Skip to content

[AMD][MI35X] 0508 DSV4 #1300

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
cquil11 merged 4 commits into from
May 8, 2026
Merged

[AMD][MI35X] 0508 DSV4 #1300

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
882e38a
Update dsv4 env vars for mi355 sglang
1am9trash May 8, 2026
e610d11
Tazhu dsv4 patch (#1299)
1am9trash May 8, 2026
3783fe9
Update log
1am9trash May 8, 2026
092f8e1
Update PR link in perf-changelog.yaml
benenzhu May 8, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions .github/configs/amd-master.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1591,7 +1591,7 @@ dsv4-fp8-mi355x-sglang:
# image tag, so bumping sglang is just an image tag bump here. Sweeps
# DP-attention on/off and EP=8.
dsv4-fp4-mi355x-sglang:
image: rocm/sgl-dev:rocm720-mi35x-a8410de-20260502-DSv4
image: rocm/sgl-dev:rocm720-mi35x-bfd32b6-20260507-DSv4
model: deepseek-ai/DeepSeek-V4-Pro
model-prefix: dsv4
runner: mi355x
Expand All @@ -1608,7 +1608,7 @@ dsv4-fp4-mi355x-sglang:
- isl: 8192
osl: 1024
search-space:
- { tp: 8, dp-attn: true, conc-start: 16, conc-end: 128 }
- { tp: 8, dp-attn: true, conc-start: 16, conc-end: 256 }
- { tp: 8, dp-attn: false, conc-start: 1, conc-end: 16 }

# vLLM with AITER MLA decode for DSv4 on MI355X (vllm-project/vllm#40889,
Expand Down
10 changes: 8 additions & 2 deletions benchmarks/single_node/dsv4_fp4_mi355x_sglang.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,7 +11,8 @@ check_env_vars \
ISL \
OSL \
RANDOM_RANGE_RATIO \
RESULT_FILENAME
RESULT_FILENAME \
MAX_MODEL_LEN

if [[ -n "$SLURM_JOB_ID" ]]; then
echo "JOB $SLURM_JOB_ID running on $SLURMD_NODENAME"
Expand Down Expand Up @@ -60,7 +61,9 @@ export SGLANG_OPT_USE_JIT_KERNEL_FUSED_TOPK=false
export SGLANG_OPT_USE_FUSED_HASH_TOPK=false
export SGLANG_OPT_DEEPGEMM_HC_PRENORM=false
export SGLANG_OPT_USE_TILELANG_MHC_PRE=false
export SGLANG_OPT_USE_TILELANG_MHC_POST=true
export SGLANG_OPT_USE_TILELANG_MHC_POST=false
export SGLANG_OPT_USE_AITER_MHC_PRE=true
export SGLANG_OPT_USE_AITER_MHC_POST=true
export SGLANG_ENABLE_THINKING=1
export SGLANG_USE_AITER=1
export SGLANG_USE_ROCM700A=1
Expand All @@ -73,6 +76,7 @@ export SGLANG_OPT_USE_FUSED_STORE_CACHE=false
export SGLANG_FORCE_TRITON_MOE_FP8=0
export SGLANG_HACK_FLASHMLA_BACKEND=tilelang
export SGLANG_OPT_USE_TILELANG_INDEXER=true
export SGLANG_OPT_USE_TRITON_SWA_PREPARE=true

SERVER_LOG=/workspace/server.log
PORT=${PORT:-8888}
Expand All @@ -92,6 +96,7 @@ if [ "${DP_ATTENTION}" = "true" ]; then
PARALLEL_ARGS+=(
--dp "$TP"
--enable-dp-attention
--enable-prefill-delayer
)
fi
if [ "${EP_SIZE:-1}" -gt 1 ]; then
Expand All @@ -109,6 +114,7 @@ python3 -m sglang.launch_server \
--max-running-requests ${CONC} \
--cuda-graph-max-bs ${CONC} \
--page-size 256 \
--context-length $MAX_MODEL_LEN \
Comment thread
benenzhu marked this conversation as resolved.
--chunked-prefill-size 8192 \
--disable-shared-experts-fusion \
--tool-call-parser deepseekv4 \
Expand Down
8 changes: 8 additions & 0 deletions perf-changelog.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2299,3 +2299,11 @@
- "Rename and consolidate the per-concurrency recipe files to `disagg-gb300-{N}p1d-{topo}-{nodes}-c{conc}.yaml`"
- "Re-enable lm-eval scoring for dsv4-fp4-gb300-dynamo-sglang now that the srt-slurm pin includes the lm-eval orchestrator path"
pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1295

- config-keys:
- dsv4-fp4-mi355x-sglang
description:
- "Bump image to rocm/sgl-dev:rocm720-mi35x-bfd32b6-20260507-DSv4."
- "Tune DSv4 FP4 MI355X SGLang runtime envs: enable aiter MHC pre/post, and enable triton swa prepare kernel."
- "Add --context-length. Add --enable-prefill-delayer for dp config"
pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1300