From: Stephan Walter <redacted>
Date: Thu, 13 Apr 2023 14:59:50 +0000 (+0000)
Subject: ggml : optimize non-SIMD Q4_0 vector dot product (#703)
X-Git-Tag: gguf-v0.4.0~972
X-Git-Url: https://git.djapps.eu/?a=commitdiff_plain;h=6232f2d7fd7a22d5eeb62182b2f21fcf01359754;p=pkg%2Fggml%2Fsources%2Fllama.cpp

ggml : optimize non-SIMD Q4_0 vector dot product (#703)
---

diff --git a/ggml.c b/ggml.c
index 281fd8ec..eb47d829 100644
--- a/ggml.c
+++ b/ggml.c
@@ -2160,18 +2160,20 @@ static void ggml_vec_dot_q4_0(const int n, float * restrict s, const void * rest
         const uint8_t * restrict p0 = x[i].qs;
         const uint8_t * restrict p1 = y[i].qs;
 
+        int sumi = 0;
         for (int j = 0; j < QK/2; j++) {
             const uint8_t v0 = p0[j];
             const uint8_t v1 = p1[j];
 
-            const float f0 = d0*((int8_t) (v0 & 0xf) - 8);
-            const float f1 = d0*((int8_t) (v0 >> 4)  - 8);
+            const int8_t i0 = (int8_t) (v0 & 0xf) - 8;
+            const int8_t i1 = (int8_t) (v0 >> 4)  - 8;
 
-            const float f2 = d1*((int8_t) (v1 & 0xf) - 8);
-            const float f3 = d1*((int8_t) (v1 >> 4)  - 8);
+            const int8_t i2 = (int8_t) (v1 & 0xf) - 8;
+            const int8_t i3 = (int8_t) (v1 >> 4)  - 8;
 
-            sumf += f0*f2 + f1*f3;
+            sumi += i0*i2 + i1*i3;
         }
+        sumf += d0 * d1 * sumi;
     }
 #endif