vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-05-09 19:46:11 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

fixed clblast. but this part might not actually be helpful speed wise

Browse source
This commit is contained in:
Concedo 2025-03-30 21:27:52 +08:00
parent e1d3c19673
commit 621aa7c825
1 changed files with 2 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

3
ggml/src/ggml-cpu/binary-ops.cpp
View file

@ -58,6 +58,7 @@ static void apply_binary_op(const ggml_compute_params * params, ggml_tensor * ds
GGML_ASSERT(ggml_can_repeat(src1, src0) && ggml_are_same_shape(src0, dst));
#if defined(GGML_USE_CLBLAST)
//do we even need this? it seems like its actually slower than just CPU
const int ith = params->ith;
if (op == op_add && src0->type == GGML_TYPE_F32 && src1->type == GGML_TYPE_F32 && src1->clblast_offload_gpu) {
// TODO: OpenCL kernel support full broadcast
@ -73,7 +74,7 @@ static void apply_binary_op(const ggml_compute_params * params, ggml_tensor * ds
static_assert(GGML_MAX_DIMS == 4, "GGML_MAX_DIMS is not 4 - update this function");
GGML_ASSERT((src1->ne[0] == src0->ne[0]) && ggml_can_repeat(src1, src0));
if (ith == 0) {
ggml_cl_add(src0, src1, dst);
ggml_cl_mul(src0, src1, dst);
}
return;
}