llama_cpp 0.13.0 → 0.14.0

data/vendor/tmp/llama.cpp/ggml-quants.c

@@ -51,6 +51,7 @@
 
 #define UNUSED GGML_UNUSED
 
+// some compilers don't provide _mm256_set_m128i, e.g. gcc 7
 #define MM256_SET_M128I(a, b) _mm256_insertf128_si256(_mm256_castsi128_si256(b), (a), 1)
 
 #if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) || defined(__SSSE3__)
@@ -463,8 +464,8 @@ inline static int8x16_t ggml_vqtbl1q_s8(int8x16_t a, uint8x16_t b) {
 }
 
 // NOTE: not tested
-inline static int8x16_t ggml_vqtbl1q_u8(uint8x16_t a, uint8x16_t b) {
-    int8x16_t res;
+inline static uint8x16_t ggml_vqtbl1q_u8(uint8x16_t a, uint8x16_t b) {
+    uint8x16_t res;
 
     res[ 0] = a[b[ 0]];
     res[ 1] = a[b[ 1]];
@@ -3818,71 +3819,71 @@ static const uint32_t iq3xxs_grid[256] = {
     0x3e1c1c1c, 0x3e1c3404, 0x3e24140c, 0x3e24240c, 0x3e2c0404, 0x3e2c0414, 0x3e2c1424, 0x3e341c04,
 };
 
-static const uint32_t iq3xs_grid[512] = {
-    0x04040404, 0x0404040c, 0x04040414, 0x0404042c, 0x0404043e, 0x04040c04, 0x04040c0c, 0x04040c14,
-    0x04040c24, 0x04040c34, 0x04041404, 0x0404140c, 0x0404142c, 0x04041c1c, 0x04042404, 0x04042414,
-    0x0404242c, 0x0404243e, 0x04042c0c, 0x04042c1c, 0x04043404, 0x04043414, 0x04043e0c, 0x04043e24,
-    0x04043e3e, 0x040c0404, 0x040c040c, 0x040c0414, 0x040c0424, 0x040c0c04, 0x040c0c0c, 0x040c0c2c,
-    0x040c1404, 0x040c141c, 0x040c143e, 0x040c1c0c, 0x040c1c2c, 0x040c2424, 0x040c340c, 0x040c342c,
-    0x040c3e14, 0x04140404, 0x0414040c, 0x0414042c, 0x0414043e, 0x04140c04, 0x04140c1c, 0x04140c34,
-    0x0414140c, 0x0414142c, 0x04141c04, 0x04141c24, 0x04142414, 0x0414242c, 0x0414243e, 0x04142c0c,
-    0x04142c1c, 0x04143e04, 0x04143e1c, 0x041c041c, 0x041c0c0c, 0x041c0c2c, 0x041c1404, 0x041c1414,
-    0x041c1c0c, 0x041c1c1c, 0x041c1c34, 0x041c2424, 0x041c2c04, 0x041c2c14, 0x041c343e, 0x041c3e0c,
-    0x041c3e2c, 0x04240404, 0x04240c1c, 0x04240c3e, 0x0424140c, 0x04241424, 0x04241c14, 0x04242404,
-    0x0424241c, 0x04242c0c, 0x04243e04, 0x042c0414, 0x042c0424, 0x042c1404, 0x042c1414, 0x042c1434,
-    0x042c1c1c, 0x042c240c, 0x042c242c, 0x042c243e, 0x042c3434, 0x042c3e1c, 0x04340434, 0x04340c0c,
-    0x04340c1c, 0x04341c0c, 0x04342c14, 0x04343e0c, 0x043e0404, 0x043e0414, 0x043e0424, 0x043e1404,
-    0x043e1414, 0x043e1434, 0x043e1c1c, 0x043e2c04, 0x043e2c24, 0x0c040404, 0x0c04040c, 0x0c040414,
-    0x0c040424, 0x0c040c04, 0x0c040c0c, 0x0c040c1c, 0x0c040c2c, 0x0c040c3e, 0x0c041404, 0x0c041414,
-    0x0c041c0c, 0x0c041c24, 0x0c041c34, 0x0c042c24, 0x0c042c34, 0x0c04340c, 0x0c043e14, 0x0c0c0404,
-    0x0c0c040c, 0x0c0c041c, 0x0c0c0434, 0x0c0c0c04, 0x0c0c0c24, 0x0c0c140c, 0x0c0c1c04, 0x0c0c1c1c,
-    0x0c0c240c, 0x0c0c2c04, 0x0c0c2c14, 0x0c0c3e04, 0x0c0c3e34, 0x0c140404, 0x0c140c14, 0x0c140c2c,
-    0x0c140c3e, 0x0c141404, 0x0c141424, 0x0c141c14, 0x0c142404, 0x0c14241c, 0x0c142c2c, 0x0c143404,
-    0x0c143e14, 0x0c1c040c, 0x0c1c0424, 0x0c1c043e, 0x0c1c0c04, 0x0c1c0c1c, 0x0c1c140c, 0x0c1c143e,
-    0x0c1c1c04, 0x0c1c1c24, 0x0c1c240c, 0x0c1c3414, 0x0c1c3e04, 0x0c24041c, 0x0c24042c, 0x0c240c14,
-    0x0c240c24, 0x0c241c0c, 0x0c241c1c, 0x0c242414, 0x0c242434, 0x0c242c04, 0x0c242c24, 0x0c2c040c,
-    0x0c2c0c04, 0x0c2c0c1c, 0x0c2c140c, 0x0c2c1c04, 0x0c2c1c14, 0x0c2c2c0c, 0x0c341404, 0x0c341424,
-    0x0c34143e, 0x0c342424, 0x0c342434, 0x0c3e040c, 0x0c3e041c, 0x0c3e0c04, 0x0c3e0c14, 0x0c3e140c,
-    0x0c3e1c2c, 0x0c3e240c, 0x0c3e3414, 0x0c3e3e04, 0x14040404, 0x1404040c, 0x1404041c, 0x1404042c,
-    0x1404043e, 0x14040c04, 0x14040c14, 0x14040c24, 0x14040c34, 0x1404140c, 0x1404141c, 0x1404143e,
-    0x14041c04, 0x14041c14, 0x1404240c, 0x1404241c, 0x1404242c, 0x14042c04, 0x14042c14, 0x1404343e,
-    0x14043e04, 0x14043e1c, 0x14043e2c, 0x140c0404, 0x140c0414, 0x140c0c04, 0x140c0c1c, 0x140c0c3e,
-    0x140c1414, 0x140c142c, 0x140c1c0c, 0x140c1c24, 0x140c2414, 0x140c2c0c, 0x1414040c, 0x14140424,
-    0x1414043e, 0x1414140c, 0x1414141c, 0x14141c04, 0x14141c3e, 0x1414240c, 0x14142c1c, 0x14142c3e,
-    0x14143e0c, 0x14143e24, 0x141c0404, 0x141c0414, 0x141c042c, 0x141c0c0c, 0x141c1414, 0x141c1424,
-    0x141c1c0c, 0x141c1c1c, 0x141c2414, 0x141c2c04, 0x141c3434, 0x1424040c, 0x1424043e, 0x14241404,
-    0x1424141c, 0x14241c14, 0x14241c2c, 0x1424240c, 0x14243e14, 0x14243e2c, 0x142c0424, 0x142c0c0c,
-    0x142c1414, 0x142c1c3e, 0x142c2404, 0x142c2c1c, 0x142c3e04, 0x14340404, 0x14340414, 0x1434043e,
-    0x1434140c, 0x14342c2c, 0x1434340c, 0x143e042c, 0x143e0c0c, 0x143e1434, 0x143e1c04, 0x143e241c,
-    0x143e2c04, 0x1c040414, 0x1c040c0c, 0x1c040c1c, 0x1c040c2c, 0x1c040c3e, 0x1c041414, 0x1c041c0c,
-    0x1c041c1c, 0x1c041c2c, 0x1c042414, 0x1c042424, 0x1c04243e, 0x1c042c0c, 0x1c04341c, 0x1c043e0c,
-    0x1c0c040c, 0x1c0c041c, 0x1c0c042c, 0x1c0c0c24, 0x1c0c140c, 0x1c0c141c, 0x1c0c2404, 0x1c0c3404,
-    0x1c0c3e14, 0x1c0c3e34, 0x1c140404, 0x1c140c14, 0x1c141404, 0x1c141c14, 0x1c141c24, 0x1c142c04,
-    0x1c1c040c, 0x1c1c0c04, 0x1c1c0c24, 0x1c1c140c, 0x1c1c141c, 0x1c1c143e, 0x1c1c1c04, 0x1c1c240c,
-    0x1c1c241c, 0x1c1c243e, 0x1c1c2c2c, 0x1c1c3e1c, 0x1c24041c, 0x1c240c0c, 0x1c240c34, 0x1c241414,
-    0x1c241c0c, 0x1c242c14, 0x1c243404, 0x1c243424, 0x1c2c040c, 0x1c2c0c04, 0x1c2c0c14, 0x1c2c142c,
-    0x1c2c1c14, 0x1c2c2424, 0x1c2c2c34, 0x1c2c3e1c, 0x1c340c34, 0x1c34240c, 0x1c3e040c, 0x1c3e041c,
-    0x1c3e1404, 0x1c3e1414, 0x1c3e1c2c, 0x24040404, 0x24040424, 0x24040c14, 0x24041404, 0x24041424,
-    0x2404143e, 0x24041c14, 0x2404240c, 0x24042c04, 0x24043e04, 0x240c0414, 0x240c043e, 0x240c0c0c,
-    0x240c0c1c, 0x240c1414, 0x240c1c04, 0x240c1c2c, 0x240c241c, 0x240c2c0c, 0x240c2c2c, 0x2414040c,
-    0x2414041c, 0x24140c04, 0x24140c2c, 0x2414140c, 0x24141c1c, 0x24142404, 0x24142c3e, 0x24143414,
-    0x24143e04, 0x241c0424, 0x241c0c0c, 0x241c0c1c, 0x241c1404, 0x241c1414, 0x241c1c0c, 0x241c1c2c,
-    0x24240404, 0x24240414, 0x24241424, 0x24241c3e, 0x24242404, 0x24243e0c, 0x242c042c, 0x242c043e,
-    0x242c140c, 0x242c3414, 0x24340c1c, 0x24341c24, 0x24343404, 0x243e0c04, 0x243e0c2c, 0x243e1c04,
-    0x243e241c, 0x243e2c0c, 0x2c040414, 0x2c040c04, 0x2c040c24, 0x2c041414, 0x2c042404, 0x2c042424,
-    0x2c04243e, 0x2c042c14, 0x2c043434, 0x2c043e24, 0x2c0c040c, 0x2c0c041c, 0x2c0c042c, 0x2c0c0c14,
-    0x2c0c140c, 0x2c0c1c14, 0x2c0c3e14, 0x2c140404, 0x2c140c0c, 0x2c14141c, 0x2c141c04, 0x2c141c34,
-    0x2c142c1c, 0x2c1c0414, 0x2c1c043e, 0x2c1c0c04, 0x2c1c143e, 0x2c1c2424, 0x2c1c2c0c, 0x2c1c342c,
-    0x2c1c3e1c, 0x2c24040c, 0x2c240424, 0x2c241404, 0x2c241c14, 0x2c242434, 0x2c2c0c14, 0x2c2c1434,
-    0x2c2c2c0c, 0x2c2c2c1c, 0x2c342414, 0x2c3e0414, 0x2c3e0424, 0x2c3e1414, 0x34040c0c, 0x34040c1c,
-    0x34040c2c, 0x34041c0c, 0x34041c1c, 0x34043404, 0x340c0404, 0x340c1404, 0x340c143e, 0x340c3424,
-    0x34140c14, 0x34141c24, 0x34142414, 0x34142c2c, 0x34143414, 0x34143e04, 0x341c0404, 0x341c0c24,
-    0x341c140c, 0x341c2404, 0x3424142c, 0x3424241c, 0x34243414, 0x342c0404, 0x342c041c, 0x342c1c24,
-    0x342c3404, 0x3434042c, 0x34342404, 0x343e0c0c, 0x343e0c1c, 0x3e040404, 0x3e040424, 0x3e04043e,
-    0x3e041404, 0x3e041414, 0x3e041c34, 0x3e042404, 0x3e042c24, 0x3e043414, 0x3e0c0414, 0x3e0c0c0c,
-    0x3e0c1424, 0x3e0c241c, 0x3e0c242c, 0x3e14040c, 0x3e140424, 0x3e140c04, 0x3e140c34, 0x3e14140c,
-    0x3e141c04, 0x3e142c0c, 0x3e1c0414, 0x3e1c1c14, 0x3e1c1c2c, 0x3e1c2c1c, 0x3e24040c, 0x3e24042c,
-    0x3e240c1c, 0x3e241404, 0x3e242c04, 0x3e2c1414, 0x3e2c2414, 0x3e340414, 0x3e341c0c, 0x3e3e0404,
+static const uint32_t iq3s_grid[512] = {
+    0x01010101, 0x01010103, 0x01010105, 0x0101010b, 0x0101010f, 0x01010301, 0x01010303, 0x01010305,
+    0x01010309, 0x0101030d, 0x01010501, 0x01010503, 0x0101050b, 0x01010707, 0x01010901, 0x01010905,
+    0x0101090b, 0x0101090f, 0x01010b03, 0x01010b07, 0x01010d01, 0x01010d05, 0x01010f03, 0x01010f09,
+    0x01010f0f, 0x01030101, 0x01030103, 0x01030105, 0x01030109, 0x01030301, 0x01030303, 0x0103030b,
+    0x01030501, 0x01030507, 0x0103050f, 0x01030703, 0x0103070b, 0x01030909, 0x01030d03, 0x01030d0b,
+    0x01030f05, 0x01050101, 0x01050103, 0x0105010b, 0x0105010f, 0x01050301, 0x01050307, 0x0105030d,
+    0x01050503, 0x0105050b, 0x01050701, 0x01050709, 0x01050905, 0x0105090b, 0x0105090f, 0x01050b03,
+    0x01050b07, 0x01050f01, 0x01050f07, 0x01070107, 0x01070303, 0x0107030b, 0x01070501, 0x01070505,
+    0x01070703, 0x01070707, 0x0107070d, 0x01070909, 0x01070b01, 0x01070b05, 0x01070d0f, 0x01070f03,
+    0x01070f0b, 0x01090101, 0x01090307, 0x0109030f, 0x01090503, 0x01090509, 0x01090705, 0x01090901,
+    0x01090907, 0x01090b03, 0x01090f01, 0x010b0105, 0x010b0109, 0x010b0501, 0x010b0505, 0x010b050d,
+    0x010b0707, 0x010b0903, 0x010b090b, 0x010b090f, 0x010b0d0d, 0x010b0f07, 0x010d010d, 0x010d0303,
+    0x010d0307, 0x010d0703, 0x010d0b05, 0x010d0f03, 0x010f0101, 0x010f0105, 0x010f0109, 0x010f0501,
+    0x010f0505, 0x010f050d, 0x010f0707, 0x010f0b01, 0x010f0b09, 0x03010101, 0x03010103, 0x03010105,
+    0x03010109, 0x03010301, 0x03010303, 0x03010307, 0x0301030b, 0x0301030f, 0x03010501, 0x03010505,
+    0x03010703, 0x03010709, 0x0301070d, 0x03010b09, 0x03010b0d, 0x03010d03, 0x03010f05, 0x03030101,
+    0x03030103, 0x03030107, 0x0303010d, 0x03030301, 0x03030309, 0x03030503, 0x03030701, 0x03030707,
+    0x03030903, 0x03030b01, 0x03030b05, 0x03030f01, 0x03030f0d, 0x03050101, 0x03050305, 0x0305030b,
+    0x0305030f, 0x03050501, 0x03050509, 0x03050705, 0x03050901, 0x03050907, 0x03050b0b, 0x03050d01,
+    0x03050f05, 0x03070103, 0x03070109, 0x0307010f, 0x03070301, 0x03070307, 0x03070503, 0x0307050f,
+    0x03070701, 0x03070709, 0x03070903, 0x03070d05, 0x03070f01, 0x03090107, 0x0309010b, 0x03090305,
+    0x03090309, 0x03090703, 0x03090707, 0x03090905, 0x0309090d, 0x03090b01, 0x03090b09, 0x030b0103,
+    0x030b0301, 0x030b0307, 0x030b0503, 0x030b0701, 0x030b0705, 0x030b0b03, 0x030d0501, 0x030d0509,
+    0x030d050f, 0x030d0909, 0x030d090d, 0x030f0103, 0x030f0107, 0x030f0301, 0x030f0305, 0x030f0503,
+    0x030f070b, 0x030f0903, 0x030f0d05, 0x030f0f01, 0x05010101, 0x05010103, 0x05010107, 0x0501010b,
+    0x0501010f, 0x05010301, 0x05010305, 0x05010309, 0x0501030d, 0x05010503, 0x05010507, 0x0501050f,
+    0x05010701, 0x05010705, 0x05010903, 0x05010907, 0x0501090b, 0x05010b01, 0x05010b05, 0x05010d0f,
+    0x05010f01, 0x05010f07, 0x05010f0b, 0x05030101, 0x05030105, 0x05030301, 0x05030307, 0x0503030f,
+    0x05030505, 0x0503050b, 0x05030703, 0x05030709, 0x05030905, 0x05030b03, 0x05050103, 0x05050109,
+    0x0505010f, 0x05050503, 0x05050507, 0x05050701, 0x0505070f, 0x05050903, 0x05050b07, 0x05050b0f,
+    0x05050f03, 0x05050f09, 0x05070101, 0x05070105, 0x0507010b, 0x05070303, 0x05070505, 0x05070509,
+    0x05070703, 0x05070707, 0x05070905, 0x05070b01, 0x05070d0d, 0x05090103, 0x0509010f, 0x05090501,
+    0x05090507, 0x05090705, 0x0509070b, 0x05090903, 0x05090f05, 0x05090f0b, 0x050b0109, 0x050b0303,
+    0x050b0505, 0x050b070f, 0x050b0901, 0x050b0b07, 0x050b0f01, 0x050d0101, 0x050d0105, 0x050d010f,
+    0x050d0503, 0x050d0b0b, 0x050d0d03, 0x050f010b, 0x050f0303, 0x050f050d, 0x050f0701, 0x050f0907,
+    0x050f0b01, 0x07010105, 0x07010303, 0x07010307, 0x0701030b, 0x0701030f, 0x07010505, 0x07010703,
+    0x07010707, 0x0701070b, 0x07010905, 0x07010909, 0x0701090f, 0x07010b03, 0x07010d07, 0x07010f03,
+    0x07030103, 0x07030107, 0x0703010b, 0x07030309, 0x07030503, 0x07030507, 0x07030901, 0x07030d01,
+    0x07030f05, 0x07030f0d, 0x07050101, 0x07050305, 0x07050501, 0x07050705, 0x07050709, 0x07050b01,
+    0x07070103, 0x07070301, 0x07070309, 0x07070503, 0x07070507, 0x0707050f, 0x07070701, 0x07070903,
+    0x07070907, 0x0707090f, 0x07070b0b, 0x07070f07, 0x07090107, 0x07090303, 0x0709030d, 0x07090505,
+    0x07090703, 0x07090b05, 0x07090d01, 0x07090d09, 0x070b0103, 0x070b0301, 0x070b0305, 0x070b050b,
+    0x070b0705, 0x070b0909, 0x070b0b0d, 0x070b0f07, 0x070d030d, 0x070d0903, 0x070f0103, 0x070f0107,
+    0x070f0501, 0x070f0505, 0x070f070b, 0x09010101, 0x09010109, 0x09010305, 0x09010501, 0x09010509,
+    0x0901050f, 0x09010705, 0x09010903, 0x09010b01, 0x09010f01, 0x09030105, 0x0903010f, 0x09030303,
+    0x09030307, 0x09030505, 0x09030701, 0x0903070b, 0x09030907, 0x09030b03, 0x09030b0b, 0x09050103,
+    0x09050107, 0x09050301, 0x0905030b, 0x09050503, 0x09050707, 0x09050901, 0x09050b0f, 0x09050d05,
+    0x09050f01, 0x09070109, 0x09070303, 0x09070307, 0x09070501, 0x09070505, 0x09070703, 0x0907070b,
+    0x09090101, 0x09090105, 0x09090509, 0x0909070f, 0x09090901, 0x09090f03, 0x090b010b, 0x090b010f,
+    0x090b0503, 0x090b0d05, 0x090d0307, 0x090d0709, 0x090d0d01, 0x090f0301, 0x090f030b, 0x090f0701,
+    0x090f0907, 0x090f0b03, 0x0b010105, 0x0b010301, 0x0b010309, 0x0b010505, 0x0b010901, 0x0b010909,
+    0x0b01090f, 0x0b010b05, 0x0b010d0d, 0x0b010f09, 0x0b030103, 0x0b030107, 0x0b03010b, 0x0b030305,
+    0x0b030503, 0x0b030705, 0x0b030f05, 0x0b050101, 0x0b050303, 0x0b050507, 0x0b050701, 0x0b05070d,
+    0x0b050b07, 0x0b070105, 0x0b07010f, 0x0b070301, 0x0b07050f, 0x0b070909, 0x0b070b03, 0x0b070d0b,
+    0x0b070f07, 0x0b090103, 0x0b090109, 0x0b090501, 0x0b090705, 0x0b09090d, 0x0b0b0305, 0x0b0b050d,
+    0x0b0b0b03, 0x0b0b0b07, 0x0b0d0905, 0x0b0f0105, 0x0b0f0109, 0x0b0f0505, 0x0d010303, 0x0d010307,
+    0x0d01030b, 0x0d010703, 0x0d010707, 0x0d010d01, 0x0d030101, 0x0d030501, 0x0d03050f, 0x0d030d09,
+    0x0d050305, 0x0d050709, 0x0d050905, 0x0d050b0b, 0x0d050d05, 0x0d050f01, 0x0d070101, 0x0d070309,
+    0x0d070503, 0x0d070901, 0x0d09050b, 0x0d090907, 0x0d090d05, 0x0d0b0101, 0x0d0b0107, 0x0d0b0709,
+    0x0d0b0d01, 0x0d0d010b, 0x0d0d0901, 0x0d0f0303, 0x0d0f0307, 0x0f010101, 0x0f010109, 0x0f01010f,
+    0x0f010501, 0x0f010505, 0x0f01070d, 0x0f010901, 0x0f010b09, 0x0f010d05, 0x0f030105, 0x0f030303,
+    0x0f030509, 0x0f030907, 0x0f03090b, 0x0f050103, 0x0f050109, 0x0f050301, 0x0f05030d, 0x0f050503,
+    0x0f050701, 0x0f050b03, 0x0f070105, 0x0f070705, 0x0f07070b, 0x0f070b07, 0x0f090103, 0x0f09010b,
+    0x0f090307, 0x0f090501, 0x0f090b01, 0x0f0b0505, 0x0f0b0905, 0x0f0d0105, 0x0f0d0703, 0x0f0f0101,
 };
 
 #define NGRID_IQ2XXS 512
@@ -4162,11 +4163,11 @@ void dequantize_row_iq3_s(const block_iq3_s * restrict x, float * restrict y, in
         const uint8_t * signs = x[i].signs;
 
         for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) {
-            const float db1 = d * (0.5f + (x[i].scales[ib32/2] & 0xf)) * 0.5f;
-            const float db2 = d * (0.5f + (x[i].scales[ib32/2] >>  4)) * 0.5f;
+            const float db1 = d * (1 + 2*(x[i].scales[ib32/2] & 0xf));
+            const float db2 = d * (1 + 2*(x[i].scales[ib32/2] >>  4));
             for (int l = 0; l < 4; ++l) {
-                const uint8_t * grid1 = (const uint8_t *)(iq3xs_grid + (qs[2*l+0] | ((qh[0] << (8-2*l)) & 256)));
-                const uint8_t * grid2 = (const uint8_t *)(iq3xs_grid + (qs[2*l+1] | ((qh[0] << (7-2*l)) & 256)));
+                const uint8_t * grid1 = (const uint8_t *)(iq3s_grid + (qs[2*l+0] | ((qh[0] << (8-2*l)) & 256)));
+                const uint8_t * grid2 = (const uint8_t *)(iq3s_grid + (qs[2*l+1] | ((qh[0] << (7-2*l)) & 256)));
                 for (int j = 0; j < 4; ++j) {
                     y[j+0] = db1 * grid1[j] * (signs[l] & kmask_iq2xs[j+0] ? -1.f : 1.f);
                     y[j+4] = db1 * grid2[j] * (signs[l] & kmask_iq2xs[j+4] ? -1.f : 1.f);
@@ -4176,8 +4177,8 @@ void dequantize_row_iq3_s(const block_iq3_s * restrict x, float * restrict y, in
             qs += 8;
             signs += 4;
             for (int l = 0; l < 4; ++l) {
-                const uint8_t * grid1 = (const uint8_t *)(iq3xs_grid + (qs[2*l+0] | ((qh[1] << (8-2*l)) & 256)));
-                const uint8_t * grid2 = (const uint8_t *)(iq3xs_grid + (qs[2*l+1] | ((qh[1] << (7-2*l)) & 256)));
+                const uint8_t * grid1 = (const uint8_t *)(iq3s_grid + (qs[2*l+0] | ((qh[1] << (8-2*l)) & 256)));
+                const uint8_t * grid2 = (const uint8_t *)(iq3s_grid + (qs[2*l+1] | ((qh[1] << (7-2*l)) & 256)));
                 for (int j = 0; j < 4; ++j) {
                     y[j+0] = db2 * grid1[j] * (signs[l] & kmask_iq2xs[j+0] ? -1.f : 1.f);
                     y[j+4] = db2 * grid2[j] * (signs[l] & kmask_iq2xs[j+4] ? -1.f : 1.f);
@@ -9563,7 +9564,7 @@ void ggml_vec_dot_iq2_xs_q8_K(int n, float * restrict s, size_t bs, const void *
 
         const __m128i odd_bits = _mm_shuffle_epi8(bit_helper, partial_sign_bits_for_counting);
         const __m128i full_sign_bits = _mm_or_si128(partial_sign_bits, odd_bits);
-        const __m256i full_signs = _mm256_set_m128i(full_sign_bits, full_sign_bits);
+        const __m256i full_signs = MM256_SET_M128I(full_sign_bits, full_sign_bits);
 
         const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)y[i].qs);
         const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)(y[i].qs+32));
@@ -9585,8 +9586,8 @@ void ggml_vec_dot_iq2_xs_q8_K(int n, float * restrict s, size_t bs, const void *
         const __m256i dot1  = _mm256_maddubs_epi16(q2_1, q8s_1);
         const __m256i dot2  = _mm256_maddubs_epi16(q2_2, q8s_2);
 
-        const __m256i sc1 = _mm256_set_m128i(_mm_set1_epi16(2*(x[i].scales[0] >> 4)+1), _mm_set1_epi16(2*(x[i].scales[0] & 0xf)+1));
-        const __m256i sc2 = _mm256_set_m128i(_mm_set1_epi16(2*(x[i].scales[1] >> 4)+1), _mm_set1_epi16(2*(x[i].scales[1] & 0xf)+1));
+        const __m256i sc1 = MM256_SET_M128I(_mm_set1_epi16(2*(x[i].scales[0] >> 4)+1), _mm_set1_epi16(2*(x[i].scales[0] & 0xf)+1));
+        const __m256i sc2 = MM256_SET_M128I(_mm_set1_epi16(2*(x[i].scales[1] >> 4)+1), _mm_set1_epi16(2*(x[i].scales[1] & 0xf)+1));
 
         const __m256i sum = _mm256_add_epi32(_mm256_madd_epi16(sc1, dot1), _mm256_madd_epi16(sc2, dot2));
 
@@ -9653,8 +9654,8 @@ void ggml_vec_dot_iq2_xs_q8_K(int n, float * restrict s, size_t bs, const void *
 
             const __m128i full_signs_l = _mm256_castsi256_si128(full_sign_bits);
             const __m128i full_signs_h = _mm256_extractf128_si256(full_sign_bits, 1);
-            const __m256i full_signs_1 = _mm256_set_m128i(full_signs_l, full_signs_l);
-            const __m256i full_signs_2 = _mm256_set_m128i(full_signs_h, full_signs_h);
+            const __m256i full_signs_1 = MM256_SET_M128I(full_signs_l, full_signs_l);
+            const __m256i full_signs_2 = MM256_SET_M128I(full_signs_h, full_signs_h);
 
             __m256i signs;
             signs = _mm256_shuffle_epi8(full_signs_1, block_sign_shuffle_1);
@@ -10089,18 +10090,34 @@ void ggml_vec_dot_iq3_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const v
 
 #if defined(__ARM_NEON)
 
+    typedef union {
+        uint16x8_t vec_index;
+        uint16_t   index[8];
+    } vec_index_t;
+
    static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
                                        0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03
    };
 
     static const uint8_t k_mask2[16] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80,};
 
-    const uint8x16x2_t mask1 = vld1q_u8_x2(k_mask1);
-    const uint8x16_t   mask2 = vld1q_u8(k_mask2);
+    static const int16_t k_shift[8] = {8, 7, 6, 5, 4, 3, 2, 1};
+
+    const uint8x16x2_t mask1  = vld1q_u8_x2(k_mask1);
+    const uint8x16_t   mask2  = vld1q_u8(k_mask2);
+    const int16x8_t    hshift = vld1q_s16(k_shift);
+    const uint16x8_t   m256   = vdupq_n_u16(256);
+    const uint8x16_t   m1     = vdupq_n_u8(1);
 
     uint8x16x2_t vs;
     ggml_int8x16x4_t q3s;
     ggml_int8x16x4_t q8b;
+    vec_index_t idx;
+
+#if QK_K == 256
+    uint32_t scales32[2];
+    const uint8_t * scales8 = (const uint8_t *)scales32;
+#endif
 
