ggml-cpu: add always_inline to tinyBLAS_PPC accumulator saves - #20791
Merged
Conversation
shalinib-ibm
force-pushed
the
ppc_fp32_inline_fix
branch
from
March 20, 2026 08:00
c0cc0a5 to
0eee946
Compare
github-actions
Bot
added
the
ggml
Contributor
Author
|
The changes in this PR (powerpc: add FP16 MMA path for Q4/Q8 matmul) caused the compiler to stop inlining add_save and add_save_acc in tinyBLAS_PPC. These functions are on the critical path and rely on inlining for optimal performance (especially in the MMA kernels). Request your review. |
taronaeo
approved these changes
Mar 20, 2026
Member
|
Triggered CI. We are experiencing some intermittent problems with our |
Member
Explicitly mark save_acc and add_save_Acc with always_inline in tinyBLAS_PPC. This ensures the compiler keeps MMA accumulator disassembly within kernel's register context, preventing un-necessary stask spills. Signed-off-by: Shalini Salomi Bodapati <Shalini.Salomi.Bodapati@ibm.com>
shalinib-ibm
force-pushed
the
ppc_fp32_inline_fix
branch
from
March 20, 2026 11:01
0eee946 to
4f845d7
Compare
taronaeo
approved these changes
Mar 20, 2026
Seunghhon
pushed a commit
to Seunghhon/llama.cpp
that referenced
this pull request
Apr 26, 2026
…rg#20791) Explicitly mark save_acc and add_save_Acc with always_inline in tinyBLAS_PPC. This ensures the compiler keeps MMA accumulator disassembly within kernel's register context, preventing un-necessary stask spills. Signed-off-by: Shalini Salomi Bodapati <Shalini.Salomi.Bodapati@ibm.com>
ljubomirj
pushed a commit
to ljubomirj/llama.cpp
that referenced
this pull request
May 6, 2026
…rg#20791) Explicitly mark save_acc and add_save_Acc with always_inline in tinyBLAS_PPC. This ensures the compiler keeps MMA accumulator disassembly within kernel's register context, preventing un-necessary stask spills. Signed-off-by: Shalini Salomi Bodapati <Shalini.Salomi.Bodapati@ibm.com>
my-other-github-account
pushed a commit
to my-other-github-account/llama.cpp
that referenced
this pull request
May 15, 2026
…rg#20791) Explicitly mark save_acc and add_save_Acc with always_inline in tinyBLAS_PPC. This ensures the compiler keeps MMA accumulator disassembly within kernel's register context, preventing un-necessary stask spills. Signed-off-by: Shalini Salomi Bodapati <Shalini.Salomi.Bodapati@ibm.com>
my-other-github-account
pushed a commit
to my-other-github-account/llama.cpp
that referenced
this pull request
May 15, 2026
…rg#20791) Explicitly mark save_acc and add_save_Acc with always_inline in tinyBLAS_PPC. This ensures the compiler keeps MMA accumulator disassembly within kernel's register context, preventing un-necessary stask spills. Signed-off-by: Shalini Salomi Bodapati <Shalini.Salomi.Bodapati@ibm.com>
fewtarius
pushed a commit
to fewtarius/CachyLLama
that referenced
this pull request
May 30, 2026
…rg#20791) Explicitly mark save_acc and add_save_Acc with always_inline in tinyBLAS_PPC. This ensures the compiler keeps MMA accumulator disassembly within kernel's register context, preventing un-necessary stask spills. Signed-off-by: Shalini Salomi Bodapati <Shalini.Salomi.Bodapati@ibm.com>
fukuro-kun
pushed a commit
to fukuro-kun/fukuro-llama-cpp-turboquant
that referenced
this pull request
Jul 5, 2026
…rg#20791) Explicitly mark save_acc and add_save_Acc with always_inline in tinyBLAS_PPC. This ensures the compiler keeps MMA accumulator disassembly within kernel's register context, preventing un-necessary stask spills. Signed-off-by: Shalini Salomi Bodapati <Shalini.Salomi.Bodapati@ibm.com>
MrLordCat
referenced
this pull request
in MrLordCat/llama.cpp-rdna-lab
Jul 16, 2026
Explicitly mark save_acc and add_save_Acc with always_inline in tinyBLAS_PPC. This ensures the compiler keeps MMA accumulator disassembly within kernel's register context, preventing un-necessary stask spills. Signed-off-by: Shalini Salomi Bodapati <Shalini.Salomi.Bodapati@ibm.com>
zommiommy
pushed a commit
to zommiommy/llama.cpp
that referenced
this pull request
Aug 18, 2026
…rg#20791) Explicitly mark save_acc and add_save_Acc with always_inline in tinyBLAS_PPC. This ensures the compiler keeps MMA accumulator disassembly within kernel's register context, preventing un-necessary stask spills. Signed-off-by: Shalini Salomi Bodapati <Shalini.Salomi.Bodapati@ibm.com>
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
2 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.
Explicitly mark save_acc and add_save_acc with always_inline in tinyBLAS_PPC. This ensures the compiler keeps MMA accumulator disassembly within kernel's register context, preventing un-necessary stask spills.
Make sure to read the contributing guidelines before submitting a PR