Skip to content

LLM inference in C/C++

C++ ggml

ggml-cpu : select the NEON fp32->fp16 conversion (#30505)

Bernard Ladenthin119f68d7d441
diff --git a/ggml/src/ggml-cpu/simd-mappings.h b/ggml/src/ggml-cpu/simd-mappings.h
index ebb54a4ab8..b7b390db42 100644
--- a/ggml/src/ggml-cpu/simd-mappings.h
+++ b/ggml/src/ggml-cpu/simd-mappings.h
@@ -42,6 +42,7 @@ extern "C" {
#define GGML_CPU_COMPUTE_FP32_TO_FP16(x) neon_compute_fp32_to_fp16(x)
#define GGML_CPU_FP16_TO_FP32(x) GGML_CPU_COMPUTE_FP16_TO_FP32(x)
+ #define GGML_CPU_FP32_TO_FP16(x) GGML_CPU_COMPUTE_FP32_TO_FP16(x)
static inline float neon_compute_fp16_to_fp32(ggml_fp16_t h) {
__fp16 tmp;
@@ -150,9 +151,15 @@ extern float ggml_table_f32_ue4m3[1 << 8];
#define GGML_CPU_FP32_TO_FP16(x) GGML_CPU_COMPUTE_FP32_TO_FP16(x)
#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.
-// This is also true for POWER9.
+// a branch that defines a GGML_CPU_COMPUTE_* macro must also select it, or the fallback takes over silently
+#if defined(GGML_CPU_COMPUTE_FP16_TO_FP32) && !defined(GGML_CPU_FP16_TO_FP32) && !defined(__F16C__) // see: https://github.com/ggml-org/llama.cpp/issues/30342
+#error "a branch defined GGML_CPU_COMPUTE_FP16_TO_FP32 without selecting it as GGML_CPU_FP16_TO_FP32"
+#endif
+#if defined(GGML_CPU_COMPUTE_FP32_TO_FP16) && !defined(GGML_CPU_FP32_TO_FP16)
+#error "a branch defined GGML_CPU_COMPUTE_FP32_TO_FP16 without selecting it as GGML_CPU_FP32_TO_FP16"
+#endif
+
+// fall back to the lookup table / the portable software conversion for whichever direction no branch selected
#if !defined(GGML_CPU_FP16_TO_FP32)
inline static float ggml_lookup_fp16_to_fp32(ggml_fp16_t f) {
uint16_t s;