Adjust MiniMax MI355X block size for TP8 EP8

[jiacao-amd]

Summary

• default MiniMax MI355X vLLM runs to block size 16 with shuffled KV cache layout enabled
• special-case TP8/EP8 to disable shuffled KV cache layout and use block size 32

Testing

• bash -n benchmarks/single_node/minimaxm2.5_fp8_mi355x.sh

[claude]

Claude Code Review

This pull request is from a fork — automated review is disabled. A repository maintainer can comment @claude review to run a one-time review.

[github-actions]

@jiacao-amd Kicking off a sweep.

Run: https://github.com/SemiAnalysisAI/InferenceX/actions/runs/25121853688
Command: test-config --config-files .github/configs/amd-master.yaml --config-keys minimaxm2.5-fp8-mi355x-vllm
Pinned ref: d66409b
Approval: not required (trusted collaborator).

[jiacao-amd]

/sweep test-config --config-files .github/configs/amd-master.yaml --config-keys minimaxm2.5-fp8-mi355x-vllm

[github-actions]

@jiacao-amd Kicking off a sweep.

Run: https://github.com/SemiAnalysisAI/InferenceX/actions/runs/25330781481
Command: test-config --config-files .github/configs/amd-master.yaml --config-keys minimaxm2.5-fp8-mi355x-vllm
Pinned ref: a34fb25
Approval: not required (trusted collaborator).

[jiacao-amd]

/sweep test-config --config-files .github/configs/amd-master.yaml --config-keys minimaxm2.5-fp8-mi355x-vllm

[github-actions]

@jiacao-amd Kicking off a sweep.

Run: https://github.com/SemiAnalysisAI/InferenceX/actions/runs/25336346089
Command: test-config --config-files .github/configs/amd-master.yaml --config-keys minimaxm2.5-fp8-mi355x-vllm
Pinned ref: 37962d4
Approval: not required (trusted collaborator).

[jiacao-amd]

Superseded by #1276. The replacement PR uses the same MiniMax MI355X vLLM scheduling change, but the branch is pushed directly to SemiAnalysisAI/InferenceX instead of the fork so CI/automation should avoid the fork-PR permission issues.