CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-09 10:17:06 +00:00
Code Issues Packages Projects Releases Wiki Activity

ggml-zdnn: clean up matmul selection

Browse Source 
Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>
This commit is contained in:
Aaron Teo
2025-07-30 17:34:15 +08:00
parent e67feafc65
commit 90d460c20b
1 changed files with 3 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
ggml/src/ggml-zdnn/ggml-zdnn-rewrite.cpp
View File

@@ -129,11 +129,7 @@ static void ggml_zdnn_mul_mat_op(ggml_backend_zdnn_context * ctx, const ggml_ten
const int64_t output_rows = ne1;
const int64_t output_cols = ne0;
const int64_t weights_dim[GGML_MAX_DIMS] = { 1, 1, weights_cols, weights_rows };
const int64_t inputs_dim[GGML_MAX_DIMS] = { 1, 1, inputs_cols, inputs_rows };
const int64_t bias_dim [GGML_MAX_DIMS] = { 1, 1, 1, output_cols };
const int64_t output_dim[GGML_MAX_DIMS] = { 1, 1, output_cols, output_rows };
ggml_zdnn_create_tensor(ptd_bias, td_bias, zt_bias, output, bias_dim, ZDNN_1D);
void * bias_data = (void *)calloc(ne0, ggml_element_size(output));
@@ -277,13 +273,12 @@ static bool ggml_zdnn_supports_op(const ggml_backend_zdnn_device_context * ctx_d
const int64_t ne0 = op->ne[0];
const int64_t ne1 = op->ne[1];
const int64_t max_batch = zdnn_get_nnpa_max_dim_idx_size();
const int64_t max_batch = ctx_dev->max_size;
return ggml_is_contiguous(src0) &&
ggml_is_contiguous(src1) &&
src1->type == GGML_TYPE_F32 &&
(ne0 <= max_batch && ne1 <= max_batch && ne10 <= max_batch) &&
(src0->type == GGML_TYPE_F32 || ggml_get_type_traits(src0->type)->to_float != NULL);
src0->type == GGML_TYPE_F32 && src1->type == GGML_TYPE_F32 &&
(ne0 <= max_batch && ne1 <= max_batch && ne10 <= max_batch);
} break;
default: