|
@@ -317,8 +317,6 @@ static bool ggml_backend_amx_device_supports_op(ggml_backend_dev_t dev, const st
|
|
|
const enum ggml_type type = src0->type;
|
|
const enum ggml_type type = src0->type;
|
|
|
const int64_t ne0 = op->ne[0];
|
|
const int64_t ne0 = op->ne[0];
|
|
|
|
|
|
|
|
- bool is_training = src0->grad || src1->grad;
|
|
|
|
|
-
|
|
|
|
|
// amx kernels enables for Q4_0, Q4_1, Q8_0, F16
|
|
// amx kernels enables for Q4_0, Q4_1, Q8_0, F16
|
|
|
// Q4_K, Q5_K, Q6_K, IQ4_XS enabled for QK_K = 256
|
|
// Q4_K, Q5_K, Q6_K, IQ4_XS enabled for QK_K = 256
|
|
|
bool has_amx_kernels = qtype_has_amx_kernels(type) || (type == GGML_TYPE_F16);
|
|
bool has_amx_kernels = qtype_has_amx_kernels(type) || (type == GGML_TYPE_F16);
|
|
@@ -326,7 +324,6 @@ static bool ggml_backend_amx_device_supports_op(ggml_backend_dev_t dev, const st
|
|
|
bool can_use_amx =
|
|
bool can_use_amx =
|
|
|
is_contiguous_2d(src0) && // src0 must be contiguous
|
|
is_contiguous_2d(src0) && // src0 must be contiguous
|
|
|
is_contiguous_2d(src1) && // src1 must be contiguous
|
|
is_contiguous_2d(src1) && // src1 must be contiguous
|
|
|
- !is_training && // inference only
|
|
|
|
|
src1->type == GGML_TYPE_F32 && // src1 must be float32
|
|
src1->type == GGML_TYPE_F32 && // src1 must be float32
|
|
|
has_amx_kernels && // with amx kernel impls
|
|
has_amx_kernels && // with amx kernel impls
|
|
|
ne0 % (TILE_N * 2) == 0; // out_features is 32x
|
|
ne0 % (TILE_N * 2) == 0; // out_features is 32x
|
Georgi Gerganov