NVRTC kernels for cast-transpose #258
Merged
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
Signed-off-by: Tim Moon <tmoon@nvidia.com>
timmoon10
force-pushed
the
nvrtc-cast-transpose
branch
from
dfc121b to
aed892a
Compare
|
/te-ci |
|
/te-ci |
|
/te-ci |
|
/te-ci |
|
/te-ci |
|
/te-ci |
|
/te-ci |
Signed-off-by: Tim Moon <tmoon@nvidia.com>
|
/te-ci |
|
/te-ci |
|
/te-ci |
|
/te-ci |
|
If there are no objections, I will merge this after the 1.6 release. |
Signed-off-by: Tim Moon <tmoon@nvidia.com>
Signed-off-by: Tim Moon <tmoon@nvidia.com>
|
/te-ci |
|
/te-ci |
| const size_t num_tiles) { | ||
| __launch_bounds__(block_size) | ||
| cast_transpose_general_kernel(const IType * __restrict__ const input, | ||
| const CType * const noop, |
There was a problem hiding this comment.
Was the restrict keyword intentionally omitted here?
Oleg-Goncharov
approved these changes
Apr 19, 2024
timmoon10
commented
Apr 19, 2024
Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
|
/te-ci |
pggPL
pushed a commit
to pggPL/TransformerEngine
that referenced
this pull request
May 15, 2024
* Add NVRTC kernels for cast-transpose Signed-off-by: Tim Moon <tmoon@nvidia.com> * Update copyright year Signed-off-by: Tim Moon <tmoon@nvidia.com> * Add noop flag to NVRTC cast-transpose kernel Signed-off-by: Tim Moon <tmoon@nvidia.com> * Apply suggestions from code review Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com> --------- Signed-off-by: Tim Moon <tmoon@nvidia.com> Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com> Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.com>
pggPL
pushed a commit
to pggPL/TransformerEngine
that referenced
this pull request
May 16, 2024
* Add NVRTC kernels for cast-transpose Signed-off-by: Tim Moon <tmoon@nvidia.com> * Update copyright year Signed-off-by: Tim Moon <tmoon@nvidia.com> * Add noop flag to NVRTC cast-transpose kernel Signed-off-by: Tim Moon <tmoon@nvidia.com> * Apply suggestions from code review Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com> --------- Signed-off-by: Tim Moon <tmoon@nvidia.com> Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com> Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.com>
pggPL
pushed a commit
to pggPL/TransformerEngine
that referenced
this pull request
May 23, 2024
* Add NVRTC kernels for cast-transpose Signed-off-by: Tim Moon <tmoon@nvidia.com> * Update copyright year Signed-off-by: Tim Moon <tmoon@nvidia.com> * Add noop flag to NVRTC cast-transpose kernel Signed-off-by: Tim Moon <tmoon@nvidia.com> * Apply suggestions from code review Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com> --------- Signed-off-by: Tim Moon <tmoon@nvidia.com> Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com> Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.com>
13 tasks
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.
This is a follow-up to #138 that adds NVRTC kernels for cast-transpose.
I've also tweaked the kernel selection heuristics to be a bit more general. There are some performance differences at intermediate scale, but nothing too drastic:
Transpose kernel times
Evaluated FP8 transposes on an H100 PCIe with square matrices with nice dims (power of two or halfway between powers of two).