Skip to content
This repository was archived by the owner on Mar 28, 2023. It is now read-only.

Conversation

@JackAKirk
Copy link

@JackAKirk JackAKirk commented Aug 29, 2022

Updated all tests to use new "unified" interfaces from intel/llvm#7077. The old legacy interface implementation is deprecated but still tested via the _legacy files.

Signed-off-by: JackAKirk jack.kirk@codeplay.com

Signed-off-by: JackAKirk <jack.kirk@codeplay.com>
JackAKirk and others added 2 commits August 30, 2022 17:00
Signed-off-by: JackAKirk <jack.kirk@codeplay.com>
Signed-off-by: JackAKirk <jack.kirk@codeplay.com>
@JackAKirk JackAKirk changed the title [SYCL][CUDA] Accumulator layout specified at load/store. [SYCL][CUDA] Unified matrix interface updated tests Oct 11, 2022
Signed-off-by: JackAKirk <jack.kirk@codeplay.com>
Signed-off-by: JackAKirk <jack.kirk@codeplay.com>
Signed-off-by: JackAKirk <jack.kirk@codeplay.com>
Signed-off-by: JackAKirk <jack.kirk@codeplay.com>
Copy link

@steffenlarsen steffenlarsen left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Just a couple stylistic comments but otherwise it looks good.

JackAKirk and others added 2 commits October 17, 2022 16:51
Co-authored-by: Steffen Larsen <steffen.larsen@intel.com>
Signed-off-by: JackAKirk <jack.kirk@codeplay.com>
@JackAKirk
Copy link
Author

/verify with intel/llvm#7077

Copy link

@steffenlarsen steffenlarsen left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM!

@steffenlarsen steffenlarsen merged commit 2834fab into intel:intel Oct 18, 2022
steffenlarsen pushed a commit to intel/llvm that referenced this pull request Dec 12, 2022
…ce (#7077)

CUDA backend implementation using the "unified" matrix extension
interface. The same interface will be used for a future Intel backend
implementation of the matrix extension.

- New "unified" interface uses SYCL_EXT_ONEAPI_MATRIX_VERSION=4
- `joint_matrix_load`, `joint_matrix_store`, `joint_matrix_mad` and
`joint_matrix` interfaces match the new spec from
#6662
- Separated `joint_matrix_*` functions into new header
matrix-unified.hpp: Intel backend implementations can be called from the
same functions in the future.
- C++17 everywhere in line with #6678
- Updated device code tests to use new interfaces
- Completely removed uint16 implementations that are replaced by
bfloat16 that is being moved out of the experimental namespace
- Updated all CUDA runtime matrix tests here:
intel/llvm-test-suite#1183

Signed-off-by: JackAKirk <jack.kirk@codeplay.com>
myler pushed a commit to myler/llvm-test-suite that referenced this pull request Mar 22, 2023
Updated all tests to use new "unified" interfaces from intel/llvm#7077. The old legacy interface implementation is deprecated but still tested via the _legacy files.

Signed-off-by: JackAKirk <jack.kirk@codeplay.com>
aelovikov-intel pushed a commit to aelovikov-intel/llvm that referenced this pull request Mar 27, 2023
…suite#1183)

Updated all tests to use new "unified" interfaces from intel#7077. The old legacy interface implementation is deprecated but still tested via the _legacy files.

Signed-off-by: JackAKirk <jack.kirk@codeplay.com>
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants