Add complex group algorithms #7120
Merged
Conversation
This file contains hidden or 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
Naghasan
reviewed
Oct 19, 2022
asudarsa
reviewed
Oct 19, 2022
| @@ -46,4 +46,16 @@ enum GroupOperation { | |||
| ExclusiveScan = 2, | |||
| }; | |||
|
|
|||
| typedef struct { | |||
There was a problem hiding this comment.
Would it make sense to add bfloat16 here? It is still 'experimental', but will be moved out soon. We can wait for that to happen. Thanks
There was a problem hiding this comment.
This has already got rather long in the tooth, so I'd like to defer that change for later
ldrumm
force-pushed
the
complex_group_algo
branch
3 times, most recently
from
6c674fa to
29279e0
Compare
romanovvlad
reviewed
Oct 24, 2022
|
Thanks for the reviews and apologies for the silence. I left this while dependent patches were merged upstream. |
ldrumm
force-pushed
the
complex_group_algo
branch
from
74007dd to
f7e1823
Compare
ldrumm
commented
Mar 10, 2023
| @@ -46,4 +46,16 @@ enum GroupOperation { | |||
| ExclusiveScan = 2, | |||
| }; | |||
|
|
|||
| typedef struct { | |||
There was a problem hiding this comment.
This has already got rather long in the tooth, so I'd like to defer that change for later
ldrumm
force-pushed
the
complex_group_algo
branch
from
f7e1823 to
f38b675
Compare
ldrumm
force-pushed
the
complex_group_algo
branch
from
f38b675 to
2d88880
Compare
Extends the implementation of group algorithms by adding support for complex types. Operations that can be done element-wise are implemented in headers and multiplication is implemented in CUDA PI. This involved changing the macros in CUDA PI to accommodate complex and defining complex types to use in PI interface. The scratch space for moving data between workgroups in the PI also had to be doubled.
ldrumm
force-pushed
the
complex_group_algo
branch
from
2d88880 to
59a3806
Compare
romanovvlad
requested review from
sergey-semenov
and removed request for
romanovvlad
Naghasan
approved these changes
Mar 22, 2023
|
friendly ping @sergey-semenov @asudarsa @intel/llvm-gatekeepers Could we get some help/reviews on this PR, it's been open for a long time. |
sergey-semenov
approved these changes
Mar 24, 2023
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.
Add complex group algorithms
Extends the implementation of group algorithms by adding support for
complex types. Operations that can be done element-wise are implemented
in headers and multiplication is implemented in CUDA PI. This involved
changing the macros in CUDA PI to accommodate complex and defining
complex types to use in PI interface. The scratch space for moving data
between workgroups in the PI also had to be doubled.
This patch depends on two fixes to the NVPTX backend which are exposed by the new code generation.
Tests are in intel/llvm-test-suitegit