ggml : support AVX512VNNI (ggerganov#6280)

This change causes some quants (e.g. Q4_0, Q8_0) to go faster on some architectures (e.g. AMD Zen 4).
luciancorlan · Mar 25, 2024 · 7733f0c · 7733f0c
1 parent a32b77c
commit 7733f0c
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/ggml-quants.c b/ggml-quants.c
@@ -132,7 +132,7 @@ static inline __m256 sum_i16_pairs_float(const __m256i x) {
 }
 
 static inline __m256 mul_sum_us8_pairs_float(const __m256i ax, const __m256i sy) {
-#if __AVXVNNI__
+#if defined(__AVXVNNI__) || defined(__AVX512VNNI__)
     const __m256i zero = _mm256_setzero_si256();
     const __m256i summed_pairs = _mm256_dpbusd_epi32(zero, ax, sy);
     return _mm256_cvtepi32_ps(summed_pairs);