We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
There was an error while loading. Please reload this page.
1 parent 615655a commit 71fdcf0Copy full SHA for 71fdcf0
ggml/src/ggml-cpu/arch/arm/repack.cpp
@@ -24,6 +24,7 @@
24
25
#define UNUSED GGML_UNUSED
26
27
+#if defined(__aarch64__) && defined(__ARM_NEON) && defined(__ARM_FEATURE_MATMUL_INT8)
28
static inline void decode_q4_Kx8_scales_mins(const uint8_t * scales_in,
29
int16x8_t * out_mins,
30
int8_t * out_scales) {
@@ -46,6 +47,7 @@ static inline void decode_q4_Kx8_scales_mins(const uint8_t * scales_in,
46
47
scales_u32[1] = (sm[2] & kmask2) | (((sm[0] >> 6) & kmask3) << 4);
48
memcpy(out_scales, scales_u32, 8);
49
}
50
+#endif
51
52
void ggml_quantize_mat_q8_0_4x4(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k) {
53
assert(QK8_0 == 32);
0 commit comments