chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

[CI] Migrate Attention Backend tests to test/registered/attention/ (#15563)

Browse Source
This commit is contained in:
Alison Shao
2025-12-22 22:17:52 -08:00
committed by GitHub
parent 989d4b3012
commit 883747ced1
18 changed files with 94 additions and 22 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/srt/run_suite.py
-19
View File
@@ -54,12 +54,10 @@ suites = {
TestFile("dllm/test_llada2_mini.py", 520),
TestFile("test_abort.py", 131),
TestFile("test_chunked_prefill.py", 312),
TestFile("test_create_kvindices.py", 7),
TestFile("test_deterministic.py", 228),
TestFile("test_constrained_decoding.py", 111),
TestFile("test_eval_fp8_accuracy.py", 250),
TestFile("test_external_models.py", 30),
TestFile("test_fa3.py", 420),
TestFile("test_flashmla.py", 230),
TestFile("test_fp8_utils.py", 9),
TestFile("rotary_embedding/test_mrope.py", 10),
@@ -67,7 +65,6 @@ suites = {
TestFile("test_gpt_oss_1gpu.py", 402),
TestFile("test_harmony_parser.py", 6),
TestFile("test_hidden_states.py", 55),
TestFile("test_hybrid_attn_backend.py", 379),
TestFile("test_input_embeddings.py", 38),
TestFile("test_io_struct.py", 8),
TestFile("test_jinja_template_utils.py", 7),
@@ -91,8 +88,6 @@ suites = {
TestFile("test_prefill_adder.py", 1),
TestFile("test_priority_scheduling.py", 130),
TestFile("test_pytorch_sampling_backend.py", 66),
TestFile("test_radix_attention.py", 105),
TestFile("test_radix_cache_unit.py", 8),
TestFile("test_reasoning_parser.py", 5),
TestFile("test_request_queue_validation.py", 47),
TestFile("test_retract_decode.py", 259),
@@ -110,13 +105,8 @@ suites = {
TestFile("test_torch_compile.py", 190),
TestFile("test_torch_compile_moe.py", 210),
TestFile("test_triton_fused_moe.py", 12),
TestFile("test_torch_native_attention_backend.py", 221),
TestFile("test_torchao.py", 103),
TestFile("test_triton_attention_kernels.py", 4),
TestFile("test_triton_attention_backend.py", 203),
TestFile("test_triton_attention_kernels.py", 4),
TestFile("test_triton_moe_channel_fp8_kernel.py", 16),
TestFile("test_triton_sliding_window.py", 84),
TestFile("test_utils_update_weights.py", 29),
TestFile("test_video_utils.py", 5),
TestFile("test_vision_chunked_prefill.py", 150),
@@ -144,7 +134,6 @@ suites = {
"per-commit-4-gpu": [
TestFile("models/test_qwen3_next_models.py", 650),
TestFile("test_gpt_oss_4gpu.py", 300),
TestFile("test_local_attn.py", 411),
TestFile("test_multi_instance_release_memory_occupation.py", 64),
TestFile("test_pp_single_node.py", 500),
TestFile("test_epd_disaggregation.py", 150),
@@ -166,7 +155,6 @@ suites = {
],
"per-commit-4-gpu-b200": [
TestFile("test_deepseek_v3_fp4_4gpu.py", 1500),
TestFile("test_flash_attention_4.py", 90),
TestFile("test_fp8_blockwise_gemm.py", 280),
TestFile("test_gpt_oss_4gpu.py", 700),
TestFile("test_llama31_fp4.py", 90),
@@ -253,7 +241,6 @@ suite_amd = {
TestFile("test_abort.py", 51),
TestFile("test_bench_typebaseddispatcher.py", 10),
TestFile("test_chunked_prefill.py", 312),
TestFile("test_create_kvindices.py", 2),
TestFile("test_eval_fp8_accuracy.py", 303),
TestFile("test_fused_moe.py", 30),
TestFile("test_harmony_parser.py", 20),
@@ -268,7 +255,6 @@ suite_amd = {
TestFile("test_page_size.py", 60),
TestFile("test_penalty.py", 180),
TestFile("test_pytorch_sampling_backend.py", 66),
TestFile("test_radix_attention.py", 105),
TestFile("test_reasoning_parser.py", 5),
TestFile("test_constrained_decoding.py", 120),
TestFile("test_retract_decode.py", 450),
@@ -279,12 +265,7 @@ suite_amd = {
TestFile("test_srt_engine.py", 261),
TestFile("test_torch_compile.py", 169),
# TestFile("test_torch_compile_moe.py", 210), # Disabled temporarily, see https://github.com/sgl-project/sglang/issues/13107
TestFile("test_torch_native_attention_backend.py", 123),
# TestFile("test_triton_attention_kernels.py", 4),
TestFile("test_triton_attention_backend.py", 150),
TestFile("test_triton_sliding_window.py", 250),
TestFile("test_type_based_dispatcher.py", 10),
TestFile("test_wave_attention_kernels.py", 2),
# Disabled temporarily
# TestFile("test_vlm_input_format.py", 300),
# TestFile("models/test_embedding_models.py", 73), # Disabled temporarily, see https://github.com/sgl-project/sglang/issues/11127