[Performance] Improving speed by splitting kernels#97
Merged
Conversation
…rder to make computation faster. If the processor is powerful enough, these two parts can be executed in parallel by out-of-order execution. However, there is some overhead that is required to split the original kernel into two. The degree of speed-up depends on the micro architecture, and the split kernels may be slower on some micro architectures. This feature is enabled by SPLIT_KERNEL macro, and currently, it is enabled only for AVX2 and AVX512F. This change affects computation speed of many functions, but for example, the ratio of execution time between the previous version and the new version of asind4_u35 is 2.08.
fpetrogalli
approved these changes
Oct 31, 2017
fpetrogalli
left a comment
fpetrogalli
left a comment
There was a problem hiding this comment.
Please cherry pick this change into cmake-transition.
Please squash the commits into a single commit.
shibatch
added a commit
that referenced
this pull request
Nov 1, 2017
With this patch, kernels can be splitted into odd and even parts in order to make computation faster. If the processor is powerful enough, these two parts can be executed in parallel by out-of-order execution. However, there is some overhead that is required to split the original kernel into two. The degree of speed-up depends on the micro architecture, and the split kernels may be slower on some micro architectures. This feature is enabled by SPLIT_KERNEL macro, and currently, it is enabled only for AVX2 and AVX512F. This change affects computation speed of many functions, but for example, the ratio of execution time between the previous version and the new version of asind4_u35 is 2.08.
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.
With this patch, kernels can be splitted into odd and even parts in order to make computation faster.
If the processor is powerful enough, these two parts can be executed in parallel by out-of-order execution.
However, there is some overhead that is required to split the original kernel into two.
The degree of speed-up depends on the micro architecture. It may be slower on some micro architectures.
This feature is enabled by SPLIT_KERNEL macro, and currently, it is enabled only for AVX2 and AVX512F.
This change affects computation speed of many functions, but for example,
the ratio of execution time between the previous version and the new version of asind4_u35 is 2.08.