[Bugfix] Reduce _npu_flash_attention mask to 128x128 for memory savings#1100
Closed
ApsarasX wants to merge 1 commit intovllm-project:mainfrom
Closed
[Bugfix] Reduce _npu_flash_attention mask to 128x128 for memory savings#1100ApsarasX wants to merge 1 commit intovllm-project:mainfrom
ApsarasX wants to merge 1 commit intovllm-project:mainfrom
Conversation
Merged
wangxiyuan
pushed a commit
that referenced
this pull request
Jun 11, 2025
### What this PR does / why we need it? Single machine 16 cards deepseekr1 attention (tp8/dp2) / moe(etp) Best performance rely on: vllm-ascend commit id:da9acfca6053352730fce75fb772e214755d0341 vllm commit id:b124e1085b1bf977e3dac96d99ffd9d8ddfdb6cc + #910 + [Reduce _npu_flash_attention mask to 128x128 for memory savings] #1100 [Reduce memory usage by splitting tokens in fused_experts] --------- Signed-off-by: ttanzhiqiang <389825161@qq.com>
momo609
pushed a commit
to momo609/vllm-ascend
that referenced
this pull request
Jun 17, 2025
### What this PR does / why we need it? Single machine 16 cards deepseekr1 attention (tp8/dp2) / moe(etp) Best performance rely on: vllm-ascend commit id:da9acfca6053352730fce75fb772e214755d0341 vllm commit id:b124e1085b1bf977e3dac96d99ffd9d8ddfdb6cc + vllm-project#910 + [Reduce _npu_flash_attention mask to 128x128 for memory savings] vllm-project#1100 [Reduce memory usage by splitting tokens in fused_experts] --------- Signed-off-by: ttanzhiqiang <389825161@qq.com>
momo609
pushed a commit
to momo609/vllm-ascend
that referenced
this pull request
Jun 17, 2025
### What this PR does / why we need it? Single machine 16 cards deepseekr1 attention (tp8/dp2) / moe(etp) Best performance rely on: vllm-ascend commit id:da9acfca6053352730fce75fb772e214755d0341 vllm commit id:b124e1085b1bf977e3dac96d99ffd9d8ddfdb6cc + vllm-project#910 + [Reduce _npu_flash_attention mask to 128x128 for memory savings] vllm-project#1100 [Reduce memory usage by splitting tokens in fused_experts] --------- Signed-off-by: ttanzhiqiang <389825161@qq.com> Signed-off-by: wangxiaoxin (A) <wangxiaoxin7@huawei.com>
momo609
pushed a commit
to momo609/vllm-ascend
that referenced
this pull request
Jun 17, 2025
### What this PR does / why we need it? Single machine 16 cards deepseekr1 attention (tp8/dp2) / moe(etp) Best performance rely on: vllm-ascend commit id:da9acfca6053352730fce75fb772e214755d0341 vllm commit id:b124e1085b1bf977e3dac96d99ffd9d8ddfdb6cc + vllm-project#910 + [Reduce _npu_flash_attention mask to 128x128 for memory savings] vllm-project#1100 [Reduce memory usage by splitting tokens in fused_experts] --------- Signed-off-by: ttanzhiqiang <389825161@qq.com> Signed-off-by: wangxiaoxin (A) <wangxiaoxin7@huawei.com>
Contributor
|
This pull request has conflicts, please resolve those before we can evaluate the pull request. |
shiyuan680
pushed a commit
to raindaywhu/vllm-ascend
that referenced
this pull request
Jul 7, 2025
### What this PR does / why we need it? Single machine 16 cards deepseekr1 attention (tp8/dp2) / moe(etp) Best performance rely on: vllm-ascend commit id:da9acfca6053352730fce75fb772e214755d0341 vllm commit id:b124e1085b1bf977e3dac96d99ffd9d8ddfdb6cc + vllm-project#910 + [Reduce _npu_flash_attention mask to 128x128 for memory savings] vllm-project#1100 [Reduce memory usage by splitting tokens in fused_experts] --------- Signed-off-by: ttanzhiqiang <389825161@qq.com>
Member
|
Is this PR still needed? If so, please rebase, otherwise please close it. Thanks |
Signed-off-by: ApsarasX <apsarax@outlook.com>
ApsarasX
force-pushed
the
community-attn_mask
branch
from
d951107 to
df4cee4
Compare
Collaborator
Author
This PR is still necessary once torch_npu._npu_flash_attention supports compressed 128x128 mask I have already rebased this PR |
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #1100 +/- ##
=======================================
Coverage 76.31% 76.31%
=======================================
Files 116 116
Lines 13238 13238
=======================================
Hits 10102 10102
Misses 3136 3136
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
Contributor
|
This pull request has conflicts, please resolve those before we can evaluate the pull request. |
chopper0126
pushed a commit
to chopper0126/vllm-ascend
that referenced
this pull request
Oct 16, 2025
### What this PR does / why we need it? Single machine 16 cards deepseekr1 attention (tp8/dp2) / moe(etp) Best performance rely on: vllm-ascend commit id:da9acfca6053352730fce75fb772e214755d0341 vllm commit id:b124e1085b1bf977e3dac96d99ffd9d8ddfdb6cc + vllm-project#910 + [Reduce _npu_flash_attention mask to 128x128 for memory savings] vllm-project#1100 [Reduce memory usage by splitting tokens in fused_experts] --------- Signed-off-by: ttanzhiqiang <389825161@qq.com>
Angazenn
pushed a commit
to Angazenn/vllm-ascend
that referenced
this pull request
Oct 21, 2025
### What this PR does / why we need it? Single machine 16 cards deepseekr1 attention (tp8/dp2) / moe(etp) Best performance rely on: vllm-ascend commit id:da9acfca6053352730fce75fb772e214755d0341 vllm commit id:b124e1085b1bf977e3dac96d99ffd9d8ddfdb6cc + vllm-project#910 + [Reduce _npu_flash_attention mask to 128x128 for memory savings] vllm-project#1100 [Reduce memory usage by splitting tokens in fused_experts] --------- Signed-off-by: ttanzhiqiang <389825161@qq.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What this PR does / why we need it?
Avoid generating a special
attn_mask(which theoretically could be as large asbf16[max_model_len, max_model_len]) whenchunked-prefillis not enabled and the user input is extremely long.However, this PR depends on modifications to the
torch_npu._npu_flash_attentionDoes this PR introduce any user-facing change?
No
How was this patch tested?
No