Code object compression via bundling #2054
Draft
+110
−51
Conversation
This file contains 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
bstefanuk
requested review from
babakpst,
yoichiyoshida,
bragadeesh and
AlexBrownAMD
as code owners
|
Once documentation and tests are added, I will remove from draft. |
ellosel
reviewed
Nov 21, 2024
|
|
||
| extObj = ".o" | ||
| extCo = ".co" | ||
| extCoRaw = ".co.raw" |
There was a problem hiding this comment.
what is the meaning of raw?
There was a problem hiding this comment.
Here, ".raw" is an extension used to identify the co file after linking but before compression. Files with this extension live in the build_tmp/ directory and are the input for the compression via bundling step.
ellosel
reviewed
Nov 21, 2024
| bundler, | ||
| "--compress", | ||
| "--type=o", | ||
| "--bundle-align=4096", |
There was a problem hiding this comment.
Where did we get 4096 and should it be fixed?
There was a problem hiding this comment.
This parameter was recommended to me by the compiler team. @yxsamliu how does this parameter affect the output? Is it needed in this case?
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Outdated suggestions cannot be applied.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.
Summary:
This PR adds a compression layer to all final code objects, thereby generating smaller libraries at the expense of build time. Includes minor refactoring.
Outcomes:
getAssemblyCodeObjectFileshas been renamed tobuildAssemblyCodeObjectFilesto match the name of source kernel functions.Compression ratio for gfx900: 2.66
Compression ratio for gfx90a: 3.74
Testing and Environment:
Docker: Ubuntu 24.04, ROCm 6.4 RC stack, AMD clang version 18.0.0, AMD clang-offload-bundler version 18.0.0
Tested with hipBLASLt bench