ggml: support concat for scalar types at cuda backend#24011
Merged
Conversation
zihaomu
changed the title
github-actions
Bot
added
testing
JohannesGaessler
approved these changes
Jun 2, 2026
Contributor
Author
|
Hi @ggerganov, just a friendly ping on this PR when you have time. |
am17an
approved these changes
Jun 10, 2026
github-actions
Bot
added
the
Apple Metal
zihaomu
force-pushed
the
candidate1-concat-non-f32
branch
from
0ff784c to
f697a07
Compare
zihaomu
commented
Jun 11, 2026
| return src0_type == src1_type && | ||
| src0_type == op->type && | ||
| (src0_type == GGML_TYPE_F32 || src0_type == GGML_TYPE_I32); | ||
| } |
Contributor
Author
There was a problem hiding this comment.
Metal concat kernel only supports float-sized data type.
fairydreaming
pushed a commit
to fairydreaming/llama.cpp
that referenced
this pull request
Jun 11, 2026
am17an
approved these changes
Jun 11, 2026
am17an
added
the
merge ready
zihaomu
force-pushed
the
candidate1-concat-non-f32
branch
from
f697a07 to
c96a285
Compare
Contributor
Author
|
Update: Fixed the Objective-C compilation error, and this patch should pass CI now. |
anaisbetts
pushed a commit
to anaisbetts/llama.cpp
that referenced
this pull request
Jun 16, 2026
* cuda: support concat for scalar types * Update concat.cu * fix metal ci issue (cherry picked from commit 85f99dc)
adrianhoehne
pushed a commit
to adrianhoehne/llama.cpp
that referenced
this pull request
Jul 5, 2026
* cuda: support concat for scalar types * Update concat.cu * fix metal ci issue
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
4 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Overview
Make CUDA CONCAT support common non-quantized scalar tensor types, not just F32.
What changed:
The CUDA CONCAT kernel now works for same-type, non-quantized scalar tensors with 1, 2, 4, or 8 byte elements.
Supported examples:
F16, BF16, I8, I16, I32, I64, and F32.
Not included:
Quantized tensor types are still excluded.
Tests:
Added backend CONCAT test coverage for F16, BF16, I8, I16, and I64.