[CUDA][Bindless] Address USM normalized type image creation failure and functionality #2056
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
przemektmalon
approved these changes
Sep 6, 2024
cppchedy
approved these changes
Sep 6, 2024
isaacault
approved these changes
Sep 10, 2024
Seanst98
force-pushed
the
sean/usm-normalized-fix
branch
3 times, most recently
from
b2ac8dc to
94cb7b0
Compare
Seanst98
force-pushed
the
sean/usm-normalized-fix
branch
2 times, most recently
from
e354592 to
1e40614
Compare
Seanst98
force-pushed
the
sean/usm-normalized-fix
branch
5 times, most recently
from
a706c4f to
0df25e2
Compare
…nd functionality
The specialised formats, CU_AD_FORMAT_{U|S}NORM_INT{8|16}X{1|2|4},
used to specify a format that is normalized, cannot be applied to
images backed by USM memory, causing a failure. This commit
addresses that issue by using the standard integer formats and
making sure the CU_TRSF_READ_AS_INTEGER flag is set appropriately.
Another issue fixed here is that users with CUDA toolkits pre 11.5
are unable to enable normalized types due to the specialised
formats being introduced in CUDA 11.5.
Seanst98
force-pushed
the
sean/usm-normalized-fix
branch
from
0df25e2 to
bcf2244
Compare
DBDuncan
approved these changes
Sep 30, 2024
martygrant
pushed a commit
to intel/llvm
that referenced
this pull request
Oct 3, 2024
Modify pre-existing read_norm_types test to add testing for 2D USM bindless images. Associated UR PR which fixes a bug in normalized USM image creation: oneapi-src/unified-runtime#2056
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 participants
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Outdated suggestions cannot be applied.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.
The specialised formats, CU_AD_FORMAT_{U|S}NORM_INT{8|16}X{1|2|4}, used to specify a format that is normalized, cannot be applied to images backed by USM memory, causing a failure. This commit addresses that issue by using the standard integer formats and making sure the CU_TRSF_READ_AS_INTEGER flag is set appropriately.
Another issue fixed here is that users with CUDA toolkits pre 11.5 are unable to enable normalized types due to the specialised formats being introduced in CUDA 11.5.
Associated DPC++ PR which tests for this: intel/llvm#15299