     float sumf = 0;
     for (int i = 0; i < nb; ++i) {
@@ -10109,47 +10126,63 @@ void ggml_vec_dot_iq3_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const v
         const uint8_t * restrict qh = x[i].qh;
         const uint16_t * restrict signs = (const uint16_t *)x[i].signs;
         const int8_t   * restrict q8 = y[i].qs;
+
+#if QK_K == 256
+        memcpy(scales32, x[i].scales, 4);
+        scales32[1] = (((scales32[0] >> 4) & 0x0f0f0f0f) << 1) | 0x01010101;
+        scales32[0] = ((scales32[0] & 0x0f0f0f0f) << 1) | 0x01010101;
+#endif
+
         int sumi1 = 0, sumi2 = 0;
         for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) {
             q8b = ggml_vld1q_s8_x4(q8); q8 += 64;
-            const uint32x4_t aux32x4_0 = {iq3xs_grid[qs[ 0] | ((qh[ib32+0] << 8) & 256)], iq3xs_grid[qs[ 1] | ((qh[ib32+0] << 7) & 256)],
-                                          iq3xs_grid[qs[ 2] | ((qh[ib32+0] << 6) & 256)], iq3xs_grid[qs[ 3] | ((qh[ib32+0] << 5) & 256)]};
-            const uint32x4_t aux32x4_1 = {iq3xs_grid[qs[ 4] | ((qh[ib32+0] << 4) & 256)], iq3xs_grid[qs[ 5] | ((qh[ib32+0] << 3) & 256)],
-                                          iq3xs_grid[qs[ 6] | ((qh[ib32+0] << 2) & 256)], iq3xs_grid[qs[ 7] | ((qh[ib32+0] << 1) & 256)]};
-            const uint32x4_t aux32x4_2 = {iq3xs_grid[qs[ 8] | ((qh[ib32+1] << 8) & 256)], iq3xs_grid[qs[ 9] | ((qh[ib32+1] << 7) & 256)],
-                                          iq3xs_grid[qs[10] | ((qh[ib32+1] << 6) & 256)], iq3xs_grid[qs[11] | ((qh[ib32+1] << 5) & 256)]};
-            const uint32x4_t aux32x4_3 = {iq3xs_grid[qs[12] | ((qh[ib32+1] << 4) & 256)], iq3xs_grid[qs[13] | ((qh[ib32+1] << 3) & 256)],
-                                          iq3xs_grid[qs[14] | ((qh[ib32+1] << 2) & 256)], iq3xs_grid[qs[15] | ((qh[ib32+1] << 1) & 256)]};
-            qs += 16;
+
+            const uint8x16_t idx_l = vld1q_u8(qs); qs += 16;
+            idx.vec_index = vorrq_u16(vmovl_u8(vget_low_u8 (idx_l)), vandq_u16(vshlq_u16(vdupq_n_u16(qh[ib32+0]), hshift), m256));
+            const uint32x4_t aux32x4_0 = {iq3s_grid[idx.index[0]], iq3s_grid[idx.index[1]],
+                                          iq3s_grid[idx.index[2]], iq3s_grid[idx.index[3]]};
+            const uint32x4_t aux32x4_1 = {iq3s_grid[idx.index[4]], iq3s_grid[idx.index[5]],
+                                          iq3s_grid[idx.index[6]], iq3s_grid[idx.index[7]]};
+            idx.vec_index = vorrq_u16(vmovl_u8(vget_high_u8(idx_l)), vandq_u16(vshlq_u16(vdupq_n_u16(qh[ib32+1]), hshift), m256));
+            const uint32x4_t aux32x4_2 = {iq3s_grid[idx.index[0]], iq3s_grid[idx.index[1]],
+                                          iq3s_grid[idx.index[2]], iq3s_grid[idx.index[3]]};
+            const uint32x4_t aux32x4_3 = {iq3s_grid[idx.index[4]], iq3s_grid[idx.index[5]],
+                                          iq3s_grid[idx.index[6]], iq3s_grid[idx.index[7]]};
+
 
             vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[0] | (signs[1] << 16)));
             vs.val[1] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2);
             vs.val[0] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2);
-            vs.val[0] = vceqq_u8(vs.val[0], mask2);
-            vs.val[1] = vceqq_u8(vs.val[1], mask2);
+            vs.val[0] = vorrq_u8(vceqq_u8(vs.val[0], mask2), m1);
+            vs.val[1] = vorrq_u8(vceqq_u8(vs.val[1], mask2), m1);
 
-            q3s.val[0] = vsubq_s8(vreinterpretq_s8_u8(veorq_u8(vs.val[0], vreinterpretq_u8_u32(aux32x4_0))), vreinterpretq_s8_u8(vs.val[0]));
-            q3s.val[1] = vsubq_s8(vreinterpretq_s8_u8(veorq_u8(vs.val[1], vreinterpretq_u8_u32(aux32x4_1))), vreinterpretq_s8_u8(vs.val[1]));
+            q3s.val[0] = vmulq_s8(vreinterpretq_s8_u8(vs.val[0]), vreinterpretq_s8_u32(aux32x4_0));
+            q3s.val[1] = vmulq_s8(vreinterpretq_s8_u8(vs.val[1]), vreinterpretq_s8_u32(aux32x4_1));
 
             vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[2] | (signs[3] << 16)));
             vs.val[1] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2);
             vs.val[0] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2);
-            vs.val[0] = vceqq_u8(vs.val[0], mask2);
-            vs.val[1] = vceqq_u8(vs.val[1], mask2);
+            vs.val[0] = vorrq_u8(vceqq_u8(vs.val[0], mask2), m1);
+            vs.val[1] = vorrq_u8(vceqq_u8(vs.val[1], mask2), m1);
 
             signs += 4;
 
-            q3s.val[2] = vsubq_s8(vreinterpretq_s8_u8(veorq_u8(vs.val[0], vreinterpretq_u8_u32(aux32x4_2))), vreinterpretq_s8_u8(vs.val[0]));
-            q3s.val[3] = vsubq_s8(vreinterpretq_s8_u8(veorq_u8(vs.val[1], vreinterpretq_u8_u32(aux32x4_3))), vreinterpretq_s8_u8(vs.val[1]));
+            q3s.val[2] = vmulq_s8(vreinterpretq_s8_u8(vs.val[0]), vreinterpretq_s8_u32(aux32x4_2));
+            q3s.val[3] = vmulq_s8(vreinterpretq_s8_u8(vs.val[1]), vreinterpretq_s8_u32(aux32x4_3));
 
             const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q3s.val[0], q8b.val[0]), q3s.val[1], q8b.val[1]);
             const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q3s.val[2], q8b.val[2]), q3s.val[3], q8b.val[3]);
+#if QK_K == 256
+            sumi1 += vaddvq_s32(p1) * scales8[ib32/2+0];
+            sumi2 += vaddvq_s32(p2) * scales8[ib32/2+4];
+#else
             sumi1 += vaddvq_s32(p1) * (1 + 2*(x[i].scales[ib32/2] & 0xf));
             sumi2 += vaddvq_s32(p2) * (1 + 2*(x[i].scales[ib32/2] >>  4));
+#endif
         }
         sumf += d*(sumi1 + sumi2);
     }
-    *s = 0.25f * sumf;
+    *s = sumf;
 
 #elif defined(__AVX2__)
 
@@ -10164,6 +10197,16 @@ void ggml_vec_dot_iq3_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const v
     const __m256i mask1 = _mm256_loadu_si256((const __m256i*)k_mask1);
     const __m256i mask2 = _mm256_loadu_si256((const __m256i*)k_mask2);
 
+    const __m256i idx_shift = _mm256_set_epi32(1, 2, 3, 4, 5, 6, 7, 8);
+    const __m256i idx_mask  = _mm256_set1_epi32(256);
+
+    typedef union {
+        __m256i  vec[2];
+        uint32_t index[16];
+    } index_t;
+
+    index_t idx;
+
     __m256 accumf = _mm256_setzero_ps();
     for (int i = 0; i < nb; ++i) {
         const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d;
@@ -10176,24 +10219,25 @@ void ggml_vec_dot_iq3_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const v
         for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) {
             const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32;
             const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32;
-            const __m256i q2_1 = _mm256_set_epi32(iq3xs_grid[qs[7] | ((qh[ib32+0] << 1) & 256)],
-                                                  iq3xs_grid[qs[6] | ((qh[ib32+0] << 2) & 256)],
-                                                  iq3xs_grid[qs[5] | ((qh[ib32+0] << 3) & 256)],
-                                                  iq3xs_grid[qs[4] | ((qh[ib32+0] << 4) & 256)],
-                                                  iq3xs_grid[qs[3] | ((qh[ib32+0] << 5) & 256)],
-                                                  iq3xs_grid[qs[2] | ((qh[ib32+0] << 6) & 256)],
-                                                  iq3xs_grid[qs[1] | ((qh[ib32+0] << 7) & 256)],
-                                                  iq3xs_grid[qs[0] | ((qh[ib32+0] << 8) & 256)]);
-            qs += 8;
-            const __m256i q2_2 = _mm256_set_epi32(iq3xs_grid[qs[7] | ((qh[ib32+1] << 1) & 256)],
-                                                  iq3xs_grid[qs[6] | ((qh[ib32+1] << 2) & 256)],
-                                                  iq3xs_grid[qs[5] | ((qh[ib32+1] << 3) & 256)],
-                                                  iq3xs_grid[qs[4] | ((qh[ib32+1] << 4) & 256)],
-                                                  iq3xs_grid[qs[3] | ((qh[ib32+1] << 5) & 256)],
-                                                  iq3xs_grid[qs[2] | ((qh[ib32+1] << 6) & 256)],
-                                                  iq3xs_grid[qs[1] | ((qh[ib32+1] << 7) & 256)],
-                                                  iq3xs_grid[qs[0] | ((qh[ib32+1] << 8) & 256)]);
-            qs += 8;
+            const __m256i idx_l = _mm256_cvtepu8_epi16(_mm_loadu_si128((const __m128i *)qs)); qs += 16;
+            idx.vec[0] = _mm256_set1_epi32(qh[ib32+0]);
+            idx.vec[1] = _mm256_set1_epi32(qh[ib32+1]);
+            idx.vec[0] = _mm256_and_si256(_mm256_sllv_epi32(idx.vec[0], idx_shift), idx_mask);
+            idx.vec[1] = _mm256_and_si256(_mm256_sllv_epi32(idx.vec[1], idx_shift), idx_mask);
+            idx.vec[0] = _mm256_or_si256(idx.vec[0], _mm256_cvtepi16_epi32(_mm256_castsi256_si128(idx_l)));
+            idx.vec[1] = _mm256_or_si256(idx.vec[1], _mm256_cvtepi16_epi32(_mm256_extractf128_si256(idx_l, 1)));
+
+            // At leat on my CPU (Ryzen 7950X), using _mm256_i32gather_epi32 is slower than _mm256_set_epi32. Strange.
+            //const __m256i q2_1 = _mm256_i32gather_epi32((const int *)iq3s_grid, idx.vec[0], 4);
+            //const __m256i q2_2 = _mm256_i32gather_epi32((const int *)iq3s_grid, idx.vec[1], 4);
+            const __m256i q2_1 = _mm256_set_epi32(
+                    iq3s_grid[idx.index[7]], iq3s_grid[idx.index[6]], iq3s_grid[idx.index[5]], iq3s_grid[idx.index[4]],
+                    iq3s_grid[idx.index[3]], iq3s_grid[idx.index[2]], iq3s_grid[idx.index[1]], iq3s_grid[idx.index[0]]
+            );
+            const __m256i q2_2 = _mm256_set_epi32(
+                    iq3s_grid[idx.index[15]], iq3s_grid[idx.index[14]], iq3s_grid[idx.index[13]], iq3s_grid[idx.index[12]],
+                    iq3s_grid[idx.index[11]], iq3s_grid[idx.index[10]], iq3s_grid[idx.index[ 9]], iq3s_grid[idx.index[ 8]]
+            );
 
             __m256i aux256 = _mm256_set1_epi32(signs[0] | (signs[1] << 16));
             aux256 = _mm256_and_si256(_mm256_shuffle_epi8(aux256,mask1), mask2);
@@ -10221,7 +10265,7 @@ void ggml_vec_dot_iq3_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const v
 
