small-amd") class TestInputEmbeds(CustomTestCase): diff --git a/test/registered/embedding/test_input_embeds_chunked.py b/test/registered/embedding/test_input_embeds_chunked.py index 8215f2378..13fc92cea 100644 --- a/test/registered/embedding/test_input_embeds_chunked.py +++ b/test/registered/embedding/test_input_embeds_chunked.py @@ -30,7 +30,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=45, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=45, suite="stage-b-test-1-gpu-small") CHUNKED_PREFILL_SIZE = 256 diff --git a/test/registered/embedding/test_openai_embedding.py b/test/registered/embedding/test_openai_embedding.py index 1432a5307..3c13198e8 100644 --- a/test/registered/embedding/test_openai_embedding.py +++ b/test/registered/embedding/test_openai_embedding.py @@ -13,8 +13,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=70, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=141, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=70, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=141, suite="stage-b-test-1-gpu-small-amd") class TestOpenAIEmbedding(CustomTestCase): diff --git a/test/registered/ep/test_deepep_small.py b/test/registered/ep/test_deepep_small.py index 78c42e1db..911915aaf 100644 --- a/test/registered/ep/test_deepep_small.py +++ b/test/registered/ep/test_deepep_small.py @@ -16,7 +16,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=531, suite="stage-c-test-deepep-4-gpu") +register_cuda_ci(est_time=531, suite="stage-c-test-deepep-4-gpu-h100") class TestPureDP(CustomTestCase): diff --git a/test/registered/ep/test_mooncake_ep_small.py b/test/registered/ep/test_mooncake_ep_small.py index 16cc4622c..a59a44c31 100644 --- a/test/registered/ep/test_mooncake_ep_small.py +++ b/test/registered/ep/test_mooncake_ep_small.py @@ -15,7 +15,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=660, suite="stage-c-test-deepep-4-gpu") +register_cuda_ci(est_time=660, suite="stage-c-test-deepep-4-gpu-h100") ib_devices = get_rdma_devices_args() diff --git a/test/registered/eval/test_eval_accuracy_large.py b/test/registered/eval/test_eval_accuracy_large.py index 6a0a39581..fa6b40bd0 100644 --- a/test/registered/eval/test_eval_accuracy_large.py +++ b/test/registered/eval/test_eval_accuracy_large.py @@ -20,8 +20,8 @@ from sglang.test.test_utils import ( write_github_step_summary, ) -register_cuda_ci(est_time=300, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=420, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=300, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=420, suite="stage-b-test-1-gpu-small-amd") class TestEvalAccuracyLarge(CustomTestCase): diff --git a/test/registered/eval/test_moe_eval_accuracy_large.py b/test/registered/eval/test_moe_eval_accuracy_large.py index 369835ef4..759b66ad1 100644 --- a/test/registered/eval/test_moe_eval_accuracy_large.py +++ b/test/registered/eval/test_moe_eval_accuracy_large.py @@ -21,8 +21,8 @@ from sglang.test.test_utils import ( write_github_step_summary, ) -register_cuda_ci(est_time=500, suite="stage-b-test-large-2-gpu") -register_amd_ci(est_time=500, suite="stage-b-test-large-2-gpu-amd") +register_cuda_ci(est_time=500, suite="stage-b-test-2-gpu-large") +register_amd_ci(est_time=500, suite="stage-b-test-2-gpu-large-amd") class TestMoEEvalAccuracyLarge(CustomTestCase): diff --git a/test/registered/function_call/test_kimik2_detector.py b/test/registered/function_call/test_kimik2_detector.py index a2ec93ea9..158223958 100644 --- a/test/registered/function_call/test_kimik2_detector.py +++ b/test/registered/function_call/test_kimik2_detector.py @@ -11,7 +11,7 @@ from sglang.srt.function_call.kimik2_detector import ( from sglang.srt.parser.reasoning_parser import KimiK2Detector as KimiK2ReasoningDetector from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(1.0, "stage-a-cpu-only") +register_cpu_ci(1.0, "stage-a-test-cpu") def _make_tool(name, parameters=None): diff --git a/test/registered/hicache/test_hicache_storage.py b/test/registered/hicache/test_hicache_storage.py index 77a6ed326..00640de91 100644 --- a/test/registered/hicache/test_hicache_storage.py +++ b/test/registered/hicache/test_hicache_storage.py @@ -1,7 +1,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=96, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=300, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=96, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=300, suite="stage-b-test-1-gpu-small-amd") import time import unittest diff --git a/test/registered/hicache/test_hicache_storage_3fs_backend.py b/test/registered/hicache/test_hicache_storage_3fs_backend.py index 8a9cb6e06..6c478767a 100644 --- a/test/registered/hicache/test_hicache_storage_3fs_backend.py +++ b/test/registered/hicache/test_hicache_storage_3fs_backend.py @@ -13,8 +13,8 @@ from test_hicache_storage_file_backend import HiCacheStorageBaseMixin from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=200, suite="stage-b-test-large-2-gpu") -register_amd_ci(est_time=300, suite="stage-b-test-large-2-gpu") +register_cuda_ci(est_time=200, suite="stage-b-test-2-gpu-large") +register_amd_ci(est_time=300, suite="stage-b-test-2-gpu-large") class HiCacheStorage3FSBackendBaseMixin(HiCacheStorageBaseMixin): diff --git a/test/registered/hicache/test_hicache_storage_file_backend.py b/test/registered/hicache/test_hicache_storage_file_backend.py index acd28dbb3..30d1239b7 100644 --- a/test/registered/hicache/test_hicache_storage_file_backend.py +++ b/test/registered/hicache/test_hicache_storage_file_backend.py @@ -32,8 +32,8 @@ from sglang.test.test_utils import ( ) from sglang.utils import wait_for_http_ready -register_cuda_ci(est_time=200, suite="stage-b-test-large-2-gpu") -register_amd_ci(est_time=526, suite="stage-b-test-large-2-gpu-amd") +register_cuda_ci(est_time=200, suite="stage-b-test-2-gpu-large") +register_amd_ci(est_time=526, suite="stage-b-test-2-gpu-large-amd") class HiCacheStorageBaseMixin: diff --git a/test/registered/hicache/test_hicache_storage_mooncake_backend.py b/test/registered/hicache/test_hicache_storage_mooncake_backend.py index 947bce776..066f14cb8 100644 --- a/test/registered/hicache/test_hicache_storage_mooncake_backend.py +++ b/test/registered/hicache/test_hicache_storage_mooncake_backend.py @@ -20,7 +20,7 @@ from sglang.test.test_utils import ( is_in_ci, ) -register_cuda_ci(est_time=300, suite="stage-b-test-large-2-gpu") +register_cuda_ci(est_time=300, suite="stage-b-test-2-gpu-large") class HiCacheStorageMooncakeBackendBaseMixin(HiCacheStorageBaseMixin): diff --git a/test/registered/hicache/test_hicache_storage_runtime_attach_detach.py b/test/registered/hicache/test_hicache_storage_runtime_attach_detach.py index 28094e850..ea5a6a861 100644 --- a/test/registered/hicache/test_hicache_storage_runtime_attach_detach.py +++ b/test/registered/hicache/test_hicache_storage_runtime_attach_detach.py @@ -28,7 +28,7 @@ from sglang.test.test_utils import ( ) from sglang.utils import wait_for_http_ready -register_cuda_ci(est_time=200, suite="stage-b-test-large-2-gpu") +register_cuda_ci(est_time=200, suite="stage-b-test-2-gpu-large") class TestHiCacheStorageRuntimeAttachDetach(CustomTestCase): diff --git a/test/registered/hicache/test_hicache_variants.py b/test/registered/hicache/test_hicache_variants.py index 9762ea6dc..6934a32fc 100644 --- a/test/registered/hicache/test_hicache_variants.py +++ b/test/registered/hicache/test_hicache_variants.py @@ -1,7 +1,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=524, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=524, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=524, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=524, suite="stage-b-test-1-gpu-small-amd") """ Consolidated HiCache variant tests. Tests HiCache with different configurations: standard, MLA, EAGLE, and page size variants. diff --git a/test/registered/kernels/test_nsa_indexer.py b/test/registered/kernels/test_nsa_indexer.py index abee474f5..789baed00 100644 --- a/test/registered/kernels/test_nsa_indexer.py +++ b/test/registered/kernels/test_nsa_indexer.py @@ -24,7 +24,7 @@ from sglang.srt.model_executor.forward_batch_info import ForwardBatch, ForwardMo from sglang.srt.server_args import ServerArgs, set_global_server_args_for_scheduler from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=2, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=2, suite="stage-b-test-1-gpu-large") # Global configuration for all indexer tests DEFAULT_CONFIG = { diff --git a/test/registered/language/test_srt_backend.py b/test/registered/language/test_srt_backend.py index e1e9d20d5..99620e2d2 100644 --- a/test/registered/language/test_srt_backend.py +++ b/test/registered/language/test_srt_backend.py @@ -19,8 +19,8 @@ from sglang.test.test_programs import ( ) from sglang.test.test_utils import DEFAULT_MODEL_NAME_FOR_TEST, CustomTestCase -register_cuda_ci(est_time=80, suite="stage-a-test-small-1-gpu") -register_amd_ci(est_time=120, suite="stage-a-test-small-1-gpu-amd") +register_cuda_ci(est_time=80, suite="stage-a-test-1-gpu-small") +register_amd_ci(est_time=120, suite="stage-a-test-1-gpu-small-amd") class TestSRTBackend(CustomTestCase): diff --git a/test/registered/layers/mamba/test_causal_conv1d.py b/test/registered/layers/mamba/test_causal_conv1d.py index c777226ea..820a6ff6f 100644 --- a/test/registered/layers/mamba/test_causal_conv1d.py +++ b/test/registered/layers/mamba/test_causal_conv1d.py @@ -1,7 +1,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=25, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=25, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=25, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=25, suite="stage-b-test-1-gpu-small-amd") # Adapted from https://github.com/vllm-project/vllm/blob/main/tests/kernels/mamba/test_causal_conv1d.py diff --git a/test/registered/layers/mamba/test_mamba2_mixer.py b/test/registered/layers/mamba/test_mamba2_mixer.py index bbc145103..d63179c1a 100644 --- a/test/registered/layers/mamba/test_mamba2_mixer.py +++ b/test/registered/layers/mamba/test_mamba2_mixer.py @@ -15,7 +15,7 @@ from sglang.srt.distributed.parallel_state import ( ) from sglang.test.ci.ci_register import register_cuda_ci -register_cuda_ci(est_time=50, suite="stage-b-test-large-2-gpu") +register_cuda_ci(est_time=50, suite="stage-b-test-2-gpu-large") NUM_GPUS = 2 diff --git a/test/registered/layers/mamba/test_mamba_ssm.py b/test/registered/layers/mamba/test_mamba_ssm.py index 0178bfa2c..9847684a5 100644 --- a/test/registered/layers/mamba/test_mamba_ssm.py +++ b/test/registered/layers/mamba/test_mamba_ssm.py @@ -1,7 +1,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=7, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=20, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=7, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=20, suite="stage-b-test-1-gpu-small-amd") # Adapted from https://github.com/vllm-project/vllm/blob/633f943e30a4444d890d26b81850f7217736f840/tests/kernels/mamba/test_mamba_ssm_ssd.py diff --git a/test/registered/layers/mamba/test_mamba_ssm_ssd.py b/test/registered/layers/mamba/test_mamba_ssm_ssd.py index 02b5f9f1e..670bceabb 100644 --- a/test/registered/layers/mamba/test_mamba_ssm_ssd.py +++ b/test/registered/layers/mamba/test_mamba_ssm_ssd.py @@ -1,7 +1,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=15, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=34, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=15, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=34, suite="stage-b-test-1-gpu-small-amd") # Adapted from https://github.com/vllm-project/vllm/blob/633f943e30a4444d890d26b81850f7217736f840/tests/kernels/mamba/test_mamba_ssm_ssd.py diff --git a/test/registered/layers/test_fla_layernorm_guard.py b/test/registered/layers/test_fla_layernorm_guard.py index 2d305cb62..c6b910a7d 100644 --- a/test/registered/layers/test_fla_layernorm_guard.py +++ b/test/registered/layers/test_fla_layernorm_guard.py @@ -18,7 +18,7 @@ from sglang.test.ci.ci_register import register_cuda_ci register_cuda_ci( est_time=60, - suite="stage-b-test-large-2-gpu", + suite="stage-b-test-2-gpu-large", disabled="Temporarily disabled", ) diff --git a/test/registered/lora/test_fused_moe_lora_kernel.py b/test/registered/lora/test_fused_moe_lora_kernel.py index 4f8664624..cf3bc1f2c 100644 --- a/test/registered/lora/test_fused_moe_lora_kernel.py +++ b/test/registered/lora/test_fused_moe_lora_kernel.py @@ -14,7 +14,7 @@ from sglang.test.ci.ci_register import register_cuda_ci # ============================================================================== -register_cuda_ci(est_time=120, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=120, suite="stage-b-test-1-gpu-large") def round_up(x, base): diff --git a/test/registered/lora/test_lora_backend.py b/test/registered/lora/test_lora_backend.py index 0cc2aca9b..5d76cfced 100644 --- a/test/registered/lora/test_lora_backend.py +++ b/test/registered/lora/test_lora_backend.py @@ -29,10 +29,10 @@ from sglang.test.lora_utils import ( ) from sglang.test.test_utils import CustomTestCase, is_in_ci -register_cuda_ci(est_time=200, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=200, suite="stage-b-test-1-gpu-small") register_amd_ci( est_time=200, - suite="stage-b-test-small-1-gpu-amd", + suite="stage-b-test-1-gpu-small-amd", disabled="see https://github.com/sgl-project/sglang/issues/13107", ) diff --git a/test/registered/lora/test_lora_eviction.py b/test/registered/lora/test_lora_eviction.py index 3ed9ea176..5a765e9f9 100644 --- a/test/registered/lora/test_lora_eviction.py +++ b/test/registered/lora/test_lora_eviction.py @@ -23,8 +23,8 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.runners import SRTRunner from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=224, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=224, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=224, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=224, suite="stage-b-test-1-gpu-small-amd") PROMPTS = [ "AI is a field of computer science focused on", diff --git a/test/registered/lora/test_lora_hf_sgl_logprob_diff.py b/test/registered/lora/test_lora_hf_sgl_logprob_diff.py index 73caf1508..c32c100a5 100644 --- a/test/registered/lora/test_lora_hf_sgl_logprob_diff.py +++ b/test/registered/lora/test_lora_hf_sgl_logprob_diff.py @@ -41,11 +41,11 @@ from sglang.test.test_utils import DEFAULT_PORT_FOR_SRT_TEST_RUNNER, CustomTestC register_cuda_ci( est_time=150, - suite="stage-b-test-small-1-gpu", + suite="stage-b-test-1-gpu-small", ) register_amd_ci( est_time=250, - suite="stage-b-test-small-1-gpu-amd", + suite="stage-b-test-1-gpu-small-amd", ) # Test configuration constants BASE_MODEL = "meta-llama/Llama-2-7b-hf" diff --git a/test/registered/lora/test_lora_overlap_loading.py b/test/registered/lora/test_lora_overlap_loading.py index 49658c83d..a35b5422b 100644 --- a/test/registered/lora/test_lora_overlap_loading.py +++ b/test/registered/lora/test_lora_overlap_loading.py @@ -29,8 +29,8 @@ from sglang.test.lora_utils import ( ) from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=75, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=75, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=75, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=75, suite="stage-b-test-1-gpu-small-amd") class TestLoRAOverlapLoading(CustomTestCase): diff --git a/test/registered/lora/test_lora_qwen3.py b/test/registered/lora/test_lora_qwen3.py index 199075f52..115597fcf 100644 --- a/test/registered/lora/test_lora_qwen3.py +++ b/test/registered/lora/test_lora_qwen3.py @@ -24,7 +24,7 @@ from sglang.test.test_utils import CustomTestCase register_amd_ci( est_time=30, - suite="stage-b-test-small-1-gpu-amd", + suite="stage-b-test-1-gpu-small-amd", disabled="see https://github.com/sgl-project/sglang/issues/13107", ) register_cuda_ci(est_time=97, suite="nightly-1-gpu", nightly=True) diff --git a/test/registered/lora/test_lora_tp.py b/test/registered/lora/test_lora_tp.py index 43a2074e3..017b4da53 100644 --- a/test/registered/lora/test_lora_tp.py +++ b/test/registered/lora/test_lora_tp.py @@ -31,11 +31,11 @@ from sglang.test.test_utils import CustomTestCase, is_in_ci register_cuda_ci( est_time=116, - suite="stage-b-test-large-2-gpu", + suite="stage-b-test-2-gpu-large", ) register_amd_ci( est_time=116, - suite="stage-b-test-large-2-gpu-amd", + suite="stage-b-test-2-gpu-large-amd", disabled="see https://github.com/sgl-project/sglang/issues/13107", ) diff --git a/test/registered/lora/test_lora_update.py b/test/registered/lora/test_lora_update.py index 218a08480..860520dba 100644 --- a/test/registered/lora/test_lora_update.py +++ b/test/registered/lora/test_lora_update.py @@ -36,7 +36,7 @@ from sglang.test.test_utils import ( register_cuda_ci( est_time=487, - suite="stage-b-test-large-1-gpu", + suite="stage-b-test-1-gpu-large", ) PROMPTS = [ diff --git a/test/registered/lora/test_multi_lora_backend.py b/test/registered/lora/test_multi_lora_backend.py index f34b9e622..a0b5db62b 100644 --- a/test/registered/lora/test_multi_lora_backend.py +++ b/test/registered/lora/test_multi_lora_backend.py @@ -25,8 +25,8 @@ from sglang.test.lora_utils import ( ) from sglang.test.test_utils import CustomTestCase, is_in_ci -register_cuda_ci(est_time=100, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=100, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=100, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=100, suite="stage-b-test-1-gpu-small-amd") class TestMultiLoRABackend(CustomTestCase): diff --git a/test/registered/metrics/test_metrics.py b/test/registered/metrics/test_metrics.py index 409ebdbcb..4aafbabae 100644 --- a/test/registered/metrics/test_metrics.py +++ b/test/registered/metrics/test_metrics.py @@ -20,8 +20,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=32, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=32, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=32, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=32, suite="stage-b-test-1-gpu-small-amd") _MODEL_NAME = "Qwen/Qwen3-0.6B" diff --git a/test/registered/metrics/test_priority_metrics.py b/test/registered/metrics/test_priority_metrics.py index b25caea41..70c561312 100644 --- a/test/registered/metrics/test_priority_metrics.py +++ b/test/registered/metrics/test_priority_metrics.py @@ -18,9 +18,9 @@ from sglang.test.test_utils import ( register_cuda_ci( est_time=60, - suite="stage-b-test-small-1-gpu", + suite="stage-b-test-1-gpu-small", ) -register_amd_ci(est_time=60, suite="stage-b-test-small-1-gpu-amd") +register_amd_ci(est_time=60, suite="stage-b-test-1-gpu-small-amd") _MODEL_NAME = "Qwen/Qwen3-0.6B" diff --git a/test/registered/mla/test_flashmla.py b/test/registered/mla/test_flashmla.py index 4b3ab577f..b270358c8 100644 --- a/test/registered/mla/test_flashmla.py +++ b/test/registered/mla/test_flashmla.py @@ -21,7 +21,7 @@ from sglang.test.test_utils import ( ) # FlashMLA attention backend tests with MTP speculative decoding -register_cuda_ci(est_time=284, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=284, suite="stage-b-test-1-gpu-large") class TestFlashMLAAttnBackend(unittest.TestCase): diff --git a/test/registered/mla/test_mla.py b/test/registered/mla/test_mla.py index 3cc401d66..55140c2ae 100644 --- a/test/registered/mla/test_mla.py +++ b/test/registered/mla/test_mla.py @@ -13,8 +13,8 @@ from sglang.test.test_utils import ( ) # MLA attention test with MGSM evaluation -register_cuda_ci(est_time=194, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=1100, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=194, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=1100, suite="stage-b-test-1-gpu-small-amd") class TestMLA(CustomTestCase): diff --git a/test/registered/mla/test_mla_deepseek_v3.py b/test/registered/mla/test_mla_deepseek_v3.py index 85236be24..392154e3d 100644 --- a/test/registered/mla/test_mla_deepseek_v3.py +++ b/test/registered/mla/test_mla_deepseek_v3.py @@ -16,10 +16,10 @@ from sglang.test.test_utils import ( ) # DeepSeek-V3 MLA tests with torch compile, FA3, and MTP speculative decoding -register_cuda_ci(est_time=442, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=442, suite="stage-b-test-1-gpu-large") register_amd_ci( est_time=221, - suite="stage-b-test-small-1-gpu-amd", + suite="stage-b-test-1-gpu-small-amd", disabled="see https://github.com/sgl-project/sglang/issues/12574", ) diff --git a/test/registered/mla/test_mla_flashinfer.py b/test/registered/mla/test_mla_flashinfer.py index f8f2e02cd..555a54e5e 100644 --- a/test/registered/mla/test_mla_flashinfer.py +++ b/test/registered/mla/test_mla_flashinfer.py @@ -15,7 +15,7 @@ from sglang.test.test_utils import ( ) # FlashInfer MLA backend tests with MTP speculative decoding -register_cuda_ci(est_time=302, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=302, suite="stage-b-test-1-gpu-large") class TestFlashinferMLA(CustomTestCase): diff --git a/test/registered/mla/test_mla_fp8.py b/test/registered/mla/test_mla_fp8.py index 0135c7edb..80528a545 100644 --- a/test/registered/mla/test_mla_fp8.py +++ b/test/registered/mla/test_mla_fp8.py @@ -13,8 +13,8 @@ from sglang.test.test_utils import ( ) # MLA FP8 KV cache test with MGSM evaluation -register_cuda_ci(est_time=77, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=800, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=77, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=800, suite="stage-b-test-1-gpu-small-amd") class TestMLA(CustomTestCase): diff --git a/test/registered/mla/test_mla_int8_deepseek_v3.py b/test/registered/mla/test_mla_int8_deepseek_v3.py index a2e14bc38..d7acb8040 100644 --- a/test/registered/mla/test_mla_int8_deepseek_v3.py +++ b/test/registered/mla/test_mla_int8_deepseek_v3.py @@ -16,7 +16,7 @@ from sglang.test.test_utils import ( ) # DeepSeek-V3 INT8 quantization tests (channel and block INT8) -register_cuda_ci(est_time=341, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=341, suite="stage-b-test-1-gpu-large") class TestMLADeepseekV3ChannelInt8(CustomTestCase): diff --git a/test/registered/model_loading/test_external_models.py b/test/registered/model_loading/test_external_models.py index a191e8f62..c4878abe6 100644 --- a/test/registered/model_loading/test_external_models.py +++ b/test/registered/model_loading/test_external_models.py @@ -5,8 +5,8 @@ from sglang.srt.environ import envs from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=30, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=45, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=30, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=45, suite="stage-b-test-1-gpu-small-amd") class TestExternalModels(CustomTestCase): diff --git a/test/registered/model_loading/test_utils_update_weights.py b/test/registered/model_loading/test_utils_update_weights.py index fcc89ac77..60f3800b0 100644 --- a/test/registered/model_loading/test_utils_update_weights.py +++ b/test/registered/model_loading/test_utils_update_weights.py @@ -12,7 +12,7 @@ from sglang.srt.weight_sync.utils import update_weights from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.test_utils import DEFAULT_SMALL_MODEL_NAME_FOR_TEST -register_cuda_ci(est_time=29, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=29, suite="stage-b-test-1-gpu-large") class AsyncEngine(Engine): diff --git a/test/registered/models/test_compressed_tensors_models.py b/test/registered/models/test_compressed_tensors_models.py index c733530ab..77e751669 100644 --- a/test/registered/models/test_compressed_tensors_models.py +++ b/test/registered/models/test_compressed_tensors_models.py @@ -13,8 +13,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=42, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=42, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=42, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=42, suite="stage-b-test-1-gpu-small-amd") class TestCompressedTensorsLlama3FP8(CustomTestCase): diff --git a/test/registered/models/test_cross_encoder_models.py b/test/registered/models/test_cross_encoder_models.py index b1dbbefce..453d67a60 100644 --- a/test/registered/models/test_cross_encoder_models.py +++ b/test/registered/models/test_cross_encoder_models.py @@ -11,8 +11,8 @@ from sglang.test.test_utils import CustomTestCase, is_in_ci # Cross encoder model tests -register_cuda_ci(est_time=100, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=150, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=100, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=150, suite="stage-b-test-1-gpu-small-amd") MODELS = [ ("cross-encoder/ms-marco-MiniLM-L6-v2", 1, 1e-2), diff --git a/test/registered/models/test_dummy_grok_models.py b/test/registered/models/test_dummy_grok_models.py index e73f55467..f8ae27cdd 100644 --- a/test/registered/models/test_dummy_grok_models.py +++ b/test/registered/models/test_dummy_grok_models.py @@ -5,7 +5,7 @@ from sglang.test.test_utils import CustomTestCase, is_in_ci, run_bench_one_batch register_cuda_ci( est_time=120, - suite="stage-b-test-large-2-gpu", + suite="stage-b-test-2-gpu-large", disabled="Temporarily disabled", ) diff --git a/test/registered/models/test_generation_models.py b/test/registered/models/test_generation_models.py index 2606dbb5e..98e2bac4b 100644 --- a/test/registered/models/test_generation_models.py +++ b/test/registered/models/test_generation_models.py @@ -1,8 +1,8 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci # Generation model tests (CUDA only) -register_cuda_ci(est_time=103, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=106, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=103, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=106, suite="stage-b-test-1-gpu-small-amd") # Copyright 2023-2024 SGLang Team # Licensed under the Apache License, Version 2.0 (the "License"); diff --git a/test/registered/models/test_gpt_oss_models_pcg.py b/test/registered/models/test_gpt_oss_models_pcg.py index 9bbd621a4..438f127c2 100644 --- a/test/registered/models/test_gpt_oss_models_pcg.py +++ b/test/registered/models/test_gpt_oss_models_pcg.py @@ -17,7 +17,7 @@ from sglang.test.test_utils import ( register_cuda_ci( est_time=400, - suite="stage-b-test-large-2-gpu", + suite="stage-b-test-2-gpu-large", ) GPT_OSS_MODEL = "openai/gpt-oss-120b" diff --git a/test/registered/models/test_kimi_linear_models.py b/test/registered/models/test_kimi_linear_models.py index 88ef6d7ad..c63f6ac32 100644 --- a/test/registered/models/test_kimi_linear_models.py +++ b/test/registered/models/test_kimi_linear_models.py @@ -11,7 +11,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=90, suite="stage-b-test-large-2-gpu") +register_cuda_ci(est_time=90, suite="stage-b-test-2-gpu-large") class TestKimiLinear(CustomTestCase): diff --git a/test/registered/models/test_kimi_linear_models_pcg.py b/test/registered/models/test_kimi_linear_models_pcg.py index c464831dd..32314f2b8 100644 --- a/test/registered/models/test_kimi_linear_models_pcg.py +++ b/test/registered/models/test_kimi_linear_models_pcg.py @@ -17,7 +17,7 @@ from sglang.test.test_utils import ( register_cuda_ci( est_time=100, - suite="stage-b-test-large-2-gpu", + suite="stage-b-test-2-gpu-large", ) KIMI_LINEAR_MODEL = "moonshotai/Kimi-Linear-48B-A3B-Instruct" diff --git a/test/registered/models/test_ministral3_models.py b/test/registered/models/test_ministral3_models.py index 0da56fb81..49184a05d 100644 --- a/test/registered/models/test_ministral3_models.py +++ b/test/registered/models/test_ministral3_models.py @@ -8,7 +8,7 @@ from sglang.test.server_fixtures.mmmu_fixture import MMMUServerBase register_cuda_ci( est_time=200, - suite="stage-b-test-small-1-gpu", + suite="stage-b-test-1-gpu-small", disabled="Temporarily disabled", ) diff --git a/test/registered/models/test_nvidia_nemotron_3_nano.py b/test/registered/models/test_nvidia_nemotron_3_nano.py index 2129abca3..0580979c4 100644 --- a/test/registered/models/test_nvidia_nemotron_3_nano.py +++ b/test/registered/models/test_nvidia_nemotron_3_nano.py @@ -4,7 +4,7 @@ from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.kits.lm_eval_kit import LMEvalMixin from sglang.test.server_fixtures.default_fixture import DefaultServerBase -register_cuda_ci(est_time=180, suite="stage-b-test-large-2-gpu") +register_cuda_ci(est_time=180, suite="stage-b-test-2-gpu-large") NEMOTRON_3_NANO_THINKING_ARGS = [ "--trust-remote-code", diff --git a/test/registered/models/test_nvidia_nemotron_nano_v2.py b/test/registered/models/test_nvidia_nemotron_nano_v2.py index 21a5a9aa8..7e510bcd9 100644 --- a/test/registered/models/test_nvidia_nemotron_nano_v2.py +++ b/test/registered/models/test_nvidia_nemotron_nano_v2.py @@ -5,7 +5,7 @@ from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.kits.gsm8k_accuracy_kit import GSM8KMixin from sglang.test.server_fixtures.default_fixture import DefaultServerBase -register_cuda_ci(est_time=132, suite="stage-b-test-large-2-gpu") +register_cuda_ci(est_time=132, suite="stage-b-test-2-gpu-large") class TestNvidiaNemotronNanoV2BF16(GSM8KMixin, DefaultServerBase): diff --git a/test/registered/models/test_nvidia_nemotron_nano_v2_vl.py b/test/registered/models/test_nvidia_nemotron_nano_v2_vl.py index 6257b5fed..abf6d91d0 100644 --- a/test/registered/models/test_nvidia_nemotron_nano_v2_vl.py +++ b/test/registered/models/test_nvidia_nemotron_nano_v2_vl.py @@ -10,7 +10,7 @@ from sglang.test.server_fixtures.mmmu_fixture import MMMUServerBase # GSM8k + MMMU evaluation -register_cuda_ci(est_time=214, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=214, suite="stage-b-test-1-gpu-large") MODEL = "nvidia/NVIDIA-Nemotron-Nano-12B-v2-VL-BF16" diff --git a/test/registered/models/test_qwen_models.py b/test/registered/models/test_qwen_models.py index e0dd5c503..45b2aeb54 100644 --- a/test/registered/models/test_qwen_models.py +++ b/test/registered/models/test_qwen_models.py @@ -13,8 +13,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=90, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=130, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=90, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=130, suite="stage-b-test-1-gpu-small-amd") class TestQwen2(CustomTestCase): diff --git a/test/registered/models/test_reward_models.py b/test/registered/models/test_reward_models.py index 0105b0e5e..a26314e62 100644 --- a/test/registered/models/test_reward_models.py +++ b/test/registered/models/test_reward_models.py @@ -24,8 +24,8 @@ from sglang.test.test_utils import CustomTestCase # ============================================================================== -register_cuda_ci(est_time=103, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=132, suite="stage-b-test-small-1-gpu-amd-nondeterministic") +register_cuda_ci(est_time=103, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=132, suite="stage-b-test-1-gpu-small-amd-nondeterministic") MODELS = [ ("LxzGordon/URM-LLaMa-3.1-8B", 1, 4e-2), diff --git a/test/registered/models/test_transformers_models.py b/test/registered/models/test_transformers_models.py index 471445ade..325db6e29 100644 --- a/test/registered/models/test_transformers_models.py +++ b/test/registered/models/test_transformers_models.py @@ -20,8 +20,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=245, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=320, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=245, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=320, suite="stage-b-test-1-gpu-small-amd") class TestTransformersFallbackEndpoint(CustomTestCase): diff --git a/test/registered/models/test_vlm_models.py b/test/registered/models/test_vlm_models.py index a2c6cabc9..e03c56e8a 100644 --- a/test/registered/models/test_vlm_models.py +++ b/test/registered/models/test_vlm_models.py @@ -16,8 +16,8 @@ from sglang.test.test_utils import is_in_ci # VLM (Vision Language Model) tests -register_cuda_ci(est_time=228, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=850, suite="stage-b-test-small-1-gpu-amd-nondeterministic") +register_cuda_ci(est_time=228, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=850, suite="stage-b-test-1-gpu-small-amd-nondeterministic") _is_hip = is_hip() # VLM models for testing diff --git a/test/registered/moe/test_fused_moe.py b/test/registered/moe/test_fused_moe.py index a5f6d9234..d67f90356 100644 --- a/test/registered/moe/test_fused_moe.py +++ b/test/registered/moe/test_fused_moe.py @@ -13,8 +13,8 @@ from sglang.srt.utils import get_device, get_device_capability, is_hip from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.test_utils import CustomTestCase, empty_gpu_cache -register_cuda_ci(est_time=80, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=30, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=80, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=30, suite="stage-b-test-1-gpu-small-amd") _is_hip = is_hip() _is_fp8_fnuz = is_fp8_fnuz() diff --git a/test/registered/moe/test_glm4_moe_models.py b/test/registered/moe/test_glm4_moe_models.py index 4d668f313..59003f5ba 100644 --- a/test/registered/moe/test_glm4_moe_models.py +++ b/test/registered/moe/test_glm4_moe_models.py @@ -11,7 +11,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=100, suite="stage-b-test-large-2-gpu") +register_cuda_ci(est_time=100, suite="stage-b-test-2-gpu-large") class TestGLM4MoE(CustomTestCase): diff --git a/test/registered/moe/test_moe_ep.py b/test/registered/moe/test_moe_ep.py index b11055c4b..155ce767c 100644 --- a/test/registered/moe/test_moe_ep.py +++ b/test/registered/moe/test_moe_ep.py @@ -12,7 +12,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=140, suite="stage-b-test-large-2-gpu") +register_cuda_ci(est_time=140, suite="stage-b-test-2-gpu-large") class TestEp(CustomTestCase): diff --git a/test/registered/moe/test_torch_compile_moe.py b/test/registered/moe/test_torch_compile_moe.py index 707967ff0..e62af0cb9 100644 --- a/test/registered/moe/test_torch_compile_moe.py +++ b/test/registered/moe/test_torch_compile_moe.py @@ -15,8 +15,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=210, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=1400, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=210, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=1400, suite="stage-b-test-1-gpu-small-amd") class TestTorchCompileMoe(CustomTestCase): diff --git a/test/registered/moe/test_triton_fused_moe.py b/test/registered/moe/test_triton_fused_moe.py index 11b62528e..e20ed609f 100644 --- a/test/registered/moe/test_triton_fused_moe.py +++ b/test/registered/moe/test_triton_fused_moe.py @@ -12,7 +12,7 @@ from sglang.srt.server_args import ServerArgs, set_global_server_args_for_schedu from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=89, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=89, suite="stage-b-test-1-gpu-large") class TestFusedMOE(CustomTestCase): diff --git a/test/registered/moe/test_triton_moe_channel_fp8_kernel.py b/test/registered/moe/test_triton_moe_channel_fp8_kernel.py index b180864ea..4acbc9ab2 100644 --- a/test/registered/moe/test_triton_moe_channel_fp8_kernel.py +++ b/test/registered/moe/test_triton_moe_channel_fp8_kernel.py @@ -11,7 +11,7 @@ from sglang.srt.server_args import ServerArgs, set_global_server_args_for_schedu from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=16, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=16, suite="stage-b-test-1-gpu-large") def native_w8a8_per_token_matmul(A, B, As, Bs, output_dtype=torch.float16): diff --git a/test/registered/openai_server/basic/test_anthropic_server.py b/test/registered/openai_server/basic/test_anthropic_server.py index fbf5b0463..2ea535f84 100644 --- a/test/registered/openai_server/basic/test_anthropic_server.py +++ b/test/registered/openai_server/basic/test_anthropic_server.py @@ -31,8 +31,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=120, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=140, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=120, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=140, suite="stage-b-test-1-gpu-small-amd") class TestAnthropicServer(CustomTestCase): diff --git a/test/registered/openai_server/basic/test_openai_server.py b/test/registered/openai_server/basic/test_openai_server.py index cdc39f29a..ce1b8ec51 100644 --- a/test/registered/openai_server/basic/test_openai_server.py +++ b/test/registered/openai_server/basic/test_openai_server.py @@ -28,8 +28,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=184, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=200, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=184, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=200, suite="stage-b-test-1-gpu-small-amd") class TestOpenAIServer(CustomTestCase): diff --git a/test/registered/openai_server/basic/test_protocol.py b/test/registered/openai_server/basic/test_protocol.py index b5a1758ff..bdbdff6a1 100644 --- a/test/registered/openai_server/basic/test_protocol.py +++ b/test/registered/openai_server/basic/test_protocol.py @@ -30,8 +30,8 @@ from sglang.srt.entrypoints.openai.protocol import ( ) from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=3, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=3, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=10, suite="stage-b-test-1-gpu-small-amd") class TestModelCard(unittest.TestCase): diff --git a/test/registered/openai_server/basic/test_serving_chat.py b/test/registered/openai_server/basic/test_serving_chat.py index 27dfa7ea0..2f63bbf5f 100644 --- a/test/registered/openai_server/basic/test_serving_chat.py +++ b/test/registered/openai_server/basic/test_serving_chat.py @@ -24,8 +24,8 @@ from sglang.srt.managers.io_struct import GenerateReqInput from sglang.srt.utils import get_or_create_event_loop from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=10, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=10, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=10, suite="stage-b-test-1-gpu-small-amd") class _MockTokenizerManager: diff --git a/test/registered/openai_server/basic/test_serving_completions.py b/test/registered/openai_server/basic/test_serving_completions.py index ecbad79d8..026f1c2d8 100644 --- a/test/registered/openai_server/basic/test_serving_completions.py +++ b/test/registered/openai_server/basic/test_serving_completions.py @@ -18,8 +18,8 @@ from sglang.srt.managers.tokenizer_manager import TokenizerManager from sglang.srt.utils import get_or_create_event_loop from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=10, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=10, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=10, suite="stage-b-test-1-gpu-small-amd") class _MockTemplateManager: diff --git a/test/registered/openai_server/basic/test_serving_rerank.py b/test/registered/openai_server/basic/test_serving_rerank.py index 4ed4599eb..40fd4a666 100644 --- a/test/registered/openai_server/basic/test_serving_rerank.py +++ b/test/registered/openai_server/basic/test_serving_rerank.py @@ -7,8 +7,8 @@ from sglang.srt.managers.tokenizer_manager_multiitem_mixin import ScoreResult from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci # Keep consistent with other openai_server/basic unit tests. -register_cuda_ci(est_time=10, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=10, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=10, suite="stage-b-test-1-gpu-small-amd") try: from sglang.srt.entrypoints.openai.serving_rerank import ( diff --git a/test/registered/openai_server/features/test_enable_thinking.py b/test/registered/openai_server/features/test_enable_thinking.py index 06cf7f4c7..6ee28866a 100644 --- a/test/registered/openai_server/features/test_enable_thinking.py +++ b/test/registered/openai_server/features/test_enable_thinking.py @@ -21,8 +21,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=103, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=200, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=103, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=200, suite="stage-b-test-1-gpu-small-amd") class TestEnableThinking(CustomTestCase): diff --git a/test/registered/openai_server/features/test_json_mode.py b/test/registered/openai_server/features/test_json_mode.py index 7f8373458..8cb922ecb 100644 --- a/test/registered/openai_server/features/test_json_mode.py +++ b/test/registered/openai_server/features/test_json_mode.py @@ -14,8 +14,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=109, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=180, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=109, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=180, suite="stage-b-test-1-gpu-small-amd") class JSONModeMixin: diff --git a/test/registered/openai_server/features/test_openai_server_ebnf.py b/test/registered/openai_server/features/test_openai_server_ebnf.py index 165bbcf78..05cf9f102 100644 --- a/test/registered/openai_server/features/test_openai_server_ebnf.py +++ b/test/registered/openai_server/features/test_openai_server_ebnf.py @@ -13,8 +13,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=7, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=20, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=7, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=20, suite="stage-b-test-1-gpu-small-amd") # ------------------------------------------------------------------------- diff --git a/test/registered/openai_server/features/test_openai_server_hidden_states.py b/test/registered/openai_server/features/test_openai_server_hidden_states.py index 0ac3d3265..ff1070385 100644 --- a/test/registered/openai_server/features/test_openai_server_hidden_states.py +++ b/test/registered/openai_server/features/test_openai_server_hidden_states.py @@ -16,10 +16,10 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=186, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=186, suite="stage-b-test-1-gpu-small") register_amd_ci( est_time=186, - suite="stage-b-test-small-1-gpu-amd", + suite="stage-b-test-1-gpu-small-amd", disabled="see https://github.com/sgl-project/sglang/issues/11127", ) diff --git a/test/registered/openai_server/features/test_reasoning_content.py b/test/registered/openai_server/features/test_reasoning_content.py index 2c59921fe..d53c41f44 100644 --- a/test/registered/openai_server/features/test_reasoning_content.py +++ b/test/registered/openai_server/features/test_reasoning_content.py @@ -23,8 +23,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=89, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=89, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=89, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=89, suite="stage-b-test-1-gpu-small-amd") class TestReasoningContentAPI(CustomTestCase): diff --git a/test/registered/openai_server/function_call/test_anthropic_tool_use.py b/test/registered/openai_server/function_call/test_anthropic_tool_use.py index 086f1b036..1653b216d 100644 --- a/test/registered/openai_server/function_call/test_anthropic_tool_use.py +++ b/test/registered/openai_server/function_call/test_anthropic_tool_use.py @@ -25,8 +25,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=120, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=140, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=120, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=140, suite="stage-b-test-1-gpu-small-amd") # System message to guide Llama3.2 to produce proper tool call format SYSTEM_MESSAGE = ( diff --git a/test/registered/openai_server/function_call/test_openai_function_calling.py b/test/registered/openai_server/function_call/test_openai_function_calling.py index 7c6a7bb31..33cde4018 100644 --- a/test/registered/openai_server/function_call/test_openai_function_calling.py +++ b/test/registered/openai_server/function_call/test_openai_function_calling.py @@ -14,8 +14,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=60, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=73, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=60, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=73, suite="stage-b-test-1-gpu-small-amd") class TestOpenAIServerFunctionCalling(CustomTestCase): diff --git a/test/registered/openai_server/function_call/test_tool_choice.py b/test/registered/openai_server/function_call/test_tool_choice.py index d463e8bbf..9b0fad76d 100644 --- a/test/registered/openai_server/function_call/test_tool_choice.py +++ b/test/registered/openai_server/function_call/test_tool_choice.py @@ -22,8 +22,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=120, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=258, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=120, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=258, suite="stage-b-test-1-gpu-small-amd") class TestToolChoiceLlama32(CustomTestCase): diff --git a/test/registered/openai_server/validation/test_large_max_new_tokens.py b/test/registered/openai_server/validation/test_large_max_new_tokens.py index 047c8d93f..981e3af72 100644 --- a/test/registered/openai_server/validation/test_large_max_new_tokens.py +++ b/test/registered/openai_server/validation/test_large_max_new_tokens.py @@ -22,8 +22,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=41, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=41, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=41, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=41, suite="stage-b-test-1-gpu-small-amd") class TestLargeMaxNewTokens(CustomTestCase): diff --git a/test/registered/openai_server/validation/test_matched_stop.py b/test/registered/openai_server/validation/test_matched_stop.py index aa890e363..0611a1ea2 100644 --- a/test/registered/openai_server/validation/test_matched_stop.py +++ b/test/registered/openai_server/validation/test_matched_stop.py @@ -11,8 +11,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=40, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=60, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=40, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=60, suite="stage-b-test-1-gpu-small-amd") class TestMatchedStop(CustomTestCase, MatchedStopMixin): diff --git a/test/registered/openai_server/validation/test_openai_server_ignore_eos.py b/test/registered/openai_server/validation/test_openai_server_ignore_eos.py index bbf696d41..9014466c0 100644 --- a/test/registered/openai_server/validation/test_openai_server_ignore_eos.py +++ b/test/registered/openai_server/validation/test_openai_server_ignore_eos.py @@ -11,8 +11,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=6, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=47, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=6, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=47, suite="stage-b-test-1-gpu-small-amd") class TestOpenAIServerIgnoreEOS(CustomTestCase): diff --git a/test/registered/openai_server/validation/test_request_length_validation.py b/test/registered/openai_server/validation/test_request_length_validation.py index 5035b6d4d..0e9fb0a3e 100644 --- a/test/registered/openai_server/validation/test_request_length_validation.py +++ b/test/registered/openai_server/validation/test_request_length_validation.py @@ -12,8 +12,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=38, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=31, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=38, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=31, suite="stage-b-test-1-gpu-small-amd") class TestRequestLengthValidation(CustomTestCase): diff --git a/test/registered/ops/test_repeat_interleave.py b/test/registered/ops/test_repeat_interleave.py index 0254cbbfa..acb45bd47 100644 --- a/test/registered/ops/test_repeat_interleave.py +++ b/test/registered/ops/test_repeat_interleave.py @@ -11,8 +11,8 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci # Ops - Repeat Interleave tests (1-GPU) -register_cuda_ci(est_time=8, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=75, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=8, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=75, suite="stage-b-test-1-gpu-small-amd") def torch_ref_impl(grid_thw: torch.Tensor) -> torch.Tensor: diff --git a/test/registered/perf/test_bench_one_batch_1gpu.py b/test/registered/perf/test_bench_one_batch_1gpu.py index 5ab4b202c..acc8ccafb 100644 --- a/test/registered/perf/test_bench_one_batch_1gpu.py +++ b/test/registered/perf/test_bench_one_batch_1gpu.py @@ -11,8 +11,8 @@ from sglang.test.test_utils import ( write_github_step_summary, ) -register_cuda_ci(est_time=120, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=120, suite="stage-b-test-large-1-gpu-amd") +register_cuda_ci(est_time=120, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=120, suite="stage-b-test-1-gpu-large-amd") class TestBenchOneBatch1GPU(CustomTestCase): diff --git a/test/registered/perf/test_bench_one_batch_2gpu.py b/test/registered/perf/test_bench_one_batch_2gpu.py index b36e775db..38fe4b3c8 100644 --- a/test/registered/perf/test_bench_one_batch_2gpu.py +++ b/test/registered/perf/test_bench_one_batch_2gpu.py @@ -11,8 +11,8 @@ from sglang.test.test_utils import ( write_github_step_summary, ) -register_cuda_ci(est_time=180, suite="stage-b-test-large-2-gpu") -register_amd_ci(est_time=630, suite="stage-b-test-large-2-gpu-amd") +register_cuda_ci(est_time=180, suite="stage-b-test-2-gpu-large") +register_amd_ci(est_time=630, suite="stage-b-test-2-gpu-large-amd") class TestBenchOneBatch2GPU(CustomTestCase): diff --git a/test/registered/perf/test_bench_serving_1gpu_large.py b/test/registered/perf/test_bench_serving_1gpu_large.py index 6dd8c4249..36f17107d 100644 --- a/test/registered/perf/test_bench_serving_1gpu_large.py +++ b/test/registered/perf/test_bench_serving_1gpu_large.py @@ -17,8 +17,8 @@ from sglang.test.test_utils import ( write_github_step_summary, ) -register_cuda_ci(est_time=300, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=300, suite="stage-b-test-large-1-gpu-amd") +register_cuda_ci(est_time=300, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=300, suite="stage-b-test-1-gpu-large-amd") class TestBenchServing1GPULarge(CustomTestCase): diff --git a/test/registered/perf/test_bench_serving_1gpu_part1.py b/test/registered/perf/test_bench_serving_1gpu_part1.py index 385647efc..f97c6e706 100644 --- a/test/registered/perf/test_bench_serving_1gpu_part1.py +++ b/test/registered/perf/test_bench_serving_1gpu_part1.py @@ -19,8 +19,8 @@ from sglang.test.test_utils import ( write_github_step_summary, ) -register_cuda_ci(est_time=1000, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=1100, suite="stage-b-test-large-1-gpu-amd") +register_cuda_ci(est_time=1000, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=1100, suite="stage-b-test-1-gpu-large-amd") class TestBenchServing1GPUPart1(CustomTestCase): diff --git a/test/registered/perf/test_bench_serving_1gpu_part2.py b/test/registered/perf/test_bench_serving_1gpu_part2.py index 6730e2e67..5bb97ef49 100644 --- a/test/registered/perf/test_bench_serving_1gpu_part2.py +++ b/test/registered/perf/test_bench_serving_1gpu_part2.py @@ -19,8 +19,8 @@ from sglang.test.test_utils import ( write_github_step_summary, ) -register_cuda_ci(est_time=900, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=900, suite="stage-b-test-large-1-gpu-amd") +register_cuda_ci(est_time=900, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=900, suite="stage-b-test-1-gpu-large-amd") class TestBenchServing1GPUPart2(CustomTestCase): diff --git a/test/registered/perf/test_bench_serving_2gpu.py b/test/registered/perf/test_bench_serving_2gpu.py index 00e8a058b..78320decf 100644 --- a/test/registered/perf/test_bench_serving_2gpu.py +++ b/test/registered/perf/test_bench_serving_2gpu.py @@ -14,8 +14,8 @@ from sglang.test.test_utils import ( write_github_step_summary, ) -register_cuda_ci(est_time=600, suite="stage-b-test-large-2-gpu") -register_amd_ci(est_time=1100, suite="stage-b-test-large-2-gpu-amd") +register_cuda_ci(est_time=600, suite="stage-b-test-2-gpu-large") +register_amd_ci(est_time=1100, suite="stage-b-test-2-gpu-large-amd") class TestBenchServing2GPU(CustomTestCase): diff --git a/test/registered/perf/test_vlm_perf_5090.py b/test/registered/perf/test_vlm_perf_5090.py index 389e4dc85..8f7d4cd04 100644 --- a/test/registered/perf/test_vlm_perf_5090.py +++ b/test/registered/perf/test_vlm_perf_5090.py @@ -13,8 +13,8 @@ from sglang.test.test_utils import ( write_github_step_summary, ) -register_cuda_ci(est_time=600, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=500, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=600, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=500, suite="stage-b-test-1-gpu-small-amd") class TestVLMPerf5090(CustomTestCase): diff --git a/test/registered/piecewise_cuda_graph/test_piecewise_cuda_graph_support_1_gpu.py b/test/registered/piecewise_cuda_graph/test_piecewise_cuda_graph_support_1_gpu.py index 14a4abeaa..aa1afba03 100644 --- a/test/registered/piecewise_cuda_graph/test_piecewise_cuda_graph_support_1_gpu.py +++ b/test/registered/piecewise_cuda_graph/test_piecewise_cuda_graph_support_1_gpu.py @@ -17,7 +17,7 @@ from sglang.test.test_utils import ( ) # CI Registration -register_cuda_ci(est_time=220, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=220, suite="stage-b-test-1-gpu-large") class TestPiecewiseCudaGraphQwen25VL(CustomTestCase): diff --git a/test/registered/profiling/test_profile_v2.py b/test/registered/profiling/test_profile_v2.py index c90c62292..33452027d 100644 --- a/test/registered/profiling/test_profile_v2.py +++ b/test/registered/profiling/test_profile_v2.py @@ -19,7 +19,7 @@ from sglang.test.test_utils import ( register_cuda_ci( est_time=120, - suite="stage-b-test-small-1-gpu", + suite="stage-b-test-1-gpu-small", disabled="Temporarily disabled", ) diff --git a/test/registered/profiling/test_start_profile.py b/test/registered/profiling/test_start_profile.py index ddd19b6bb..a2d25e8e9 100644 --- a/test/registered/profiling/test_start_profile.py +++ b/test/registered/profiling/test_start_profile.py @@ -29,8 +29,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=41, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=60, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=41, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=60, suite="stage-b-test-1-gpu-small-amd") OUTPUT_DIR = "./profiler_dir" diff --git a/test/registered/quant/test_autoround.py b/test/registered/quant/test_autoround.py index f2baf0617..89374ec56 100644 --- a/test/registered/quant/test_autoround.py +++ b/test/registered/quant/test_autoround.py @@ -17,7 +17,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=77, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=77, suite="stage-b-test-1-gpu-large") class TestAutoRound(CustomTestCase): diff --git a/test/registered/quant/test_awq.py b/test/registered/quant/test_awq.py index 005d7bd84..06c860a63 100644 --- a/test/registered/quant/test_awq.py +++ b/test/registered/quant/test_awq.py @@ -13,8 +13,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=700, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=200, suite="stage-b-test-large-1-gpu-amd") +register_cuda_ci(est_time=700, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=200, suite="stage-b-test-1-gpu-large-amd") class TestAWQ(CustomTestCase): diff --git a/test/registered/quant/test_awq_dequant.py b/test/registered/quant/test_awq_dequant.py index 589aad2c4..ee0e80f1f 100644 --- a/test/registered/quant/test_awq_dequant.py +++ b/test/registered/quant/test_awq_dequant.py @@ -21,7 +21,7 @@ from sglang.srt.utils import get_device from sglang.test.ci.ci_register import register_amd_ci from sglang.test.test_utils import CustomTestCase -register_amd_ci(est_time=2, suite="stage-a-test-small-1-gpu-amd") +register_amd_ci(est_time=2, suite="stage-a-test-1-gpu-small-amd") device = get_device() diff --git a/test/registered/quant/test_block_int8.py b/test/registered/quant/test_block_int8.py index 91f2cd474..68ddf2fc9 100644 --- a/test/registered/quant/test_block_int8.py +++ b/test/registered/quant/test_block_int8.py @@ -10,8 +10,8 @@ from sglang.srt.server_args import ServerArgs, set_global_server_args_for_schedu from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=44, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=22, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=44, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=22, suite="stage-b-test-1-gpu-small-amd") # For test diff --git a/test/registered/quant/test_bnb.py b/test/registered/quant/test_bnb.py index 814ec6a5e..8ce583f22 100644 --- a/test/registered/quant/test_bnb.py +++ b/test/registered/quant/test_bnb.py @@ -22,7 +22,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=5, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=5, suite="stage-b-test-1-gpu-small") VISION_MODELS = [ "unsloth/Qwen2.5-VL-7B-Instruct-bnb-4bit", diff --git a/test/registered/quant/test_eval_fp8_accuracy.py b/test/registered/quant/test_eval_fp8_accuracy.py index 22b183785..f60fcc7af 100644 --- a/test/registered/quant/test_eval_fp8_accuracy.py +++ b/test/registered/quant/test_eval_fp8_accuracy.py @@ -14,8 +14,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=250, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=600, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=250, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=600, suite="stage-b-test-1-gpu-small-amd") class TestEvalFP8Accuracy(CustomTestCase): diff --git a/test/registered/quant/test_fp8_kernel.py b/test/registered/quant/test_fp8_kernel.py index 44e9b5be0..57a636235 100644 --- a/test/registered/quant/test_fp8_kernel.py +++ b/test/registered/quant/test_fp8_kernel.py @@ -9,7 +9,7 @@ from sglang.srt.layers.quantization.fp8_kernel import ( from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=132, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=132, suite="stage-b-test-1-gpu-large") class TestFP8Base(CustomTestCase): diff --git a/test/registered/quant/test_fp8_utils.py b/test/registered/quant/test_fp8_utils.py index d5c16e42b..563a9123b 100644 --- a/test/registered/quant/test_fp8_utils.py +++ b/test/registered/quant/test_fp8_utils.py @@ -10,7 +10,7 @@ from sglang.srt.layers.quantization.fp8_utils import ( from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=9, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=9, suite="stage-b-test-1-gpu-large") class TestInverseTransformScaleUe8m0(CustomTestCase): diff --git a/test/registered/quant/test_fp8kv_triton.py b/test/registered/quant/test_fp8kv_triton.py index 7fb81f3e3..3158e8e24 100644 --- a/test/registered/quant/test_fp8kv_triton.py +++ b/test/registered/quant/test_fp8kv_triton.py @@ -12,7 +12,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=520, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=520, suite="stage-b-test-1-gpu-large") class TestFP8KVCacheTritonBackend(CustomTestCase): diff --git a/test/registered/quant/test_fused_rms_fp8_group_quant.py b/test/registered/quant/test_fused_rms_fp8_group_quant.py index c9aac1de0..02dbfbbaf 100644 --- a/test/registered/quant/test_fused_rms_fp8_group_quant.py +++ b/test/registered/quant/test_fused_rms_fp8_group_quant.py @@ -7,7 +7,7 @@ import torch.nn.functional as F from sglang.test.ci.ci_register import register_amd_ci from sglang.test.test_utils import CustomTestCase -register_amd_ci(est_time=10, suite="stage-a-test-small-1-gpu-amd") +register_amd_ci(est_time=10, suite="stage-a-test-1-gpu-small-amd") def _fp8_available() -> bool: diff --git a/test/registered/quant/test_gguf.py b/test/registered/quant/test_gguf.py index 14448bf9b..853d3dda7 100644 --- a/test/registered/quant/test_gguf.py +++ b/test/registered/quant/test_gguf.py @@ -6,7 +6,7 @@ import sglang as sgl from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=96, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=96, suite="stage-b-test-1-gpu-small") class TestGGUF(CustomTestCase): diff --git a/test/registered/quant/test_gptqmodel_dynamic.py b/test/registered/quant/test_gptqmodel_dynamic.py index 9d08cace6..6fdbf676f 100644 --- a/test/registered/quant/test_gptqmodel_dynamic.py +++ b/test/registered/quant/test_gptqmodel_dynamic.py @@ -14,7 +14,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=102, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=102, suite="stage-b-test-1-gpu-large") def check_quant_method(model_path: str, use_marlin_kernel: bool): diff --git a/test/registered/quant/test_int4fp8_moe.py b/test/registered/quant/test_int4fp8_moe.py index 48e03d440..f5a8f6dca 100644 --- a/test/registered/quant/test_int4fp8_moe.py +++ b/test/registered/quant/test_int4fp8_moe.py @@ -9,7 +9,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_amd_ci(est_time=313, suite="stage-b-test-small-1-gpu-amd") +register_amd_ci(est_time=313, suite="stage-b-test-1-gpu-small-amd") class TestMixtralAccuracy(CustomTestCase): diff --git a/test/registered/quant/test_int8_kernel.py b/test/registered/quant/test_int8_kernel.py index d45f82efb..c15de1d4a 100644 --- a/test/registered/quant/test_int8_kernel.py +++ b/test/registered/quant/test_int8_kernel.py @@ -11,7 +11,7 @@ from sglang.srt.server_args import ServerArgs, set_global_server_args_for_schedu from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=8, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=8, suite="stage-b-test-1-gpu-small") def native_w8a8_per_token_matmul(A, B, As, Bs, output_dtype=torch.float16): diff --git a/test/registered/quant/test_marlin_moe.py b/test/registered/quant/test_marlin_moe.py index a37cf9fd1..310e56196 100644 --- a/test/registered/quant/test_marlin_moe.py +++ b/test/registered/quant/test_marlin_moe.py @@ -12,7 +12,7 @@ from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.test_marlin_utils import awq_marlin_quantize, marlin_quantize from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=200, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=200, suite="stage-b-test-1-gpu-small") set_global_server_args_for_scheduler(object.__new__(ServerArgs)) diff --git a/test/registered/quant/test_modelopt_fp8.py b/test/registered/quant/test_modelopt_fp8.py index 80ba8b7c5..b13adbe26 100644 --- a/test/registered/quant/test_modelopt_fp8.py +++ b/test/registered/quant/test_modelopt_fp8.py @@ -12,7 +12,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=120, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=120, suite="stage-b-test-1-gpu-large") class TestModeloptFP8(CustomTestCase): diff --git a/test/registered/quant/test_quant_config_parsing.py b/test/registered/quant/test_quant_config_parsing.py index f863a13c7..f2aeba712 100644 --- a/test/registered/quant/test_quant_config_parsing.py +++ b/test/registered/quant/test_quant_config_parsing.py @@ -5,7 +5,7 @@ from sglang.srt.configs.model_config import ModelConfig from sglang.test.ci.ci_register import register_cpu_ci from sglang.test.test_utils import CustomTestCase -register_cpu_ci(est_time=5, suite="stage-a-cpu-only") +register_cpu_ci(est_time=5, suite="stage-a-test-cpu") class TestQuantLogString(CustomTestCase): diff --git a/test/registered/quant/test_quantization.py b/test/registered/quant/test_quantization.py index 770b3855a..4ab483144 100644 --- a/test/registered/quant/test_quantization.py +++ b/test/registered/quant/test_quantization.py @@ -16,7 +16,7 @@ from sglang.test.test_utils import ( write_results_to_json, ) -register_cuda_ci(est_time=185, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=185, suite="stage-b-test-1-gpu-large") MODEL_SCORE_THRESHOLDS = { "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4": 0.825, diff --git a/test/registered/quant/test_torchao.py b/test/registered/quant/test_torchao.py index f6e70ee06..21b9d30a3 100644 --- a/test/registered/quant/test_torchao.py +++ b/test/registered/quant/test_torchao.py @@ -6,8 +6,8 @@ import requests from sglang import Engine from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=103, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=230, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=103, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=230, suite="stage-b-test-1-gpu-small-amd") from sglang.lang.chat_template import get_chat_template_by_model_path from sglang.srt.utils import kill_process_tree from sglang.test.run_eval import run_eval diff --git a/test/registered/quant/test_triton_scaled_mm.py b/test/registered/quant/test_triton_scaled_mm.py index 35a30d710..2a4010240 100644 --- a/test/registered/quant/test_triton_scaled_mm.py +++ b/test/registered/quant/test_triton_scaled_mm.py @@ -8,8 +8,8 @@ from sglang.srt.layers.quantization.fp8_kernel import triton_scaled_mm from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=8, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=12, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=8, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=12, suite="stage-b-test-1-gpu-small-amd") def torch_scaled_mm( diff --git a/test/registered/quant/test_w8a8_quantization.py b/test/registered/quant/test_w8a8_quantization.py index df9124cd8..88e344831 100644 --- a/test/registered/quant/test_w8a8_quantization.py +++ b/test/registered/quant/test_w8a8_quantization.py @@ -14,7 +14,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=160, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=160, suite="stage-b-test-1-gpu-large") class BaseW8A8Test(CustomTestCase): diff --git a/test/registered/radix_cache/test_radix_attention.py b/test/registered/radix_cache/test_radix_attention.py index 5931a0921..871dac26e 100644 --- a/test/registered/radix_cache/test_radix_attention.py +++ b/test/registered/radix_cache/test_radix_attention.py @@ -14,8 +14,8 @@ from sglang.test.test_utils import ( ) # RadixAttention server integration tests -register_cuda_ci(est_time=100, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=100, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=100, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=100, suite="stage-b-test-1-gpu-small-amd") class TestRadixCacheFCFS(CustomTestCase): diff --git a/test/registered/radix_cache/test_radix_cache_hit.py b/test/registered/radix_cache/test_radix_cache_hit.py index cc48e19ae..43f63cad7 100644 --- a/test/registered/radix_cache/test_radix_cache_hit.py +++ b/test/registered/radix_cache/test_radix_cache_hit.py @@ -11,7 +11,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=120, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=120, suite="stage-b-test-1-gpu-small") MODEL = DEFAULT_SMALL_MODEL_NAME_FOR_TEST diff --git a/test/registered/radix_cache/test_swa_radix_cache_kl.py b/test/registered/radix_cache/test_swa_radix_cache_kl.py index 617219ceb..db60d0055 100644 --- a/test/registered/radix_cache/test_swa_radix_cache_kl.py +++ b/test/registered/radix_cache/test_swa_radix_cache_kl.py @@ -6,7 +6,7 @@ from sglang.test.server_fixtures.default_fixture import DefaultServerBase MODEL = "openai/gpt-oss-20b" -register_cuda_ci(est_time=100, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=100, suite="stage-b-test-1-gpu-large") class TestSWARadixCacheKL(KLDivergenceMixin, DefaultServerBase): diff --git a/test/registered/rl/test_fp32_lm_head.py b/test/registered/rl/test_fp32_lm_head.py index 5f19332e7..8b721cc4e 100644 --- a/test/registered/rl/test_fp32_lm_head.py +++ b/test/registered/rl/test_fp32_lm_head.py @@ -15,8 +15,8 @@ from sglang.srt.server_args import ( from sglang.srt.utils import get_device from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=9, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=15, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=9, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=15, suite="stage-b-test-1-gpu-small-amd") class LMHeadStub(nn.Module): diff --git a/test/registered/rl/test_lora_load_from_tensor.py b/test/registered/rl/test_lora_load_from_tensor.py index fd0113a73..690eb266a 100644 --- a/test/registered/rl/test_lora_load_from_tensor.py +++ b/test/registered/rl/test_lora_load_from_tensor.py @@ -10,8 +10,8 @@ import sglang as sgl from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=90, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=90, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=90, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=90, suite="stage-b-test-1-gpu-small-amd") MODEL_PATH = "Qwen/Qwen3-0.6B" LORA_REPO = "charent/self_cognition_Alice" diff --git a/test/registered/rl/test_patch_torch.py b/test/registered/rl/test_patch_torch.py index 793877c28..2c46e6754 100644 --- a/test/registered/rl/test_patch_torch.py +++ b/test/registered/rl/test_patch_torch.py @@ -10,9 +10,9 @@ from sglang.srt.utils.patch_torch import monkey_patch_torch_reductions from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci register_amd_ci( - est_time=19, suite="stage-b-test-large-2-gpu-amd", disabled="see #11127" + est_time=19, suite="stage-b-test-2-gpu-large-amd", disabled="see #11127" ) -register_cuda_ci(est_time=19, suite="stage-b-test-large-2-gpu") +register_cuda_ci(est_time=19, suite="stage-b-test-2-gpu-large") class TestReleaseMemoryOccupation(unittest.TestCase): diff --git a/test/registered/rl/test_update_weights_from_disk.py b/test/registered/rl/test_update_weights_from_disk.py index 080210659..076bbe1bb 100644 --- a/test/registered/rl/test_update_weights_from_disk.py +++ b/test/registered/rl/test_update_weights_from_disk.py @@ -19,9 +19,9 @@ from sglang.test.test_utils import ( ) register_amd_ci( - est_time=210, suite="stage-b-test-small-1-gpu-amd", disabled="see #14021" + est_time=210, suite="stage-b-test-1-gpu-small-amd", disabled="see #14021" ) -register_cuda_ci(est_time=210, suite="stage-b-test-large-1-gpu", disabled="see #14021") +register_cuda_ci(est_time=210, suite="stage-b-test-1-gpu-large", disabled="see #14021") ############################################################################### diff --git a/test/registered/rl/test_update_weights_from_distributed.py b/test/registered/rl/test_update_weights_from_distributed.py index 42e3a28ae..d0cad0285 100644 --- a/test/registered/rl/test_update_weights_from_distributed.py +++ b/test/registered/rl/test_update_weights_from_distributed.py @@ -43,8 +43,8 @@ from sglang.test.test_utils import ( ) from sglang.utils import terminate_process -register_cuda_ci(est_time=103, suite="stage-b-test-large-2-gpu") -register_amd_ci(est_time=103, suite="stage-b-test-large-2-gpu-amd") +register_cuda_ci(est_time=103, suite="stage-b-test-2-gpu-large") +register_amd_ci(est_time=103, suite="stage-b-test-2-gpu-large-amd") mp.set_start_method("spawn", force=True) diff --git a/test/registered/rl/test_update_weights_from_tensor.py b/test/registered/rl/test_update_weights_from_tensor.py index 7cb023071..aebe71e2a 100644 --- a/test/registered/rl/test_update_weights_from_tensor.py +++ b/test/registered/rl/test_update_weights_from_tensor.py @@ -1,7 +1,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=195, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=195, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=195, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=195, suite="stage-b-test-1-gpu-small-amd") import gc import json diff --git a/test/registered/rotary/test_mrope.py b/test/registered/rotary/test_mrope.py index 3b2c8c56f..64a0b017b 100644 --- a/test/registered/rotary/test_mrope.py +++ b/test/registered/rotary/test_mrope.py @@ -20,8 +20,8 @@ from sglang.srt.utils import ( ) from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=10, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=15, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=10, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=15, suite="stage-b-test-1-gpu-small-amd") _is_cuda = is_cuda() _is_hip = is_hip() diff --git a/test/registered/rotary/test_rope_rocm.py b/test/registered/rotary/test_rope_rocm.py index 7ab788555..3a982051c 100644 --- a/test/registered/rotary/test_rope_rocm.py +++ b/test/registered/rotary/test_rope_rocm.py @@ -7,7 +7,7 @@ from sglang.srt.utils import get_bool_env_var, is_hip from sglang.test.ci.ci_register import register_amd_ci from sglang.test.test_utils import CustomTestCase -register_amd_ci(est_time=3, suite="stage-b-test-small-1-gpu-amd") +register_amd_ci(est_time=3, suite="stage-b-test-1-gpu-small-amd") torch.manual_seed(0) diff --git a/test/registered/sampling/test_original_logprobs.py b/test/registered/sampling/test_original_logprobs.py index ed78daa60..47b4a9c57 100644 --- a/test/registered/sampling/test_original_logprobs.py +++ b/test/registered/sampling/test_original_logprobs.py @@ -25,8 +25,8 @@ import sglang as sgl from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.test_utils import DEFAULT_SMALL_MODEL_NAME_FOR_TEST -register_cuda_ci(est_time=41, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=60, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=41, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=60, suite="stage-b-test-1-gpu-small-amd") # ------------------------- Configurable via env ------------------------- # MODEL_ID = DEFAULT_SMALL_MODEL_NAME_FOR_TEST diff --git a/test/registered/sampling/test_penalty.py b/test/registered/sampling/test_penalty.py index 11e7e13a4..df456e0f7 100644 --- a/test/registered/sampling/test_penalty.py +++ b/test/registered/sampling/test_penalty.py @@ -16,8 +16,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=82, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=82, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=82, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=82, suite="stage-b-test-1-gpu-small-amd") class TestPenalty(CustomTestCase): diff --git a/test/registered/sampling/test_pytorch_sampling_backend.py b/test/registered/sampling/test_pytorch_sampling_backend.py index b6953990c..501d70678 100644 --- a/test/registered/sampling/test_pytorch_sampling_backend.py +++ b/test/registered/sampling/test_pytorch_sampling_backend.py @@ -14,8 +14,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=66, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=66, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=66, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=66, suite="stage-b-test-1-gpu-small-amd") class TestPyTorchSamplingBackend(CustomTestCase): diff --git a/test/registered/scheduler/test_abort.py b/test/registered/scheduler/test_abort.py index e2f5d16f5..ce53425d9 100644 --- a/test/registered/scheduler/test_abort.py +++ b/test/registered/scheduler/test_abort.py @@ -19,8 +19,8 @@ from sglang.test.test_utils import ( run_and_check_memory_leak, ) -register_cuda_ci(est_time=131, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=300, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=131, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=300, suite="stage-b-test-1-gpu-small-amd") class TestAbort(CustomTestCase): diff --git a/test/registered/scheduler/test_abort_with_metrics.py b/test/registered/scheduler/test_abort_with_metrics.py index c2adfdd54..5144d03ea 100644 --- a/test/registered/scheduler/test_abort_with_metrics.py +++ b/test/registered/scheduler/test_abort_with_metrics.py @@ -18,7 +18,7 @@ from sglang.srt.utils.http_middleware_patch import _PureASGIDispatch from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=10, suite="stage-a-cpu-only") +register_cuda_ci(est_time=10, suite="stage-a-test-cpu") _HTTP_SCOPE = { "type": "http", diff --git a/test/registered/scheduler/test_chunked_prefill.py b/test/registered/scheduler/test_chunked_prefill.py index 951b09f74..9b72ce21f 100644 --- a/test/registered/scheduler/test_chunked_prefill.py +++ b/test/registered/scheduler/test_chunked_prefill.py @@ -7,8 +7,8 @@ import unittest from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.test_utils import CustomTestCase, run_mmlu_test, run_mulit_request_test -register_cuda_ci(est_time=312, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=312, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=312, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=312, suite="stage-b-test-1-gpu-small-amd") class TestChunkedPrefill(CustomTestCase): diff --git a/test/registered/scheduler/test_no_chunked_prefill.py b/test/registered/scheduler/test_no_chunked_prefill.py index 2e871b69c..a255346d0 100644 --- a/test/registered/scheduler/test_no_chunked_prefill.py +++ b/test/registered/scheduler/test_no_chunked_prefill.py @@ -8,8 +8,8 @@ from sglang.test.test_utils import ( run_mmlu_test, ) -register_cuda_ci(est_time=108, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=108, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=108, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=108, suite="stage-b-test-1-gpu-small-amd") class TestNoChunkedPrefill(CustomTestCase): diff --git a/test/registered/scheduler/test_no_overlap_scheduler.py b/test/registered/scheduler/test_no_overlap_scheduler.py index 2f461faf7..8ae06fe16 100644 --- a/test/registered/scheduler/test_no_overlap_scheduler.py +++ b/test/registered/scheduler/test_no_overlap_scheduler.py @@ -9,8 +9,8 @@ import unittest from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.test_utils import CustomTestCase, run_mmlu_test -register_cuda_ci(est_time=245, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=275, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=245, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=275, suite="stage-b-test-1-gpu-small-amd") class TestOverlapSchedule(CustomTestCase): diff --git a/test/registered/scheduler/test_priority_scheduling.py b/test/registered/scheduler/test_priority_scheduling.py index b3ba23ea0..339db350a 100644 --- a/test/registered/scheduler/test_priority_scheduling.py +++ b/test/registered/scheduler/test_priority_scheduling.py @@ -17,8 +17,8 @@ from sglang.test.test_utils import ( send_concurrent_generate_requests_with_custom_params, ) -register_cuda_ci(est_time=130, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=195, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=130, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=195, suite="stage-b-test-1-gpu-small-amd") class TestPriorityScheduling(CustomTestCase): diff --git a/test/registered/scheduler/test_retract_decode.py b/test/registered/scheduler/test_retract_decode.py index 2c59ec247..15f333104 100644 --- a/test/registered/scheduler/test_retract_decode.py +++ b/test/registered/scheduler/test_retract_decode.py @@ -17,8 +17,8 @@ from sglang.test.test_utils import ( ) from sglang.utils import is_in_ci -register_cuda_ci(est_time=311, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=600, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=311, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=600, suite="stage-b-test-1-gpu-small-amd") class TestRetractDecode(CustomTestCase): diff --git a/test/registered/sessions/test_session_control.py b/test/registered/sessions/test_session_control.py index b9c4a4459..12dbad41b 100644 --- a/test/registered/sessions/test_session_control.py +++ b/test/registered/sessions/test_session_control.py @@ -24,7 +24,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=60, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=60, suite="stage-b-test-1-gpu-large") def remove_prefix(text: str, prefix: str) -> str: diff --git a/test/registered/sessions/test_session_latency.py b/test/registered/sessions/test_session_latency.py index c4ec4d6a9..8195b9f56 100644 --- a/test/registered/sessions/test_session_latency.py +++ b/test/registered/sessions/test_session_latency.py @@ -33,7 +33,7 @@ from sglang.test.test_utils import ( register_cuda_ci( est_time=122, - suite="stage-b-test-large-1-gpu", + suite="stage-b-test-1-gpu-large", ) NUM_TURNS = 150 diff --git a/test/registered/sessions/test_streaming_session.py b/test/registered/sessions/test_streaming_session.py index 14668b6da..ed2da9cc3 100644 --- a/test/registered/sessions/test_streaming_session.py +++ b/test/registered/sessions/test_streaming_session.py @@ -28,7 +28,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=70, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=70, suite="stage-b-test-1-gpu-large") # --------------------------------------------------------------------------- # Logprob leak constants diff --git a/test/registered/spec/eagle/test_eagle3_basic.py b/test/registered/spec/eagle/test_eagle3_basic.py index 98526c7e4..442eefcc6 100644 --- a/test/registered/spec/eagle/test_eagle3_basic.py +++ b/test/registered/spec/eagle/test_eagle3_basic.py @@ -12,8 +12,8 @@ from sglang.test.test_utils import ( DEFAULT_TARGET_MODEL_EAGLE3, ) -register_cuda_ci(est_time=50, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=50, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=50, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=50, suite="stage-b-test-1-gpu-small") _is_hip = is_hip() diff --git a/test/registered/spec/eagle/test_eagle_constrained_decoding.py b/test/registered/spec/eagle/test_eagle_constrained_decoding.py index 6cdc1e9f5..6fa1dd5a3 100644 --- a/test/registered/spec/eagle/test_eagle_constrained_decoding.py +++ b/test/registered/spec/eagle/test_eagle_constrained_decoding.py @@ -14,7 +14,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=100, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=100, suite="stage-b-test-1-gpu-large") class TestEagleConstrainedDecoding( diff --git a/test/registered/spec/eagle/test_eagle_infer_a.py b/test/registered/spec/eagle/test_eagle_infer_a.py index d3a4a8f8c..1845fc1b2 100644 --- a/test/registered/spec/eagle/test_eagle_infer_a.py +++ b/test/registered/spec/eagle/test_eagle_infer_a.py @@ -22,7 +22,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=561, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=561, suite="stage-b-test-1-gpu-large") torch_dtype = torch.float16 prefill_tolerance = 5e-2 diff --git a/test/registered/spec/eagle/test_eagle_infer_b.py b/test/registered/spec/eagle/test_eagle_infer_b.py index 011aee68e..014cc5e27 100644 --- a/test/registered/spec/eagle/test_eagle_infer_b.py +++ b/test/registered/spec/eagle/test_eagle_infer_b.py @@ -22,7 +22,7 @@ from sglang.test.kits.radix_cache_server_kit import run_radix_attention_test from sglang.test.server_fixtures.eagle_fixture import EagleServerBase from sglang.test.test_utils import DEFAULT_TARGET_MODEL_EAGLE, run_logprob_check -register_cuda_ci(est_time=1100, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=1100, suite="stage-b-test-1-gpu-large") class TestEAGLEServerBasic(EagleServerBase): diff --git a/test/registered/spec/eagle/test_eagle_infer_beta.py b/test/registered/spec/eagle/test_eagle_infer_beta.py index 3124877a9..252062611 100644 --- a/test/registered/spec/eagle/test_eagle_infer_beta.py +++ b/test/registered/spec/eagle/test_eagle_infer_beta.py @@ -19,7 +19,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=283, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=283, suite="stage-b-test-1-gpu-small") class TestEagleServerBase(CustomTestCase, MatchedStopMixin): diff --git a/test/registered/spec/test_constrained_decoding_spec_reasoning.py b/test/registered/spec/test_constrained_decoding_spec_reasoning.py index 18931ec4d..07f59bd2d 100644 --- a/test/registered/spec/test_constrained_decoding_spec_reasoning.py +++ b/test/registered/spec/test_constrained_decoding_spec_reasoning.py @@ -14,7 +14,7 @@ from sglang.test.test_utils import ( ) # Constrained decoding with EAGLE3 speculative reasoning (tp=2) -register_cuda_ci(est_time=60, suite="stage-b-test-large-2-gpu") +register_cuda_ci(est_time=60, suite="stage-b-test-2-gpu-large") class ServerWithGrammar(CustomTestCase): diff --git a/test/registered/spec/test_ngram_speculative_decoding.py b/test/registered/spec/test_ngram_speculative_decoding.py index 1b87c86f5..2b16c1ba2 100644 --- a/test/registered/spec/test_ngram_speculative_decoding.py +++ b/test/registered/spec/test_ngram_speculative_decoding.py @@ -12,7 +12,7 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=230, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=230, suite="stage-b-test-1-gpu-large") GSM_DATASET_PATH = None diff --git a/test/registered/spec/test_standalone_speculative_decoding.py b/test/registered/spec/test_standalone_speculative_decoding.py index 3ab9272c2..1a3cc0647 100644 --- a/test/registered/spec/test_standalone_speculative_decoding.py +++ b/test/registered/spec/test_standalone_speculative_decoding.py @@ -18,7 +18,7 @@ from sglang.test.test_utils import ( ) # Standalone speculative decoding tests (FA3, Triton, FlashInfer backends) -register_cuda_ci(est_time=308, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=308, suite="stage-b-test-1-gpu-large") GSM_DATASET_PATH = None diff --git a/test/registered/spec/utils/test_build_eagle_tree.py b/test/registered/spec/utils/test_build_eagle_tree.py index 4a60e8475..f13ede2bf 100644 --- a/test/registered/spec/utils/test_build_eagle_tree.py +++ b/test/registered/spec/utils/test_build_eagle_tree.py @@ -9,8 +9,8 @@ from sglang.srt.speculative.eagle_utils import ( from sglang.srt.utils import get_device from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=3, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=3, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=3, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=3, suite="stage-b-test-1-gpu-small-amd") class TestBuildEagleTree(unittest.TestCase): diff --git a/test/registered/spec/utils/test_ngram_corpus.py b/test/registered/spec/utils/test_ngram_corpus.py index b34895885..6f2427a40 100644 --- a/test/registered/spec/utils/test_ngram_corpus.py +++ b/test/registered/spec/utils/test_ngram_corpus.py @@ -6,7 +6,7 @@ from sglang.srt.speculative.cpp_ngram.ngram_corpus import NgramCorpus from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=30, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=30, suite="stage-b-test-1-gpu-small") def _make_corpus(match_type="BFS", **kwargs): diff --git a/test/registered/tokenizer/test_multi_tokenizer.py b/test/registered/tokenizer/test_multi_tokenizer.py index 1084b0a2f..2e07a2f73 100644 --- a/test/registered/tokenizer/test_multi_tokenizer.py +++ b/test/registered/tokenizer/test_multi_tokenizer.py @@ -17,8 +17,8 @@ from sglang.test.test_utils import ( write_github_step_summary, ) -register_cuda_ci(est_time=230, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=345, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=230, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=345, suite="stage-b-test-1-gpu-small-amd") class TestMultiTokenizer(CustomTestCase): diff --git a/test/registered/tokenizer/test_skip_tokenizer_init.py b/test/registered/tokenizer/test_skip_tokenizer_init.py index 852d8bd78..5f821030f 100644 --- a/test/registered/tokenizer/test_skip_tokenizer_init.py +++ b/test/registered/tokenizer/test_skip_tokenizer_init.py @@ -23,8 +23,8 @@ from sglang.test.test_utils import ( popen_launch_server, ) -register_cuda_ci(est_time=77, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=117, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=77, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=117, suite="stage-b-test-1-gpu-small-amd") class TestSkipTokenizerInit(CustomTestCase): diff --git a/test/registered/unit/README.md b/test/registered/unit/README.md index b2010f14f..7d5b36fd0 100644 --- a/test/registered/unit/README.md +++ b/test/registered/unit/README.md @@ -14,8 +14,8 @@ Tests can use CPU or GPU — the key criterion is **no server process**. 3. Register for CI at the **top of the file** (after imports, before test classes): ```python from sglang.test.ci.ci_register import register_cpu_ci - register_cpu_ci(est_time=5, suite="stage-a-cpu-only") - # or: register_cuda_ci(est_time=10, suite="stage-b-test-small-1-gpu") + register_cpu_ci(est_time=5, suite="stage-a-test-cpu") + # or: register_cuda_ci(est_time=10, suite="stage-b-test-1-gpu-small") ``` 4. Run locally: ```bash @@ -39,7 +39,7 @@ Tests can use CPU or GPU — the key criterion is **no server process**. from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=5, suite="stage-a-cpu-only") +register_cpu_ci(est_time=5, suite="stage-a-test-cpu") import unittest diff --git a/test/registered/unit/entrypoints/openai/test_serving_embedding.py b/test/registered/unit/entrypoints/openai/test_serving_embedding.py index 3b9a1bcc1..0e9dd8c4a 100644 --- a/test/registered/unit/entrypoints/openai/test_serving_embedding.py +++ b/test/registered/unit/entrypoints/openai/test_serving_embedding.py @@ -23,7 +23,7 @@ from sglang.srt.entrypoints.openai.serving_embedding import OpenAIServingEmbeddi from sglang.srt.managers.io_struct import EmbeddingReqInput from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=10, suite="stage-a-cpu-only") +register_cpu_ci(est_time=10, suite="stage-a-test-cpu") # Mock TokenizerManager for embedding tests diff --git a/test/registered/unit/entrypoints/test_ssl_cert_refresher.py b/test/registered/unit/entrypoints/test_ssl_cert_refresher.py index c34a8a738..8b8283d06 100644 --- a/test/registered/unit/entrypoints/test_ssl_cert_refresher.py +++ b/test/registered/unit/entrypoints/test_ssl_cert_refresher.py @@ -8,7 +8,7 @@ from sglang.srt.entrypoints.ssl_utils import SSLCertRefresher from sglang.test.ci.ci_register import register_cpu_ci from sglang.test.test_utils import CustomTestCase -register_cpu_ci(est_time=9, suite="stage-a-cpu-only") +register_cpu_ci(est_time=9, suite="stage-a-test-cpu") def _make_temp_pem(content: bytes) -> str: diff --git a/test/registered/unit/function_call/test_function_call_parser.py b/test/registered/unit/function_call/test_function_call_parser.py index 0ba19cc13..c7b3fb173 100644 --- a/test/registered/unit/function_call/test_function_call_parser.py +++ b/test/registered/unit/function_call/test_function_call_parser.py @@ -18,7 +18,7 @@ from sglang.srt.function_call.pythonic_detector import PythonicDetector from sglang.srt.function_call.qwen3_coder_detector import Qwen3CoderDetector from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(1.0, "stage-a-cpu-only") +register_cpu_ci(1.0, "stage-a-test-cpu") class TestPythonicDetector(unittest.TestCase): diff --git a/test/registered/unit/function_call/test_glm47_moe_detector.py b/test/registered/unit/function_call/test_glm47_moe_detector.py index 788580177..357514e19 100644 --- a/test/registered/unit/function_call/test_glm47_moe_detector.py +++ b/test/registered/unit/function_call/test_glm47_moe_detector.py @@ -10,7 +10,7 @@ from sglang.srt.function_call.glm47_moe_detector import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(1.0, "stage-a-cpu-only") +register_cpu_ci(1.0, "stage-a-test-cpu") class TestGlm47MoeDetector(unittest.TestCase): diff --git a/test/registered/unit/function_call/test_json_schema_constraint.py b/test/registered/unit/function_call/test_json_schema_constraint.py index 444864897..bc6a9fa13 100644 --- a/test/registered/unit/function_call/test_json_schema_constraint.py +++ b/test/registered/unit/function_call/test_json_schema_constraint.py @@ -18,7 +18,7 @@ from sglang.srt.function_call.utils import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(1.0, "stage-a-cpu-only") +register_cpu_ci(1.0, "stage-a-test-cpu") class TestJsonSchemaConstraint(unittest.TestCase): diff --git a/test/registered/unit/function_call/test_parallel_tool_calls.py b/test/registered/unit/function_call/test_parallel_tool_calls.py index 8019065df..bf1e18a7b 100644 --- a/test/registered/unit/function_call/test_parallel_tool_calls.py +++ b/test/registered/unit/function_call/test_parallel_tool_calls.py @@ -23,7 +23,7 @@ from sglang.srt.entrypoints.openai.protocol import Function, Tool from sglang.srt.function_call.json_array_parser import JsonArrayParser from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(1.0, "stage-a-cpu-only") +register_cpu_ci(1.0, "stage-a-test-cpu") class TestParallelToolCalls(unittest.TestCase): diff --git a/test/registered/unit/function_call/test_unknown_tool_name.py b/test/registered/unit/function_call/test_unknown_tool_name.py index 922d5f61b..e7a8394ff 100644 --- a/test/registered/unit/function_call/test_unknown_tool_name.py +++ b/test/registered/unit/function_call/test_unknown_tool_name.py @@ -9,7 +9,7 @@ from sglang.srt.function_call.base_format_detector import BaseFormatDetector from sglang.srt.function_call.core_types import StreamingParseResult from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(1.0, "stage-a-cpu-only") +register_cpu_ci(1.0, "stage-a-test-cpu") class DummyDetector(BaseFormatDetector): diff --git a/test/registered/unit/layers/test_conv_layer.py b/test/registered/unit/layers/test_conv_layer.py index e974754d2..b2520f4d2 100644 --- a/test/registered/unit/layers/test_conv_layer.py +++ b/test/registered/unit/layers/test_conv_layer.py @@ -1,6 +1,6 @@ from sglang.test.ci.ci_register import register_cuda_ci -register_cuda_ci(est_time=10, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=10, suite="stage-b-test-1-gpu-small") import unittest diff --git a/test/registered/unit/layers/test_mamba_state_scatter_triton.py b/test/registered/unit/layers/test_mamba_state_scatter_triton.py index 5ff131168..1a94dc6aa 100644 --- a/test/registered/unit/layers/test_mamba_state_scatter_triton.py +++ b/test/registered/unit/layers/test_mamba_state_scatter_triton.py @@ -1,6 +1,6 @@ from sglang.test.ci.ci_register import register_cuda_ci -register_cuda_ci(est_time=10, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=10, suite="stage-b-test-1-gpu-small") import os import unittest diff --git a/test/registered/unit/managers/test_io_struct.py b/test/registered/unit/managers/test_io_struct.py index 037a93e57..47cdf5740 100644 --- a/test/registered/unit/managers/test_io_struct.py +++ b/test/registered/unit/managers/test_io_struct.py @@ -9,8 +9,8 @@ from sglang.test.test_utils import ( CustomTestCase, ) -register_cuda_ci(est_time=8, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=8, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=8, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=8, suite="stage-b-test-1-gpu-small-amd") class TestGenerateReqInputNormalization(CustomTestCase): diff --git a/test/registered/unit/managers/test_prefill_adder.py b/test/registered/unit/managers/test_prefill_adder.py index b37990c3e..e153d9d9c 100644 --- a/test/registered/unit/managers/test_prefill_adder.py +++ b/test/registered/unit/managers/test_prefill_adder.py @@ -12,8 +12,8 @@ from sglang.srt.server_args import ServerArgs, set_global_server_args_for_schedu from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=1, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=2, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=1, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=2, suite="stage-b-test-1-gpu-small-amd") class TestPrefillAdder(CustomTestCase): diff --git a/test/registered/unit/managers/test_profile_merger_http_api.py b/test/registered/unit/managers/test_profile_merger_http_api.py index 9bf656f26..1270fdf83 100644 --- a/test/registered/unit/managers/test_profile_merger_http_api.py +++ b/test/registered/unit/managers/test_profile_merger_http_api.py @@ -4,8 +4,8 @@ import unittest from sglang.srt.managers.io_struct import ProfileReqInput from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=9, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=9, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=9, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=9, suite="stage-b-test-1-gpu-small-amd") class TestProfileMergerHTTPAPI(unittest.TestCase): diff --git a/test/registered/unit/mem_cache/test_evict_policy.py b/test/registered/unit/mem_cache/test_evict_policy.py index 3b4471081..0dd2eed5d 100644 --- a/test/registered/unit/mem_cache/test_evict_policy.py +++ b/test/registered/unit/mem_cache/test_evict_policy.py @@ -2,7 +2,7 @@ from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=5, suite="stage-a-cpu-only") +register_cpu_ci(est_time=5, suite="stage-a-test-cpu") import unittest from unittest.mock import MagicMock diff --git a/test/registered/unit/mem_cache/test_mamba_unittest.py b/test/registered/unit/mem_cache/test_mamba_unittest.py index f067f141e..695572499 100644 --- a/test/registered/unit/mem_cache/test_mamba_unittest.py +++ b/test/registered/unit/mem_cache/test_mamba_unittest.py @@ -21,8 +21,8 @@ from sglang.srt.server_args import ServerArgs, set_global_server_args_for_schedu from sglang.srt.utils import get_device from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=9, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=9, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=9, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=9, suite="stage-b-test-1-gpu-small-amd") class TestMamba(unittest.TestCase): diff --git a/test/registered/unit/mem_cache/test_nsa_pool_host_unit.py b/test/registered/unit/mem_cache/test_nsa_pool_host_unit.py index c945b3684..f75819557 100644 --- a/test/registered/unit/mem_cache/test_nsa_pool_host_unit.py +++ b/test/registered/unit/mem_cache/test_nsa_pool_host_unit.py @@ -11,7 +11,7 @@ from sglang.srt.mem_cache.memory_pool_host import ( from sglang.srt.utils import is_cuda, is_hip, is_npu, is_xpu from sglang.test.ci.ci_register import register_cuda_ci -register_cuda_ci(est_time=3, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=3, suite="stage-b-test-1-gpu-small") class TestNSAHiCacheTransfer(unittest.TestCase): diff --git a/test/registered/unit/mem_cache/test_radix_cache_slru_accuracy.py b/test/registered/unit/mem_cache/test_radix_cache_slru_accuracy.py index 3723f3c0f..7d496f2e9 100644 --- a/test/registered/unit/mem_cache/test_radix_cache_slru_accuracy.py +++ b/test/registered/unit/mem_cache/test_radix_cache_slru_accuracy.py @@ -13,7 +13,7 @@ from sglang.srt.mem_cache.memory_pool import MHATokenToKVPool, ReqToTokenPool from sglang.srt.mem_cache.radix_cache import RadixCache, RadixKey from sglang.test.ci.ci_register import register_cuda_ci -register_cuda_ci(est_time=5, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=5, suite="stage-b-test-1-gpu-small") class TestSLRUAccuracy(unittest.TestCase): diff --git a/test/registered/unit/mem_cache/test_radix_cache_unit.py b/test/registered/unit/mem_cache/test_radix_cache_unit.py index 41d202ad6..33dc31f41 100644 --- a/test/registered/unit/mem_cache/test_radix_cache_unit.py +++ b/test/registered/unit/mem_cache/test_radix_cache_unit.py @@ -21,8 +21,8 @@ from sglang.srt.mem_cache.common import available_and_evictable_str from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci # CPU-based unit test, runs quickly on any GPU runner -register_cuda_ci(est_time=5, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=5, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=5, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=5, suite="stage-b-test-1-gpu-small-amd") import random import time diff --git a/test/registered/unit/mem_cache/test_swa_unittest.py b/test/registered/unit/mem_cache/test_swa_unittest.py index f027c1fae..009f956df 100644 --- a/test/registered/unit/mem_cache/test_swa_unittest.py +++ b/test/registered/unit/mem_cache/test_swa_unittest.py @@ -17,8 +17,8 @@ from sglang.srt.mem_cache.swa_radix_cache import SWARadixCache from sglang.srt.utils import get_device from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=8, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=8, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=10, suite="stage-b-test-1-gpu-small-amd") class TestSWA(unittest.TestCase): diff --git a/test/registered/unit/model_executor/test_model_hooks.py b/test/registered/unit/model_executor/test_model_hooks.py index ae14a2abd..33c9ee82a 100644 --- a/test/registered/unit/model_executor/test_model_hooks.py +++ b/test/registered/unit/model_executor/test_model_hooks.py @@ -9,8 +9,8 @@ from sglang.srt.server_args import ServerArgs from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.test_utils import CustomTestCase -register_cuda_ci(est_time=6, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=6, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=10, suite="stage-b-test-1-gpu-small-amd") HOOK_CALLS = [] diff --git a/test/registered/unit/model_loader/test_modelopt_export.py b/test/registered/unit/model_loader/test_modelopt_export.py index 21173c113..be448d467 100644 --- a/test/registered/unit/model_loader/test_modelopt_export.py +++ b/test/registered/unit/model_loader/test_modelopt_export.py @@ -19,8 +19,8 @@ from sglang.srt.configs.model_config import ModelConfig from sglang.srt.model_loader.loader import ModelOptModelLoader from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=9, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=9, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=9, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=9, suite="stage-b-test-1-gpu-small-amd") # Note: PYTHONPATH=python should be set when running tests diff --git a/test/registered/unit/model_loader/test_modelopt_loader.py b/test/registered/unit/model_loader/test_modelopt_loader.py index 2658a87a5..7f9652c0e 100644 --- a/test/registered/unit/model_loader/test_modelopt_loader.py +++ b/test/registered/unit/model_loader/test_modelopt_loader.py @@ -30,7 +30,7 @@ CALIBRATION_BATCH_SIZE = 36 CALIBRATION_NUM_SAMPLES = 512 DEFAULT_DEVICE = "cuda:0" -register_cuda_ci(est_time=11, suite="stage-b-test-small-1-gpu") +register_cuda_ci(est_time=11, suite="stage-b-test-1-gpu-small") class TestModelOptModelLoader(CustomTestCase): diff --git a/test/registered/unit/observability/test_cpu_monitor.py b/test/registered/unit/observability/test_cpu_monitor.py index fe84c0413..e4d34d46d 100644 --- a/test/registered/unit/observability/test_cpu_monitor.py +++ b/test/registered/unit/observability/test_cpu_monitor.py @@ -3,7 +3,7 @@ import unittest from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=60, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=60, suite="stage-a-test-cpu", nightly=True) class TestCpuMonitor(unittest.TestCase): diff --git a/test/registered/unit/observability/test_metrics_utils.py b/test/registered/unit/observability/test_metrics_utils.py index fc2c8bc74..1d20fd944 100644 --- a/test/registered/unit/observability/test_metrics_utils.py +++ b/test/registered/unit/observability/test_metrics_utils.py @@ -6,7 +6,7 @@ from sglang.srt.observability.utils import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=1, suite="stage-a-cpu-only") +register_cpu_ci(est_time=1, suite="stage-a-test-cpu") class TestMetricsUtils(unittest.TestCase): diff --git a/test/registered/unit/parser/test_harmony_parser.py b/test/registered/unit/parser/test_harmony_parser.py index cee13af88..9a736a64f 100644 --- a/test/registered/unit/parser/test_harmony_parser.py +++ b/test/registered/unit/parser/test_harmony_parser.py @@ -12,7 +12,7 @@ from