Closed
Conversation
vadiklyutiy
added a commit
that referenced
this pull request
Dec 19, 2024
Right now `pow` with const exp argument is implemented simply. We convert const to const tensor and run elementwise `pow` of 2 tensors. It is simply but not always efficient. llama2 (RMSNorm part) has `x*x` that implemented as `tensor.pow(2)`. Convert `pow(x,2)` to `x*x`. Improvement on llama2-7B is around **0.237%**
vadiklyutiy
added a commit
that referenced
this pull request
Dec 20, 2024
Right now `pow` with const exp argument is implemented simply. We convert const to const tensor and run elementwise `pow` of 2 tensors. It is simply but not always efficient. llama2 (RMSNorm part) has `x*x` that implemented as `tensor.pow(2)`. Convert `pow(x,2)` to `x*x`. Improvement on llama2-7B is around **0.237%**
vadiklyutiy
added a commit
that referenced
this pull request
Dec 26, 2024
Right now `pow` with const exp argument is implemented simply. We convert const to const tensor and run elementwise `pow` of 2 tensors. It is simply but not always efficient. llama2 (RMSNorm part) has `x*x` that implemented as `tensor.pow(2)`. Convert `pow(x,2)` to `x*x`. Improvement on llama2-7B is around **0.237%**
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This is a special version for the current transpose operator. The current transpose operator will handle a general N-dimension transpose, while this PR implement a 2D version to speed up 2D transpose.
Thread coarsening and (static) shared memory have been used.
Benchmark result: