feat(gpu): implement PBSs that use new Hopper's features #987
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
pdroalves
force-pushed
the
feat/gpu/tbc
branch
11 times, most recently
from
e9813dc to
4c6bb0f
Compare
pdroalves
force-pushed
the
feat/gpu/tbc
branch
2 times, most recently
from
3b6c448 to
268d2f0
Compare
agnesLeroy
reviewed
Apr 19, 2024
There was a problem hiding this comment.
Hi @pdroalves! Thanks a lot for this! I haven't looked in details at the tbc PBS code itself, would be way too long for me to do so, but here's a my review on the rest.
backends/tfhe-cuda-backend/cuda/include/programmable_bootstrap_multibit.h
Show resolved
Hide resolved
backends/tfhe-cuda-backend/cuda/src/pbs/programmable_bootstrap_cg_classic.cuh
Show resolved
Hide resolved
backends/tfhe-cuda-backend/cuda/src/pbs/programmable_bootstrap_classic.cu
Show resolved
Hide resolved
backends/tfhe-cuda-backend/cuda/src/pbs/programmable_bootstrap_classic.cu
Show resolved
Hide resolved
backends/tfhe-cuda-backend/cuda/src/pbs/programmable_bootstrap_classic.cu
Show resolved
Hide resolved
backends/tfhe-cuda-backend/cuda/src/pbs/programmable_bootstrap_classic.cu
Outdated
Show resolved
Hide resolved
backends/tfhe-cuda-backend/cuda/src/pbs/programmable_bootstrap_classic.cu
Outdated
Show resolved
Hide resolved
backends/tfhe-cuda-backend/cuda/src/pbs/programmable_bootstrap_multibit.cu
Outdated
Show resolved
Hide resolved
|
Since TBC PBS will only run (and be compiled) on H100s, I rebased this branch on top of |
|
@slab-ci gpu_test |
|
@slab-ci integer_gpu_full_bench |
|
@agnesLeroy Looks like this breaks the CI. Should I undo this rebase? |
|
@pdroalves you should target main instead of David's branch, let's run the H100 tests manually instead of in the CI exceptionally. |
pdroalves
changed the base branch from
main
to
chre/gpu/integer_mult_gpu_params
|
Indeed. I rebased on |
pdroalves
force-pushed
the
feat/gpu/tbc
branch
3 times, most recently
from
601baaf to
e59909c
Compare
pdroalves
force-pushed
the
chre/gpu/integer_mult_gpu_params
branch
4 times, most recently
from
b34e148 to
c757fb3
Compare
agnesLeroy
force-pushed
the
chre/gpu/integer_mult_gpu_params
branch
from
c757fb3 to
60f00d7
Compare
pdroalves
force-pushed
the
chre/gpu/integer_mult_gpu_params
branch
from
60f00d7 to
f08da6e
Compare
agnesLeroy
force-pushed
the
chre/gpu/integer_mult_gpu_params
branch
from
f08da6e to
c80d5fc
Compare
cluster and distributed shared memory
agnesLeroy
force-pushed
the
feat/gpu/tbc
branch
from
e59909c to
20e11ea
Compare
…cluster and distributed shared memory
|
@agnesLeroy This branch passes integer tests on H100. |
agnesLeroy
merged commit 7c10dae
into
chre/gpu/integer_mult_gpu_params
13 checks passed
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.
https://github.com/zama-ai/tfhe-rs-internal/issues/450
https://github.com/zama-ai/concrete-internal/issues/345
closes: please link all relevant issues
PR content/description
Check-list: