Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

cuda : add support for f16->f16 GGML_OP_SET_ROWS (#25367)

Browse Source
This commit is contained in:
fairydreaming
2026-07-08 19:24:20 +08:00
committed by GitHub
parent 90e0f5cfcb
commit ed8c26150e
2 changed files with 74 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-cuda/ggml-cuda.cu
+10 -4
View File
@@ -4709,10 +4709,16 @@ static bool ggml_backend_cuda_device_supports_op(ggml_backend_dev_t dev, const g
} break;
case GGML_OP_SET_ROWS:
{
return (op->type == GGML_TYPE_F32 || op->type == GGML_TYPE_F16 || op->type == GGML_TYPE_BF16 ||
op->type == GGML_TYPE_Q4_0 || op->type == GGML_TYPE_Q4_1 || op->type == GGML_TYPE_Q5_0 ||
op->type == GGML_TYPE_Q5_1 || op->type == GGML_TYPE_Q8_0 || op->type == GGML_TYPE_IQ4_NL) &&
op->src[0]->type == GGML_TYPE_F32 &&
return (
(
(op->type == GGML_TYPE_F32 || op->type == GGML_TYPE_F16 || op->type == GGML_TYPE_BF16 ||
op->type == GGML_TYPE_Q4_0 || op->type == GGML_TYPE_Q4_1 || op->type == GGML_TYPE_Q5_0 ||
op->type == GGML_TYPE_Q5_1 || op->type == GGML_TYPE_Q8_0 || op->type == GGML_TYPE_IQ4_NL) &&
op->src[0]->type == GGML_TYPE_F32
) || (
op->type == GGML_TYPE_F16 && op->src[0]->type == GGML_TYPE_F16
)
) &&
(op->src[1]->type == GGML_TYPE_I64 || op->src[1]->type == GGML_TYPE_I32);
} break;
case GGML_OP_SET: