[CI] fix amd 1 gpu basic test (#13551)

.github/workflows/pr-test-amd.yml

@@ -116,7 +116,7 @@ jobs:
       - name: Run test
         timeout-minutes: 10
         run: |
-          docker exec -w /sglang-checkout/test ci_sglang python3 run_suite.py --hw cuda --suite stage-a-test-1
+          docker exec -w /sglang-checkout/test ci_sglang python3 run_suite.py --hw amd --suite stage-a-test-1
 
   unit-test-backend-1-gpu-amd:
     needs: [check-changes, stage-a-test-1-amd]
@@ -149,7 +149,7 @@ jobs:
           bash scripts/ci/amd_ci_exec.sh python3 run_suite.py --suite per-commit-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 12
 
   unit-test-backend-2-gpu-amd:
-    needs: [check-changes]
+    needs: [check-changes, stage-a-test-1-amd]
     if: always() && !failure() && !cancelled() &&
       ((needs.check-changes.outputs.main_package == 'true') || (needs.check-changes.outputs.sgl_kernel == 'true'))
     strategy:
@@ -211,7 +211,7 @@ jobs:
           bash scripts/ci/amd_ci_exec.sh python3 run_suite.py --suite per-commit-8-gpu-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 3 --timeout-per-file 3600
 
   performance-test-1-gpu-part-1-amd:
-    needs: [check-changes]
+    needs: [check-changes, stage-a-test-1-amd]
     if: always() && !failure() && !cancelled() &&
       ((needs.check-changes.outputs.main_package == 'true') || (needs.check-changes.outputs.sgl_kernel == 'true'))
     strategy:
@@ -256,7 +256,7 @@ jobs:
           bash scripts/ci/amd_ci_exec.sh python3 -m unittest test_bench_serving.TestBenchServing.test_offline_throughput_non_stream_small_batch_size
 
   performance-test-1-gpu-part-2-amd:
-    needs: [check-changes]
+    needs: [check-changes, stage-a-test-1-amd]
     if: always() && !failure() && !cancelled() &&
       ((needs.check-changes.outputs.main_package == 'true') || (needs.check-changes.outputs.sgl_kernel == 'true'))
     strategy:
@@ -344,7 +344,7 @@ jobs:
           bash scripts/ci/amd_ci_exec.sh python3 -m unittest test_bench_serving.TestBenchServing.test_moe_offline_throughput_without_radix_cache
 
   accuracy-test-1-gpu-amd:
-    needs: [check-changes]
+    needs: [check-changes, stage-a-test-1-amd]
     if: always() && !failure() && !cancelled() &&
       ((needs.check-changes.outputs.main_package == 'true') || (needs.check-changes.outputs.sgl_kernel == 'true'))
     strategy:

test/per_commit/test_srt_backend.py

@@ -1,7 +1,7 @@
 import unittest
 
 import sglang as sgl
-from sglang.test.ci.ci_register import register_cuda_ci
+from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
 from sglang.test.test_programs import (
     test_decode_int,
     test_decode_json_regex,
@@ -20,6 +20,7 @@ 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-1")
+register_amd_ci(est_time=120, suite="stage-a-test-1")
 
 
 class TestSRTBackend(CustomTestCase):

test/run_suite.py

@@ -8,11 +8,13 @@ from sglang.test.ci.ci_utils import TestFile, run_unittest_files
 HW_MAPPING = {
     "cpu": HWBackend.CPU,
     "cuda": HWBackend.CUDA,
+    "amd": HWBackend.AMD,
 }
 
 LABEL_MAPPING = {
-    HWBackend.CUDA: ["stage-a-test-1"],
     HWBackend.CPU: ["default"],
+    HWBackend.AMD: ["stage-a-test-1"],
+    HWBackend.CUDA: ["stage-a-test-1"],
 }
 
 
@@ -45,7 +47,7 @@ def main():
     parser.add_argument(
         "--hw",
         type=str,
-        choices=["cpu", "cuda"],
+        choices=["cpu", "cuda", "amd"],
         required=True,
         help="Hardware backend to run tests on.",
     )