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

metal : consolidate unary ops (#19490)

Browse Source
This commit is contained in:
Georgi Gerganov
2026-02-11 07:51:12 +02:00
committed by GitHub
parent 2cce9fddb7
commit ceaa89b786
7 changed files with 377 additions and 671 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-metal/ggml-metal-device.m
+10 -11
View File
@@ -1011,6 +1011,15 @@ bool ggml_metal_device_supports_op(ggml_metal_device_t dev, const struct ggml_te
}
switch (op->op) {
case GGML_OP_SCALE:
case GGML_OP_FILL:
case GGML_OP_CLAMP:
case GGML_OP_SQR:
case GGML_OP_SQRT:
case GGML_OP_SIN:
case GGML_OP_COS:
case GGML_OP_LOG:
return ggml_is_contiguous_rows(op->src[0]) && op->src[0]->type == GGML_TYPE_F32;
case GGML_OP_UNARY:
switch (ggml_get_unary_op(op)) {
case GGML_UNARY_OP_TANH:
@@ -1030,7 +1039,7 @@ bool ggml_metal_device_supports_op(ggml_metal_device_t dev, const struct ggml_te
case GGML_UNARY_OP_EXP:
case GGML_UNARY_OP_SOFTPLUS:
case GGML_UNARY_OP_EXPM1:
return ggml_is_contiguous(op->src[0]) && op->src[0]->type == GGML_TYPE_F32;
return ggml_is_contiguous_rows(op->src[0]) && op->src[0]->type == GGML_TYPE_F32;
default:
return false;
}
@@ -1061,8 +1070,6 @@ bool ggml_metal_device_supports_op(ggml_metal_device_t dev, const struct ggml_te
return ggml_is_contiguous_rows(op->src[0]) && ggml_is_contiguous_rows(op->src[1]) && op->src[0]->type == GGML_TYPE_F32;
case GGML_OP_ACC:
case GGML_OP_REPEAT:
case GGML_OP_SCALE:
case GGML_OP_FILL:
case GGML_OP_CONV_TRANSPOSE_1D:
return true;
case GGML_OP_CONV_TRANSPOSE_2D:
@@ -1070,14 +1077,6 @@ bool ggml_metal_device_supports_op(ggml_metal_device_t dev, const struct ggml_te
(op->src[0]->type == GGML_TYPE_F16 || op->src[0]->type == GGML_TYPE_F32) &&
op->src[1]->type == GGML_TYPE_F32 &&
op->type == GGML_TYPE_F32;
case GGML_OP_CLAMP:
return op->src[0]->type == GGML_TYPE_F32;
case GGML_OP_SQR:
case GGML_OP_SQRT:
case GGML_OP_SIN:
case GGML_OP_COS:
case GGML_OP_LOG:
return ggml_is_contiguous(op->src[0]) && op->src[0]->type == GGML_TYPE_F32;
case GGML_OP_SUM:
return has_simdgroup_reduction && ggml_is_contiguous(op->src[0]);
case GGML_OP_TRI: