diff options
| author | Georgi Gerganov <ggerganov@gmail.com> | 2024-02-22 23:21:39 +0200 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-02-22 23:21:39 +0200 |
| commit | 7e4f339c404dbe029d4a117c03b37a9bf646cf0e (patch) | |
| tree | ff0774611cd6f110fa83c2fa469920994db96ae6 /ggml.c | |
| parent | 334f76fa385ed81095165e5ae068756214893901 (diff) | |
ggml : always define ggml_fp16_t as uint16_t (#5666)
* ggml : always define ggml_fp16_t as uint16_t
ggml-ci
* ggml : cont
ggml-ci
* ggml : cont
* ggml : cont
ggml-ci
* ggml : cont
ggml-ci
* cuda : no longer ggml headers last
ggml-ci
* ggml : fix q6_K FP16 -> FP32 conversion
ggml-ci
* ggml : more FP16 -> FP32 conversion fixes
ggml-ci
Diffstat (limited to 'ggml.c')
| -rw-r--r-- | ggml.c | 6 |
1 files changed, 3 insertions, 3 deletions
@@ -323,7 +323,7 @@ float ggml_table_f32_f16[1 << 16]; // note: do not use these inside ggml.c // these are meant to be used via the ggml.h API float ggml_fp16_to_fp32(ggml_fp16_t x) { - return (float) GGML_FP16_TO_FP32(x); + return GGML_FP16_TO_FP32(x); } ggml_fp16_t ggml_fp32_to_fp16(float x) { @@ -798,7 +798,7 @@ inline static float vaddvq_f32(float32x4_t v) { #define GGML_F16x8 float16x8_t #define GGML_F16x8_ZERO vdupq_n_f16(0.0f) #define GGML_F16x8_SET1(x) vdupq_n_f16(x) - #define GGML_F16x8_LOAD vld1q_f16 + #define GGML_F16x8_LOAD(x) vld1q_f16((const __fp16 *)(x)) #define GGML_F16x8_STORE vst1q_f16 #define GGML_F16x8_FMA(a, b, c) vfmaq_f16(a, b, c) #define GGML_F16x8_ADD vaddq_f16 @@ -841,7 +841,7 @@ inline static float vaddvq_f32(float32x4_t v) { #define GGML_F32Cx4 float32x4_t #define GGML_F32Cx4_ZERO vdupq_n_f32(0.0f) #define GGML_F32Cx4_SET1(x) vdupq_n_f32(x) - #define GGML_F32Cx4_LOAD(x) vcvt_f32_f16(vld1_f16(x)) + #define GGML_F32Cx4_LOAD(x) vcvt_f32_f16(vld1_f16((const __fp16 *)(x))) #define GGML_F32Cx4_STORE(x, y) vst1_f16(x, vcvt_f16_f32(y)) #define GGML_F32Cx4_FMA(a, b, c) vfmaq_f32(a, b, c) #define GGML_F32Cx4_ADD vaddq_f32 |