vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

ggml : fix BLAS with unsupported types (#9775)

Browse source 
* ggml : do not use BLAS with types without to_float

* ggml : return pointer from ggml_internal_get_type_traits to avoid unnecessary copies

* ggml : rename ggml_internal_get_type_traits -> ggml_get_type_traits

it's not really internal if everybody uses it
This commit is contained in:
Diego Devesa 2024-10-08 14:21:43 +02:00 • committed by GitHub
parent 458367a906
commit dca1d4b58a
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
13 changed files with 75 additions and 74 deletions
Download patch file Download diff file Expand all files Collapse all files

26
ggml/src/ggml-blas.cpp
View file

@ -65,8 +65,8 @@ static void ggml_backend_blas_mul_mat(ggml_backend_blas_context * ctx, struct gg
// convert src0 to float
if (type != GGML_TYPE_F32) {
ggml_type_traits_t type_traits = ggml_internal_get_type_traits(type);
ggml_to_float_t const to_float = type_traits.to_float;
const auto * type_traits = ggml_get_type_traits(type);
ggml_to_float_t const to_float = type_traits->to_float;
for (int64_t i03 = 0; i03 < ne03; i03++) {
for (int64_t i02 = 0; i02 < ne02; i02++) {
@ -420,19 +420,21 @@ static bool ggml_backend_blas_device_supports_op(ggml_backend_dev_t dev, const s
// TODO: find the optimal value
const int64_t min_batch = 32;
return (ggml_is_contiguous(src0) &&
ggml_is_contiguous(src1) &&
src1->type == GGML_TYPE_F32 &&
(ne0 >= min_batch && ne1 >= min_batch && ne10 >= min_batch));
return ggml_is_contiguous(src0) &&
ggml_is_contiguous(src1) &&
src1->type == GGML_TYPE_F32 &&
(ne0 >= min_batch && ne1 >= min_batch && ne10 >= min_batch) &&
(src0->type == GGML_TYPE_F32 || ggml_get_type_traits(src0->type)->to_float != NULL);
}
case GGML_OP_OUT_PROD:
return (op->src[0]->type == GGML_TYPE_F32 &&
op->src[1]->type == GGML_TYPE_F32 &&
ggml_is_matrix(src0) &&
ggml_is_matrix(src1) &&
ggml_is_contiguous(src0) &&
(ggml_is_contiguous(src1) || ggml_is_transposed(src1)));
return op->src[0]->type == GGML_TYPE_F32 &&
op->src[1]->type == GGML_TYPE_F32 &&
ggml_is_matrix(src0) &&
ggml_is_matrix(src1) &&
ggml_is_contiguous(src0) &&
(ggml_is_contiguous(src1) || ggml_is_transposed(src1)) &&
(src0->type == GGML_TYPE_F32 || ggml_get_type_traits(src0->type)->to_float != NULL);
default:
return false;