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

metal : add CONV_2D_DW (depthwise convolution) support (#21565)

Browse Source 
* metal : add CONV_2D_DW (depthwise 2D convolution) support

* test : add perf cases for CONV_2D_DW

* metal : use 3D dispatch for CONV_2D_DW kernel

* metal : add channel-tiled CONV_2D_DW kernel for non-contiguous layouts

* metal : simplify CONV_2D_DW dispatch and trim comments

* metal : merge duplicate CONV_2D_DW pipeline getters

* tests : add F16 CONV2D_DW tests

* cpu : fix F16 kernel support for CONV_2D_DW

* tests : remove commented-out CONV_2D_DW test block

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit is contained in:
Sou-ly
2026-07-09 12:29:15 +03:00
committed by GitHub
co-authored by Georgi Gerganov
parent ccb0c34223
commit 92b187c97e
9 changed files with 381 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-metal/ggml-metal-device.m
+4
View File
@@ -1198,6 +1198,10 @@ bool ggml_metal_device_supports_op(ggml_metal_device_t dev, const struct ggml_te
op->src[1]->type == GGML_TYPE_F32 &&
op->type == GGML_TYPE_F32 &&
(op->src[0]->type == GGML_TYPE_F16 || op->src[0]->type == GGML_TYPE_F32);
case GGML_OP_CONV_2D_DW:
return op->src[1]->type == GGML_TYPE_F32 &&
op->type == GGML_TYPE_F32 &&
(op->src[0]->type == GGML_TYPE_F16 || op->src[0]->type == GGML_TYPE_F32);
case GGML_OP_UPSCALE:
return op->src[0]->type == GGML_TYPE_F32;
case GGML_OP_POOL_1D: