CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-08 10:07:01 +00:00
Code Issues Packages Projects Releases Wiki Activity

ggml-zdnn: bring load ztensor back to init routine

Browse Source 
Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>
This commit is contained in:
Aaron Teo
2025-07-30 17:21:04 +08:00
parent 1eb7c35e3a
commit 70224e6cb7
1 changed files with 8 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
ggml/src/ggml-zdnn/ggml-zdnn-rewrite.cpp
View File

@@ -137,13 +137,13 @@ static void ggml_zdnn_mul_mat_op(ggml_backend_zdnn_context * ctx, const ggml_ten
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));
if (weights_extra->ztensor.is_transformed == false) {
ggml_zdnn_load_tensor(weights_extra->ztensor, weights->data);
}
// if (weights_extra->ztensor.is_transformed == false) {
// ggml_zdnn_load_tensor(weights_extra->ztensor, weights->data);
// }
if (inputs_extra->ztensor.is_transformed == false) {
ggml_zdnn_load_tensor(inputs_extra->ztensor, inputs->data);
}
// if (inputs_extra->ztensor.is_transformed == false) {
// ggml_zdnn_load_tensor(inputs_extra->ztensor, inputs->data);
// }
ggml_zdnn_load_tensor(zt_bias, bias_data);
// GGML_LOG_INFO("%s: tensor '%s' tensor dimensions: [%ld, %ld, %ld, %ld] pre_tfm_desc dimensions: [%ld, %ld, %ld, %ld]\n",
@@ -432,8 +432,8 @@ static void ggml_backend_zdnn_buffer_memset_tensor(ggml_backend_buffer_t buffer,
static void ggml_backend_zdnn_buffer_set_tensor(ggml_backend_buffer_t buffer, ggml_tensor * tensor, const void * data, size_t offset, size_t size) {
memcpy((char *)tensor->data + offset, data, size);
// ggml_backend_zdnn_buffer * zdnn_buffer = (ggml_backend_zdnn_buffer *)tensor->extra;
// ggml_zdnn_load_tensor(zdnn_buffer->ztensor, (void *)data);
ggml_backend_zdnn_buffer * zdnn_buffer = (ggml_backend_zdnn_buffer *)tensor->extra;
ggml_zdnn_load_tensor(zdnn_buffer->ztensor, (void *)data);
GGML_UNUSED(buffer);
}