tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-28 20:25:20 +00:00
Code Issues Packages Projects Releases Wiki Activity

ggml-cpu: support IQ4_NL_4_4 by runtime repack (#10541)

Browse Source 
* ggml-cpu: support IQ4_NL_4_4 by runtime repack

* ggml-cpu: add __ARM_FEATURE_DOTPROD guard
This commit is contained in:
Shupei Fan
2024-11-28 20:52:03 +08:00
committed by GitHub
parent 2025fa67e9
commit c202cef168
8 changed files with 352 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
ggml/src/ggml-cpu/ggml-cpu.cpp
View File

@ -457,7 +457,7 @@ static bool ggml_backend_cpu_device_supports_op(ggml_backend_dev_t dev, const st
const struct ggml_tensor * src1 = op->src[1];
if (src0 && src0->buffer && ggml_backend_cpu_buft_is_aarch64(src0->buffer->buft)) {
if (op->op != GGML_OP_MUL_MAT || src0->type != GGML_TYPE_Q4_0 || ggml_aarch64_get_optimal_repack_type(src0) == GGML_TYPE_Q4_0) {
if (op->op != GGML_OP_MUL_MAT || src0->type == ggml_aarch64_get_optimal_repack_type(src0)) {
return false;
}
}