Skip to content

[DeepSeek][ROCm] V4.1-Flash: enable sparse-MLA Gluon kernel + MXFP4 sparse indexer on MI355X #1049

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
Fangzhou-Ai merged 1 commit into from
Sep 30, 2026
Merged

[DeepSeek][ROCm] V4.1-Flash: enable sparse-MLA Gluon kernel + MXFP4 sparse indexer on MI355X #1049

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
13 changes: 12 additions & 1 deletion models/deepseek-ai/DeepSeek-V4.1-Flash.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,7 +5,7 @@ meta:
provider: "DeepSeek"
description: "DeepSeek V4.1 Flash vision-language MoE (552B backbone; 8B active per prompt token, 16B per output token) combining sliding-window plus compressed sparse attention with a two-level indexer, engram n-gram memory, hyper-connections, and a DSpark multi-token draft head."
date_added: 2026-09-09
date_updated: 2026-09-22
date_updated: 2026-09-29
difficulty: advanced
tasks:
- text
Expand Down Expand Up @@ -252,6 +252,17 @@ strategy_overrides:
extra_args:
- "--engram-config"
- '{"cpu_offload":true}'
# ROCm paged MXFP4 sparse indexer (vllm-project/vllm#58671).
- "--attention-config"
- '{"indexer_kv_dtype":"mxfp4","indexer_sparse_logits":true}'
# The fp4 indexer prefers block-size 128; auto-resolution
# otherwise falls back to 64.
- "--block-size"
- "128"
extra_env:
# Aiter's Gluon sparse-MLA kernel, gfx950 only
# (vllm-project/vllm#53492).
VLLM_ROCM_USE_AITER_TRITON_SPARSE_MLA: "1"
Comment on lines +262 to +265

@gemini-code-assist gemini-code-assist Bot Sep 29, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

medium

The environment variable VLLM_ROCM_USE_AITER_TRITON_SPARSE_MLA: "1" is a hardware-specific optimization for gfx950 (MI355X) and is independent of the serving strategy. Currently, it is defined under strategy_overrides.single_node_tp.hardware_overrides.mi355x, which means it will only be applied when using the single_node_tp strategy. To ensure this optimization is applied across all compatible strategies (such as multi_node_tp or pd_cluster) when running on mi355x, it should be moved to the top-level hardware_overrides.mi355x block instead.

blackwell:
extra_args:
- "--attention-config"
Expand Down