sglang.srt.parser.harmony_parser import ( from sglang.test.ci.ci_register import register_cpu_ci from sglang.test.test_utils import CustomTestCase -register_cpu_ci(est_time=6, suite="stage-a-cpu-only") +register_cpu_ci(est_time=6, suite="stage-a-test-cpu") class TestEvent(CustomTestCase): diff --git a/test/registered/unit/parser/test_jinja_template_utils.py b/test/registered/unit/parser/test_jinja_template_utils.py index 0bb4161fe..dfb6ff79c 100644 --- a/test/registered/unit/parser/test_jinja_template_utils.py +++ b/test/registered/unit/parser/test_jinja_template_utils.py @@ -11,7 +11,7 @@ from sglang.srt.parser.jinja_template_utils import ( from sglang.test.ci.ci_register import register_cpu_ci from sglang.test.test_utils import CustomTestCase -register_cpu_ci(est_time=7, suite="stage-a-cpu-only") +register_cpu_ci(est_time=7, suite="stage-a-test-cpu") class TestTemplateContentFormatDetection(CustomTestCase): diff --git a/test/registered/unit/parser/test_reasoning_parser.py b/test/registered/unit/parser/test_reasoning_parser.py index 3e47281b4..607a6ca52 100644 --- a/test/registered/unit/parser/test_reasoning_parser.py +++ b/test/registered/unit/parser/test_reasoning_parser.py @@ -14,7 +14,7 @@ from sglang.srt.parser.reasoning_parser import ( from sglang.test.ci.ci_register import register_cpu_ci from sglang.test.test_utils import CustomTestCase -register_cpu_ci(est_time=5, suite="stage-a-cpu-only") +register_cpu_ci(est_time=5, suite="stage-a-test-cpu") class TestStreamingParseResult(CustomTestCase): diff --git a/test/registered/unit/sampling/test_custom_logit_processor.py b/test/registered/unit/sampling/test_custom_logit_processor.py index bd2df14f3..00922b45b 100644 --- a/test/registered/unit/sampling/test_custom_logit_processor.py +++ b/test/registered/unit/sampling/test_custom_logit_processor.py @@ -2,7 +2,7 @@ from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=5, suite="stage-a-cpu-only") +register_cpu_ci(est_time=5, suite="stage-a-test-cpu") import json import unittest diff --git a/test/registered/unit/sampling/test_penaltylib.py b/test/registered/unit/sampling/test_penaltylib.py index 3b62524ba..8681bcb34 100644 --- a/test/registered/unit/sampling/test_penaltylib.py +++ b/test/registered/unit/sampling/test_penaltylib.py @@ -2,7 +2,7 @@ from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=5, suite="stage-a-cpu-only") +register_cpu_ci(est_time=5, suite="stage-a-test-cpu") import unittest from unittest.mock import MagicMock diff --git a/test/registered/unit/sampling/test_sampling_batch_info.py b/test/registered/unit/sampling/test_sampling_batch_info.py index f6f48089a..bc923932c 100644 --- a/test/registered/unit/sampling/test_sampling_batch_info.py +++ b/test/registered/unit/sampling/test_sampling_batch_info.py @@ -2,7 +2,7 @@ from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=5, suite="stage-a-cpu-only") +register_cpu_ci(est_time=5, suite="stage-a-test-cpu") import unittest from unittest.mock import MagicMock, patch diff --git a/test/registered/unit/sampling/test_sampling_params.py b/test/registered/unit/sampling/test_sampling_params.py index bb16da796..4d2431bf9 100644 --- a/test/registered/unit/sampling/test_sampling_params.py +++ b/test/registered/unit/sampling/test_sampling_params.py @@ -2,7 +2,7 @@ from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=5, suite="stage-a-cpu-only") +register_cpu_ci(est_time=5, suite="stage-a-test-cpu") import unittest from unittest.mock import MagicMock diff --git a/test/registered/unit/server_args/test_server_args.py b/test/registered/unit/server_args/test_server_args.py index 0b4cdda3f..1276493d3 100644 --- a/test/registered/unit/server_args/test_server_args.py +++ b/test/registered/unit/server_args/test_server_args.py @@ -10,7 +10,7 @@ from sglang.test.test_utils import ( CustomTestCase, ) -register_cpu_ci(est_time=1, suite="stage-a-cpu-only") +register_cpu_ci(est_time=1, suite="stage-a-test-cpu") # Mock get_device() so all tests run on CPU-only CI runners _mock_device = patch("sglang.srt.server_args.get_device", return_value="cuda") diff --git a/test/registered/unit/utils/test_gauge_histogram.py b/test/registered/unit/utils/test_gauge_histogram.py index 5450e7d6b..5f8dab542 100644 --- a/test/registered/unit/utils/test_gauge_histogram.py +++ b/test/registered/unit/utils/test_gauge_histogram.py @@ -1,6 +1,6 @@ from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=5, suite="stage-a-cpu-only") +register_cpu_ci(est_time=5, suite="stage-a-test-cpu") import unittest diff --git a/test/registered/unit/utils/test_json_response.py b/test/registered/unit/utils/test_json_response.py index 20fbc8108..f201ae735 100644 --- a/test/registered/unit/utils/test_json_response.py +++ b/test/registered/unit/utils/test_json_response.py @@ -10,7 +10,7 @@ from sglang.srt.utils.json_response import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=2, suite="stage-a-cpu-only") +register_cpu_ci(est_time=2, suite="stage-a-test-cpu") class TestJSONResponseUtils(unittest.TestCase): diff --git a/test/registered/unit/utils/test_patch_tokenizer.py b/test/registered/unit/utils/test_patch_tokenizer.py index 824b65959..ea0ba7533 100644 --- a/test/registered/unit/utils/test_patch_tokenizer.py +++ b/test/registered/unit/utils/test_patch_tokenizer.py @@ -10,7 +10,7 @@ from sglang.srt.utils.patch_tokenizer import ( ) from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=30, suite="stage-a-cpu-only", nightly=True) +register_cpu_ci(est_time=30, suite="stage-a-test-cpu", nightly=True) class TestPatchTokenizerEndToEndTest(unittest.TestCase): diff --git a/test/registered/unit/utils/test_profile_merger.py b/test/registered/unit/utils/test_profile_merger.py index 412ae8046..5c00e793e 100644 --- a/test/registered/unit/utils/test_profile_merger.py +++ b/test/registered/unit/utils/test_profile_merger.py @@ -17,8 +17,8 @@ from sglang.srt.managers.io_struct import ProfileReq, ProfileReqInput, ProfileRe from sglang.srt.utils.profile_merger import ProfileMerger from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=8, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=8, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=8, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=8, suite="stage-b-test-1-gpu-small-amd") class TestProfileMerger(unittest.TestCase): diff --git a/test/registered/utils/test_bench_typebaseddispatcher.py b/test/registered/utils/test_bench_typebaseddispatcher.py index 02cccc4dc..dc0365a6f 100644 --- a/test/registered/utils/test_bench_typebaseddispatcher.py +++ b/test/registered/utils/test_bench_typebaseddispatcher.py @@ -4,7 +4,7 @@ from typing import Any, Callable, List, Tuple, Type from sglang.test.ci.ci_register import register_amd_ci from sglang.utils import TypeBasedDispatcher -register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu-amd") +register_amd_ci(est_time=10, suite="stage-b-test-1-gpu-small-amd") class TypeBasedDispatcherList: diff --git a/test/registered/utils/test_log_utils.py b/test/registered/utils/test_log_utils.py index cfefe8f43..810c12f36 100644 --- a/test/registered/utils/test_log_utils.py +++ b/test/registered/utils/test_log_utils.py @@ -9,7 +9,7 @@ from pathlib import Path from sglang.srt.utils.log_utils import create_log_targets, log_json from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=1, suite="stage-a-cpu-only") +register_cpu_ci(est_time=1, suite="stage-a-test-cpu") class TestLogUtils(unittest.TestCase): diff --git a/test/registered/utils/test_network_address.py b/test/registered/utils/test_network_address.py index 922866f26..967345fae 100644 --- a/test/registered/utils/test_network_address.py +++ b/test/registered/utils/test_network_address.py @@ -6,7 +6,7 @@ from sglang.srt.server_args import PortArgs, ServerArgs from sglang.srt.utils.network import NetworkAddress from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=1, suite="stage-a-cpu-only") +register_cpu_ci(est_time=1, suite="stage-a-test-cpu") # Mock get_device() so ServerArgs tests run on CPU-only CI runners _mock_device = patch("sglang.srt.server_args.get_device", return_value="cuda") diff --git a/test/registered/utils/test_socket_utils.py b/test/registered/utils/test_socket_utils.py index 8b7e7fafd..42c12fd88 100644 --- a/test/registered/utils/test_socket_utils.py +++ b/test/registered/utils/test_socket_utils.py @@ -15,7 +15,7 @@ from sglang.test.ci.ci_register import register_cpu_ci from sglang.test.test_utils import CustomTestCase from sglang.utils import normalize_base_url, release_port, reserve_port -register_cpu_ci(est_time=1, suite="stage-a-cpu-only") +register_cpu_ci(est_time=1, suite="stage-a-test-cpu") class TestTryBindSocket(CustomTestCase): diff --git a/test/registered/utils/test_type_based_dispatcher.py b/test/registered/utils/test_type_based_dispatcher.py index 89dc37282..37480a75f 100644 --- a/test/registered/utils/test_type_based_dispatcher.py +++ b/test/registered/utils/test_type_based_dispatcher.py @@ -11,7 +11,7 @@ from sglang.srt.managers.io_struct import SamplingParams from sglang.test.ci.ci_register import register_amd_ci from sglang.utils import TypeBasedDispatcher -register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu-amd") +register_amd_ci(est_time=10, suite="stage-b-test-1-gpu-small-amd") class TestTypeBasedDispatcher(unittest.TestCase): diff --git a/test/registered/vlm/test_evs.py b/test/registered/vlm/test_evs.py index 7ae4d1aa2..1f1da070f 100644 --- a/test/registered/vlm/test_evs.py +++ b/test/registered/vlm/test_evs.py @@ -6,8 +6,8 @@ import pytest from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci from sglang.test.test_utils import run_doctests -register_cuda_ci(est_time=20, suite="stage-b-test-small-1-gpu") -register_amd_ci(est_time=20, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=20, suite="stage-b-test-1-gpu-small") +register_amd_ci(est_time=20, suite="stage-b-test-1-gpu-small-amd") def test_resolve_evs_config(): diff --git a/test/registered/vlm/test_patch_embed_perf.py b/test/registered/vlm/test_patch_embed_perf.py index b8724ff0c..ce2f38ca2 100644 --- a/test/registered/vlm/test_patch_embed_perf.py +++ b/test/registered/vlm/test_patch_embed_perf.py @@ -8,7 +8,7 @@ import torch.nn as nn from sglang.srt.models.glm4v import Glm4vVisionPatchEmbed from sglang.test.ci.ci_register import register_cuda_ci -register_cuda_ci(est_time=120, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=120, suite="stage-b-test-1-gpu-large") PATCH_SIZE = 14 TEMPORAL_PATCH_SIZE = 2 diff --git a/test/registered/vlm/test_video_utils.py b/test/registered/vlm/test_video_utils.py index 7ba166609..5163f3519 100644 --- a/test/registered/vlm/test_video_utils.py +++ b/test/registered/vlm/test_video_utils.py @@ -5,7 +5,7 @@ import pytest from sglang.srt.utils import sample_video_frames from sglang.test.ci.ci_register import register_cpu_ci -register_cpu_ci(est_time=5, suite="stage-a-cpu-only") +register_cpu_ci(est_time=5, suite="stage-a-test-cpu") class DummyVideo: diff --git a/test/registered/vlm/test_vision_chunked_prefill.py b/test/registered/vlm/test_vision_chunked_prefill.py index 27cfab3a3..f078fab02 100644 --- a/test/registered/vlm/test_vision_chunked_prefill.py +++ b/test/registered/vlm/test_vision_chunked_prefill.py @@ -1,7 +1,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci -register_cuda_ci(est_time=150, suite="stage-b-test-large-1-gpu") -register_amd_ci(est_time=270, suite="stage-b-test-small-1-gpu-amd") +register_cuda_ci(est_time=150, suite="stage-b-test-1-gpu-large") +register_amd_ci(est_time=270, suite="stage-b-test-1-gpu-small-amd") """ Usage: python3 -m unittest test_vision_chunked_prefill.TestVisionChunkedPrefill.test_chunked_prefill diff --git a/test/registered/vlm/test_vision_openai_server_a.py b/test/registered/vlm/test_vision_openai_server_a.py index e89f3bbfc..9bf29aa6b 100644 --- a/test/registered/vlm/test_vision_openai_server_a.py +++ b/test/registered/vlm/test_vision_openai_server_a.py @@ -19,7 +19,7 @@ from sglang.test.vlm_utils import ( VideoOpenAITestMixin, ) -register_cuda_ci(est_time=957, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=957, suite="stage-b-test-1-gpu-large") class TestLlavaServer(ImageOpenAITestMixin): diff --git a/test/registered/vlm/test_vlm_input_format.py b/test/registered/vlm/test_vlm_input_format.py index 92477c799..de5369d80 100644 --- a/test/registered/vlm/test_vlm_input_format.py +++ b/test/registered/vlm/test_vlm_input_format.py @@ -37,7 +37,7 @@ from sglang.srt.entrypoints.openai.protocol import ChatCompletionRequest from sglang.srt.parser.conversation import generate_chat_conv from sglang.srt.utils.hf_transformers_utils import _fix_added_tokens_encoding -register_cuda_ci(est_time=447, suite="stage-b-test-large-1-gpu") +register_cuda_ci(est_time=447, suite="stage-b-test-1-gpu-large") IMAGE_MAN_IRONING_URL = "https://raw.githubusercontent.com/sgl-project/sgl-test-files/refs/heads/main/images/man_ironing_on_back_of_suv.png" IMAGE_SGL_LOGO_URL = "https://raw.githubusercontent.com/sgl-project/sgl-test-files/refs/heads/main/images/sgl_logo.png" diff --git a/test/run_suite.py b/test/run_suite.py index 89cd529ee..876fe9021 100644 --- a/test/run_suite.py +++ b/test/run_suite.py @@ -22,34 +22,34 @@ HW_MAPPING = { # Per-commit test suites (run on every PR) PER_COMMIT_SUITES = { - HWBackend.CPU: ["stage-a-cpu-only"], + HWBackend.CPU: ["stage-a-test-cpu"], HWBackend.AMD: [ - "stage-a-test-small-1-gpu-amd", - "stage-b-test-small-1-gpu-amd", - "stage-b-test-small-1-gpu-amd-nondeterministic", - "stage-b-test-small-1-gpu-amd-mi35x", + "stage-a-test-1-gpu-small-amd", + "stage-b-test-1-gpu-small-amd", + "stage-b-test-1-gpu-small-amd-nondeterministic", + "stage-b-test-1-gpu-small-amd-mi35x", "stage-b-test-large-8-gpu-35x-disaggregation-amd", - "stage-b-test-large-1-gpu-amd", - "stage-b-test-large-2-gpu-amd", + "stage-b-test-1-gpu-large-amd", + "stage-b-test-2-gpu-large-amd", "stage-c-test-large-8-gpu-amd", "stage-c-test-large-8-gpu-amd-mi35x", ], HWBackend.CUDA: [ - "stage-a-test-small-1-gpu", - "stage-b-test-small-1-gpu", - "stage-b-test-large-1-gpu", - "stage-b-test-large-2-gpu", + "stage-a-test-1-gpu-small", + "stage-b-test-1-gpu-small", + "stage-b-test-1-gpu-large", + "stage-b-test-2-gpu-large", "stage-c-test-4-gpu-h100", "stage-c-test-4-gpu-b200", "stage-c-test-4-gpu-gb200", - "stage-c-test-deepep-4-gpu", + "stage-c-test-deepep-4-gpu-h100", "stage-c-test-8-gpu-h20", "stage-c-test-8-gpu-h200", "stage-c-test-8-gpu-b200", "stage-c-test-deepep-8-gpu-h200", ], HWBackend.NPU: [ - "stage-a-test-small-1-gpu", + "stage-a-test-1-gpu-small", "stage-b-test-1-npu-a2", "stage-b-test-2-npu-a2", "stage-b-test-4-npu-a3",