tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-26 19:55:04 +00:00
Code Issues Packages Projects Releases Wiki Activity

ggml-cpu: extern c ggml_table_f32_f16 + chore docs

Browse Source 
Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>
This commit is contained in:
Aaron Teo
2025-06-25 23:29:08 +08:00
parent 6cebee25d0
commit 5f2a09a8f6
2 changed files with 6 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
ggml/src/ggml-cpu/ggml-cpu.c
View File

@ -3460,13 +3460,6 @@ int ggml_cpu_has_sme(void) {
}
void ggml_cpu_init(void) {
// needed to initialize f16 tables
{
struct ggml_init_params params = { 0, NULL, false };
struct ggml_context * ctx = ggml_init(params);
ggml_free(ctx);
}
ggml_critical_section_start();
static bool is_first_call = true;

8
ggml/src/ggml-cpu/simd-mappings.h
View File

@ -138,8 +138,12 @@
#endif
// precomputed f32 table for f16 (256 KB)
// defined in ggml.c, initialized in ggml_init()
GGML_API float ggml_table_f32_f16[1 << 16];
// defined in ggml-cpu.c, initialized in ggml_cpu_init()
#ifdef __cplusplus
extern "C" float ggml_table_f32_f16[1 << 16];
#else
extern float ggml_table_f32_f16[1 << 16];
#endif
// On ARM NEON, it's quicker to directly convert x -> x instead of calling into ggml_lookup_fp16_to_fp32,
// so we define GGML_CPU_FP16_TO_FP32 and GGML_CPU_FP32_TO_FP16 elsewhere for NEON.