-
Notifications
You must be signed in to change notification settings - Fork 8
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.
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
Use the RoCM/HIP device to accelerate certain DPLASMA kernels #57
base: master
Are you sure you want to change the base?
Conversation
012b53b
to
d8a604f
Compare
As discussed on 03/31/23 we need to rebase and check the result. This will be tested next week on Frontier, we need it to be ready. |
5deec12
to
fae375c
Compare
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]> Conflicts: src/CMakeLists.txt
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Use proper error checks instead of asserts
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
please squash to fewer commits. |
Signed-off-by: Aurelien Bouteiller <[email protected]>
Signed-off-by: Aurelien Bouteiller <[email protected]>
This is in ready to merge state beside the 'squash to less commits'. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why only certain kernels are HIP-enabled instead of having all CUDA-enabled kernel also be HIP-enabled ?
@@ -109,8 +109,7 @@ extern void *dplasma_pcomm; | |||
#define dplasma_error(__func, __msg) do { fprintf(stderr, "%s: %s\n", (__func), (__msg)); } while(0) | |||
#endif /* defined(DPLASMA_DEBUG) */ | |||
|
|||
#if defined(DPLASMA_HAVE_CUDA) | |||
#include "dplasmaaux_cuda.h" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why are these headers not protected ?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
protection is self contained in the header itself
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We had proper protection in place avoiding to include unnecessary files. What justify the need to made this change ?
stage_in=stage_in_lapack | ||
stage_out=stage_out_lapack] | ||
stage_in=dplasma_cuda_lapack_stage_in | ||
stage_out=dplasma_cuda_lapack_stage_out] |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why there is no HIP chore for the normal GEMM (aka. not the summa version) ?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Not all kernels are implemented (tracking issue #98); those that are implemented have all of their particular variants implemented.
This PR adds RoCM enabled kernels to the GEMM, PORTF and memory-aware GEMM operations.