CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-13 10:57:15 +00:00
Code Issues Packages Projects Releases Wiki Activity

cuda: add SET operation support (#16804)

Browse Source 
* feat(cuda): add GGML_OP_SET support

Implement CUDA kernel for SET operation with f32 support.

All tests passing (14598/14598).

* cuda(set): add I32 support; keep F32

* refactor(cuda): use ggml_cuda_cpy to unify SET operator logic and remove code duplication

* Update ggml/src/ggml-cuda/ggml-cuda.cu

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>

* Update ggml/src/ggml-cuda/set.cu

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>

---------

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
This commit is contained in:
YaelGitAccount
2025-10-28 21:10:28 +02:00
committed by GitHub
parent 85a7d8677b
commit 851553ea6b
3 changed files with 57 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
ggml/src/ggml-cuda/ggml-cuda.cu
View File

@@ -50,6 +50,7 @@
#include "ggml-cuda/upscale.cuh"
#include "ggml-cuda/wkv.cuh"
#include "ggml-cuda/gla.cuh"
#include "ggml-cuda/set.cuh"
#include "ggml-cuda/set-rows.cuh"
#include "ggml-cuda/pad_reflect_1d.cuh"
#include "ggml.h"
@@ -2416,6 +2417,9 @@ static bool ggml_cuda_compute_forward(ggml_backend_cuda_context & ctx, struct gg
case GGML_OP_SET_ROWS:
ggml_cuda_op_set_rows(ctx, dst);
break;
case GGML_OP_SET:
ggml_cuda_op_set(ctx, dst);
break;
case GGML_OP_DUP:
ggml_cuda_dup(ctx, dst);
break;
@@ -3842,6 +3846,13 @@ static bool ggml_backend_cuda_device_supports_op(ggml_backend_dev_t dev, const g
op->src[0]->type == GGML_TYPE_F32 &&
(op->src[1]->type == GGML_TYPE_I64 || op->src[1]->type == GGML_TYPE_I32);
} break;
case GGML_OP_SET:
{
const ggml_type t = op->type;
return (t == GGML_TYPE_F32 || t == GGML_TYPE_I32) &&
t == op->src[0]->type &&
t == op->src[1]->type;
} break;
case GGML_OP_CPY:
{
ggml_type src0_type = op->src[0]->type;