Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Rahul/semi structured #39

Draft
wants to merge 69 commits into
base: main
Choose a base branch
from
Draft

Rahul/semi structured #39

wants to merge 69 commits into from

Conversation

dsikka
Copy link

@dsikka dsikka commented Dec 7, 2024 •
edited by github-actions bot
Loading

No description provided.

Faraz9877 and others added 30 commits October 22, 2024 15:49
@Faraz9877
Add cutlass 2:4 infrastructure
5d51361
@Faraz9877
Update with test code
17f5b96
@Faraz9877
Clean up a bit; both fp8 and int8 working
471a03c
@Faraz9877
Add fp16 and bf16 support to sparse cutlass mm
0b332fb
@ilmarkov
semi_structured for fp16 and bf16 and int8
da31648
@ilmarkov
Fix A100 int8 tests
e655f94
@ilmarkov
Add fp8 cusparseLt
5fc3c1c
@ilmarkov
wip
9cf36d6
@ilmarkov
Fix signatures
ad09e79
@ilmarkov
Fix compilation and tests
e75eabc
@ilmarkov
Update for older platforms
0306390
@ilmarkov
Add benchmarks
1021acb
@ilmarkov
Fix typo
19ce358
@ilmarkov
Added scaled_mm for fp8.
959408c 
Removed cmake check for cusparseLt, needs to be reverted when the cmake issue is resolved.
@ilmarkov
Add docstrings
117b87b
@ilmarkov
Update for torch 2.5
2c7e68e
@ilmarkov
Add handling contiguous dense input for int8 and fp8
922f4f8
@ilmarkov
Add fp8 cusparseLt
beca038
@ilmarkov
Fix compilation and tests
5d9cd25
@ilmarkov
Add caching of cusparseLT meta
39ad9d4
@ilmarkov
Cached cusparseLt
520eb62
@ilmarkov
Fix destroy function
20956e6
@ilmarkov
Prepare for reproduce
87c8088
@ilmarkov
Fix cusparseLt caching
4ea58b1
@ilmarkov
Make cached version default function
f0551ef
@ilmarkov
Fixes and polishing after rebase
d7476e8
@dsikka @rahul-tuli
add sparse 2:4 weight loading suport
681ea5e
@rahul-tuli
Some more changes!
ecf878f
@rahul-tuli
Cleanup
80952dc
@dsikka @rahul-tuli
get uncompressed to work; update gemm to use contiguous; use alex's u…
8462c9d 
…tils instead of our decompressor
Faraz9877 and others added 29 commits November 14, 2024 20:36
@Faraz9877
Clean up the benchmarking
13fccf4
@Faraz9877
Clean up the cutlass benchmarking
b345cc8
@Faraz9877
Fix cmake errors
2d03e1d
@Faraz9877
Fix the cmake TAG
e9439cc
@robertgshaw2-neuralmagic
Merge branch 'buildable' into rahul-quant-merged-rs
4ba7c0f
@robertgshaw2-neuralmagic
update
f74ef37
@robertgshaw2-neuralmagic
fixed
f5bc9eb
@robertgshaw2-neuralmagic
updated
1316076
@robertgshaw2-neuralmagic
updated
4d2b12c
@robertgshaw2-neuralmagic
updated, calling things properly
fe30b53
@robertgshaw2-neuralmagic
running end to end but not passing
4c61b19
@robertgshaw2-neuralmagic
updated
86716f8
@robertgshaw2-neuralmagic
stash
349d904
@robertgshaw2-neuralmagic
update
b860f9e
@Faraz9877 @robertgshaw2-neuralmagic
Fix batch size and zeros issue
df462b5
@robertgshaw2-neuralmagic
working e2e
14f6141
@Faraz9877 @robertgshaw2-neuralmagic
Enable other datatypes
abfd85d
@Faraz9877 @robertgshaw2-neuralmagic
Add the heuristics for fp8
540d0ce
@robertgshaw2-neuralmagic
updated
b45c158
@rahul-tuli
Cleanup
5f7339e
@rahul-tuli
Move model compressor to scheme
ed15777
@rahul-tuli
Cleanup + updates for compressed_tensor changes
34a84a4
@Faraz9877 @robertgshaw2-neuralmagic
Add cherry-picked heuristic for Llama3 8B model
4045bda
@robertgshaw2-neuralmagic
updated with latest kernel
7c61ab0
@dsikka
Merge pull request #35 from neuralmagic/rob/semi-structured
a8a1b57 
[ 2:4 Sparse ] Cherry Pick Latest Kernel From Faraz
@dsikka
remove compressed support; validate against ct models for tp=1,24
4e060df
@dsikka
add testing cases
7a6d027
@dsikka
add support for all cases; update tests
0987c98
@dsikka
Merge pull request #37 from neuralmagic/dipika/sem-struc-uncomp
c7d1cc3 
Update 2:4 Support
@github-actions GitHub Actions
Copy link

github-actions bot commented Dec 7, 2024

👋 Hi! Thank you for contributing to the vLLM project.
Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can do one of these:

  • Add ready label to the PR
  • Enable auto-merge.

🚀

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@mgoin mgoin Awaiting requested review from mgoin mgoin will be requested when the pull request is marked ready for review mgoin is a code owner

@youkaichao youkaichao Awaiting requested review from youkaichao youkaichao will be requested when the pull request is marked ready for review youkaichao is a code owner

@robertgshaw2-neuralmagic robertgshaw2-neuralmagic Awaiting requested review from robertgshaw2-neuralmagic robertgshaw2-neuralmagic will be requested when the pull request is marked ready for review robertgshaw2-neuralmagic is a code owner

@tlrmchlsmth tlrmchlsmth Awaiting requested review from tlrmchlsmth tlrmchlsmth will be requested when the pull request is marked ready for review tlrmchlsmth is a code owner

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
5 participants
@dsikka @Faraz9877 @ilmarkov @rahul-tuli @robertgshaw2-neuralmagic