Skip to content

[AMD] Add Minimax tp8 with ep and update vllm image for MI355x #927

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
chunfangamd merged 9 commits into from
Mar 23, 2026
Merged

[AMD] Add Minimax tp8 with ep and update vllm image for MI355x #927

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
1ca7a2b
add minimax tp8 with ep and remove tp-4
benenzhu Mar 5, 2026
4edc25b
update changelog
benenzhu Mar 5, 2026
8ae8114
Update amd-master.yaml for 1k8k & 8k1k CONC to 256
benenzhu Mar 11, 2026
1f108c5
Merge remote-tracking branch 'upstream/main' into minimax-mi355-opt
benenzhu Mar 21, 2026
26c5c5f
fix
benenzhu Mar 21, 2026
d6e9723
Update perf-changelog.yaml
benenzhu Mar 23, 2026
c35e700
Update minimaxm2.5_fp8_mi355x.sh to remove log requests
benenzhu Mar 23, 2026
9d2248d
Update perf-changelog.yaml
benenzhu Mar 23, 2026
db03dc0
Merge branch 'main' into minimax-mi355-opt
cquil11 Mar 23, 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
5 changes: 4 additions & 1 deletion .github/configs/amd-master.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -360,7 +360,7 @@ kimik2.5-fp4-mi355x-vllm:
- { tp: 8, conc-start: 4, conc-end: 64 }

minimaxm2.5-fp8-mi355x-vllm:
image: vllm/vllm-openai-rocm:v0.15.1
image: vllm/vllm-openai-rocm:v0.18.0
model: MiniMaxAI/MiniMax-M2.5
model-prefix: minimaxm2.5
runner: mi355x
Expand All @@ -373,16 +373,19 @@ minimaxm2.5-fp8-mi355x-vllm:
search-space:
- { tp: 2, conc-start: 4, conc-end: 64 }
- { tp: 4, conc-start: 4, conc-end: 64 }
- { tp: 8, ep: 8, conc-start: 32, conc-end: 256 }
- isl: 1024
osl: 8192
search-space:
- { tp: 2, conc-start: 4, conc-end: 64 }
- { tp: 4, conc-start: 4, conc-end: 64 }
- { tp: 8, ep: 8, conc-start: 32, conc-end: 256 }
- isl: 8192
osl: 1024
search-space:
- { tp: 2, conc-start: 4, conc-end: 64 }
- { tp: 4, conc-start: 4, conc-end: 64 }
- { tp: 8, ep: 8, conc-start: 32, conc-end: 256 }

minimaxm2.5-fp8-mi300x-vllm:
image: vllm/vllm-openai-rocm:v0.16.0
Expand Down
9 changes: 8 additions & 1 deletion benchmarks/single_node/minimaxm2.5_fp8_mi355x.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,7 @@ source "$(dirname "$0")/../benchmark_lib.sh"
check_env_vars \
MODEL \
TP \
EP_SIZE \
CONC \
ISL \
OSL \
Expand All @@ -28,16 +29,22 @@ export VLLM_ROCM_USE_AITER=1
SERVER_LOG=/workspace/server.log
PORT=${PORT:-8888}

if [ "$EP_SIZE" -gt 1 ]; then
EP=" --enable-expert-parallel"
else
EP=" "
fi

# Start GPU monitoring (power, temperature, clocks every second)
start_gpu_monitor

set -x
vllm serve $MODEL --port $PORT \
--tensor-parallel-size=$TP \
$EP \
--gpu-memory-utilization 0.95 \
--max-model-len $MAX_MODEL_LEN \
--block-size=32 \
--disable-log-requests \
--trust-remote-code > $SERVER_LOG 2>&1 &

SERVER_PID=$!
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 @@ -1007,3 +1007,11 @@
- "EAGLE speculative decoding: num-steps 3, draft-tokens 4, topk 1"
- "New script: benchmarks/single_node/qwen3.5_fp8_b200_mtp.sh"
pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/898

- config-keys:
- minimaxm2.5-fp8-mi355x-vllm
description:
- "ADD minimax TP=8 with EP, in config of 1k1k, 1k8k, and 8k1k sequence lengths"
- "Config concurrency: 32-256"
- "update image to vllm/vllm-openai-rocm:v0.18.0"
pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/927