Skip to content

[AMD] feat: support multiple cluster for mi355x DI CI workflow #433

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

Closed
billishyahao wants to merge 2 commits into from
Closed

[AMD] feat: support multiple cluster for mi355x DI CI workflow #433

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
835c4e1
[AMD] feat: support multiple cluster for mi355x distributed inference…
billishyahao Jan 15, 2026
5aa5670
fix nodename
billishyahao Jan 15, 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
2 changes: 1 addition & 1 deletion .github/configs/amd-master.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -251,7 +251,7 @@ dsr1-fp8-mi355x-atom:
- { tp: 8, conc-start: 4, conc-end: 128 }

dsr1-fp8-mi355x-sglang-disagg:
image: rocm/sgl-dev:sglang-0.5.6.post1-rocm700-mi35x-mori-1224
image: rocm/sgl-dev:sglang-0.5.6.post1-rocm700-mi35x-mori-0113
model: deepseek-ai/DeepSeek-R1-0528
model-prefix: dsr1
# NOTE: Until this cluster has more nodes, switch to a single runner because
Expand Down
13 changes: 10 additions & 3 deletions benchmarks/dsr1_fp8_mi355x_sglang-disagg_slurm.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,7 +28,7 @@ check_env_vars \
# git clone --branch cam/sa-251219 https://github.com/cquil11/sglang_disagg.git

# Switch to origin repo url for supporting wide ep configs
git clone --branch sa-260107 https://github.com/billishyahao/sglang_disagg.git
git clone --branch sa-260114 https://github.com/billishyahao/sglang_disagg.git

cd "$SGL_SLURM_JOBS_PATH" || exit 1

Expand Down Expand Up @@ -58,11 +58,18 @@ fi
# Replace ' ' in CONC_LIST with 'x' such that the concurrency list is represented
# by a list of numbers delimited by 'x'. This is because of how the underlying launch script
# expects the concurrencies.
bash ./submit_disagg.sh $PREFILL_NODES \
JOB_ID=$(bash ./submit_disagg.sh $PREFILL_NODES \
$PREFILL_NUM_WORKERS \
$DECODE_NODES \
$DECODE_NUM_WORKERS \
$ISL $OSL "${CONC_LIST// /x}" inf \
${PREFILL_ENABLE_EP} ${PREFILL_ENABLE_DP} \
${DECODE_ENABLE_EP} ${DECODE_ENABLE_DP} \
${RANDOM_RANGE_RATIO}
${RANDOM_RANGE_RATIO})

if [[ $? -ne 0 ]]; then
echo "Failed to submit job" >&2
exit 1
fi

echo "$JOB_ID"
14 changes: 12 additions & 2 deletions runners/launch_mi355x-amd.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -38,15 +38,25 @@ if [[ "$IS_MULTINODE" == "true" ]]; then

export MODEL_NAME="DeepSeek-R1"
export MODEL_PATH="/nfsdata"

NODENAME=$(sinfo -N -h -t idle,mix -o "%N" | head -1)
if [[ $NODENAME == GPU* ]]; then
export MODEL_PATH="/nfsdata"
elif [[ $NODENAME == mia1* ]]; then
export MODEL_PATH="/it-share/data"
else
echo "[Error] No available nodes for launching slurm jobs"
exit 1
fi

export ISL="$ISL"
export OSL="$OSL"

sudo rm -rf "$SGL_SLURM_JOBS_PATH/logs" 2>/dev/null || true

bash benchmarks/"${EXP_NAME%%_*}_${PRECISION}_mi355x_${FRAMEWORK}_slurm.sh"
JOB_ID=$(bash benchmarks/"${EXP_NAME%%_*}_${PRECISION}_mi355x_${FRAMEWORK}_slurm.sh")

# Wait for job to complete
JOB_ID=$(squeue -u $USER --noheader --format='%i')
LOG_FILE="$SGL_SLURM_JOBS_PATH/slurm_job-${JOB_ID}.out"

# Give slurm time to start the job and create log file
Expand Down