Fix numVgprs issue when LDSTrInst is enabled#596
Merged
Conversation
amgddm
requested review from
TonyYHsieh,
aazz44ss,
babakpst,
msujon-AMD,
vin-huang and
yoichiyoshida
amgddm
requested review from
AlexBrownAMD,
Jinp800125,
KKyang,
Serge45,
bnemanich,
davidd-amd,
hcman2,
imcarsonliao,
jichangjichang and
solaslin
as code owners
hcman2
reviewed
Jul 11, 2025
…void using LDSTr and multiple LDS buffer for HSS or BSS
amgddm
force-pushed
the
LDSTr-NumVgpr-HSS-MLDSB
branch
from
3756e88 to
1fd355d
Compare
amgddm
changed the title
nakajee
reviewed
Jul 16, 2025
amgddm
force-pushed
the
LDSTr-NumVgpr-HSS-MLDSB
branch
from
1fd355d to
968ddc6
Compare
Contributor
Author
What do you mean? |
b-shi
approved these changes
Jul 17, 2025
assistant-librarian Bot
pushed a commit
to ROCm/hipBLASLt
that referenced
this pull request
Jul 17, 2025
Fix numVgprs issue when LDSTrInst is enabled (#596) numVgprs is not calculated correctly which prevents users from using larger DepthU. hipBLASLT complains about the number of registers is greater than 256 if DU is set to 64 for a GEMM shape of 4096x4096x8192. This PR fix that issue.
ammallya
pushed a commit
that referenced
this pull request
Jul 22, 2025
* Adding codecoverage to CI * Fix incorrect use of runCompileCOmmand in codecov groovy file * Add missing runCoverageCommand to common groovy * Fix code coverage test filter to run all hipsparse tests [ROCm/hipSPARSE commit: 9cb0596]
aferoz21
pushed a commit
that referenced
this pull request
Sep 4, 2025
numVgprs is not calculated correctly which prevents users from using larger DepthU. hipBLASLT complains about the number of registers is greater than 256 if DU is set to 64 for a GEMM shape of 4096x4096x8192. This PR fix that issue.
3 tasks
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 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.
numVgprs is not calculated correctly which prevents users from using larger DepthU. hipBLASLT complains about number of registers is greater than 256 if DU is set to 64 for a GEMM shape of 4096x4096x8192. This PR fix that issue.