[V0.9.1] optimize rope in qwen3 #1719
Merged
+292
−41
Conversation
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
David9857
force-pushed
the
pr-rope
branch
2 times, most recently
from
1874f93 to
efc563b
Compare
|
Please add unittest for this rope impl |
Signed-off-by: David9857 <985700846@qq.com> use npu_apply_rotary_pos_emb when head_size is 128 and is noex_style Signed-off-by: David9857 <985700846@qq.com>
David9857
force-pushed
the
pr-rope
branch
2 times, most recently
from
5b30648 to
e568126
Compare
Signed-off-by: David9857 <985700846@qq.com>
added |
ganyi1996ppo
approved these changes
Jul 14, 2025
David9857
added a commit
to rjg-lyh/vllm-ascend
that referenced
this pull request
Jul 14, 2025
use npu_apply_rotary_pos_emb when head_size is 128 and is noex_style ### What this PR does / why we need it? Optimize rope by extracting index_select from layers into model, which can reduce (layer_num -1) * 2 Gather ops in each prefill/decode stage. ### Does this PR introduce _any_ user-facing change? NA ### How was this patch tested? NA --------- Signed-off-by: David9857 <985700846@qq.com>
David9857
added a commit
to rjg-lyh/vllm-ascend
that referenced
this pull request
Jul 14, 2025
use npu_apply_rotary_pos_emb when head_size is 128 and is noex_style ### What this PR does / why we need it? Optimize rope by extracting index_select from layers into model, which can reduce (layer_num -1) * 2 Gather ops in each prefill/decode stage. ### Does this PR introduce _any_ user-facing change? NA ### How was this patch tested? NA --------- Signed-off-by: David9857 <985700846@qq.com>
David9857
added a commit
to rjg-lyh/vllm-ascend
that referenced
this pull request
Jul 14, 2025
use npu_apply_rotary_pos_emb when head_size is 128 and is noex_style ### What this PR does / why we need it? Optimize rope by extracting index_select from layers into model, which can reduce (layer_num -1) * 2 Gather ops in each prefill/decode stage. ### Does this PR introduce _any_ user-facing change? NA ### How was this patch tested? NA --------- Signed-off-by: David9857 <985700846@qq.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
use npu_apply_rotary_pos_emb when head_size is 128 and is noex_style
What this PR does / why we need it?
Optimize rope by extracting index_select from layers into model, which can reduce (layer_num -1) * 2 Gather ops in each prefill/decode stage.
Does this PR introduce any user-facing change?
NA
How was this patch tested?
NA