chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 11 Packages Projects Releases Wiki Activity

Upgrade to ROCm 7.0 image (#13105)

Browse Source
This commit is contained in:
yctseng0211
2025-11-12 11:14:28 +08:00
committed by GitHub
parent 7b877ab83d
commit 28b8c5792d
2 changed files with 7 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
scripts/ci/amd_ci_start_container.sh
View File

@@ -25,7 +25,7 @@ rm -f "$TMP_VERSION_FILE"
# Default base tags (can be overridden by command line arguments)
DEFAULT_MI30X_BASE_TAG="${SGLANG_VERSION}-rocm630-mi30x"
DEFAULT_MI30X_BASE_TAG="${SGLANG_VERSION}-rocm700-mi30x"
DEFAULT_MI35X_BASE_TAG="${SGLANG_VERSION}-rocm700-mi35x"
# Parse command line arguments
@@ -121,7 +121,7 @@ find_latest_image() {
if [[ "${gpu_arch}" == "mi35x" ]]; then
echo "rocm/sgl-dev:v0.5.5-rocm700-mi35x-20251110"
else
echo "rocm/sgl-dev:v0.5.5-rocm630-mi30x-20251110"
echo "rocm/sgl-dev:v0.5.5-rocm700-mi30x-20251110"
fi
}

10
test/srt/run_suite.py
View File

@@ -395,9 +395,9 @@ suite_amd = {
# TestFile("hicache/test_hicache_storage.py", 127), # Disabled temporarily, see https://github.com/sgl-project/sglang/issues/12575
TestFile("lora/test_lora.py", 150),
TestFile("lora/test_lora_backend.py", 99),
TestFile("lora/test_lora_cuda_graph.py", 250),
# TestFile("lora/test_lora_cuda_graph.py", 250), # Disabled temporarily, see https://github.com/sgl-project/sglang/issues/13107
TestFile("lora/test_lora_eviction.py", 240),
TestFile("lora/test_lora_qwen3.py", 97),
# TestFile("lora/test_lora_qwen3.py", 97), # Disabled temporarily, see https://github.com/sgl-project/sglang/issues/13107
TestFile("lora/test_multi_lora_backend.py", 60),
TestFile("models/test_compressed_tensors_models.py", 42),
TestFile("models/test_qwen_models.py", 82),
@@ -436,7 +436,7 @@ suite_amd = {
TestFile("test_jinja_template_utils.py", 1),
TestFile("test_metrics.py", 32),
TestFile("test_metrics_utils.py", 1),
TestFile("test_mla.py", 242),
# TestFile("test_mla.py", 242), # Disabled temporarily, see https://github.com/sgl-project/sglang/issues/13107
# TestFile("test_mla_deepseek_v3.py", 221), # Temporarily disabled, see https://github.com/sgl-project/sglang/issues/12574
TestFile("test_no_chunked_prefill.py", 108),
TestFile("test_page_size.py", 60),
@@ -452,7 +452,7 @@ suite_amd = {
TestFile("test_srt_endpoint.py", 130),
TestFile("test_srt_engine.py", 261),
TestFile("test_torch_compile.py", 169),
TestFile("test_torch_compile_moe.py", 210),
# 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),
@@ -472,7 +472,7 @@ suite_amd = {
TestFile("test_mla.py", 242),
],
"per-commit-2-gpu-amd": [
TestFile("lora/test_lora_tp.py", 116),
# TestFile("lora/test_lora_tp.py", 116), # Disabled temporarily, see https://github.com/sgl-project/sglang/issues/13107
TestFile("rl/test_update_weights_from_distributed.py", 103),
TestFile("test_data_parallelism.py", 73),
TestFile("test_load_weights_from_remote_instance.py", 72),