     }
 
-    *s = 0.25f * hsum_float_8(accumf);
+    *s = hsum_float_8(accumf);
 
 #else
 
@@ -10238,8 +10282,8 @@ void ggml_vec_dot_iq3_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const v
             const uint32_t ls2 = 2*(x[i].scales[ib32/2] >>  4) + 1;
             int32_t sumi = 0;
             for (int l = 0; l < 4; ++l) {
-                const uint8_t * grid1 = (const uint8_t *)(iq3xs_grid + (qs[2*l+0] | ((qh[ib32+0] << (8-2*l)) & 256)));
-                const uint8_t * grid2 = (const uint8_t *)(iq3xs_grid + (qs[2*l+1] | ((qh[ib32+0] << (7-2*l)) & 256)));
+                const uint8_t * grid1 = (const uint8_t *)(iq3s_grid + (qs[2*l+0] | ((qh[ib32+0] << (8-2*l)) & 256)));
+                const uint8_t * grid2 = (const uint8_t *)(iq3s_grid + (qs[2*l+1] | ((qh[ib32+0] << (7-2*l)) & 256)));
                 for (int j = 0; j < 4; ++j) {
                     sumi += grid1[j] * q8[j+0] * (signs[l] & kmask_iq2xs[j+0] ? -1 : 1);
                     sumi += grid2[j] * q8[j+4] * (signs[l] & kmask_iq2xs[j+4] ? -1 : 1);
@@ -10251,8 +10295,8 @@ void ggml_vec_dot_iq3_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const v
             bsum += sumi * ls1;
             sumi = 0;
             for (int l = 0; l < 4; ++l) {
-                const uint8_t * grid1 = (const uint8_t *)(iq3xs_grid + (qs[2*l+0] | ((qh[ib32+1] << (8-2*l)) & 256)));
-                const uint8_t * grid2 = (const uint8_t *)(iq3xs_grid + (qs[2*l+1] | ((qh[ib32+1] << (7-2*l)) & 256)));
+                const uint8_t * grid1 = (const uint8_t *)(iq3s_grid + (qs[2*l+0] | ((qh[ib32+1] << (8-2*l)) & 256)));
+                const uint8_t * grid2 = (const uint8_t *)(iq3s_grid + (qs[2*l+1] | ((qh[ib32+1] << (7-2*l)) & 256)));
                 for (int j = 0; j < 4; ++j) {
                     sumi += grid1[j] * q8[j+0] * (signs[l] & kmask_iq2xs[j+0] ? -1 : 1);
                     sumi += grid2[j] * q8[j+4] * (signs[l] & kmask_iq2xs[j+4] ? -1 : 1);
@@ -10265,7 +10309,7 @@ void ggml_vec_dot_iq3_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const v
         }
         sumf += d * bsum;
     }
-    *s = 0.25f * sumf;
+    *s = sumf;
 #endif
 }
 
@@ -10508,10 +10552,10 @@ void ggml_vec_dot_iq4_nl_q8_0(int n, float * restrict s, size_t bs, const void *
         const __m128i q4bits_2 = _mm_loadu_si128((const __m128i*)x[1].qs);
         const __m256i q8b_1 = _mm256_loadu_si256((const __m256i *)y[0].qs);
         const __m256i q8b_2 = _mm256_loadu_si256((const __m256i *)y[1].qs);
-        const __m256i q4b_1 = _mm256_set_m128i(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_1, 4), m4b)),
-                                               _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_1, m4b)));
-        const __m256i q4b_2 = _mm256_set_m128i(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_2, 4), m4b)),
-                                               _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_2, m4b)));
+        const __m256i q4b_1 = MM256_SET_M128I(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_1, 4), m4b)),
+                                              _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_1, m4b)));
+        const __m256i q4b_2 = MM256_SET_M128I(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_2, 4), m4b)),
+                                              _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_2, m4b)));
         const __m256i p16_1 = mul_add_epi8(q4b_1, q8b_1);
         const __m256i p16_2 = mul_add_epi8(q4b_2, q8b_2);
         const __m256i p_1 = _mm256_madd_epi16(p16_1, mone);
@@ -10618,10 +10662,10 @@ void ggml_vec_dot_iq4_xs_q8_K(int n, float * restrict s, size_t bs, const void *
             const __m128i q4bits_2 = _mm_loadu_si128((const __m128i*)qs);  qs += 16;
             const __m256i q8b_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32;
             const __m256i q8b_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32;
-            const __m256i q4b_1 = _mm256_set_m128i(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_1, 4), m4b)),
-                                                   _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_1, m4b)));
-            const __m256i q4b_2 = _mm256_set_m128i(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_2, 4), m4b)),
-                                                   _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_2, m4b)));
+            const __m256i q4b_1 = MM256_SET_M128I(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_1, 4), m4b)),
+                                                  _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_1, m4b)));
+            const __m256i q4b_2 = MM256_SET_M128I(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_2, 4), m4b)),
+                                                  _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_2, m4b)));
             const __m256i p16_1 = mul_add_epi8(q4b_1, q8b_1);
             const __m256i p16_2 = mul_add_epi8(q4b_2, q8b_2);
             const int16_t ls1 = ((x[ibl].scales_l[ib/2] & 0xf) | ((sh << 4) & 0x30)) - 32;
@@ -11912,7 +11956,8 @@ static void quantize_row_iq3_s_impl(int block_size, const float * restrict x, vo
             }
             float best = 0;
             float scale = max/(2*kMaxQ-1);
-            for (int is = -15; is <= 15; ++is) {
+            for (int k = 0; k < bs4; ++k) is_on_grid[k] = false;
+            for (int is = -9; is <= 9; ++is) {
                 float id = (2*kMaxQ-1+is*0.2f)/max;
                 float this_scale = 1/id;
                 for (int k = 0; k < bs4; ++k) {
@@ -11948,7 +11993,7 @@ static void quantize_row_iq3_s_impl(int block_size, const float * restrict x, vo
             if (n_not_ongrid > 0 && scale > 0) {
                 float id = 1/scale;
                 for (int k = 0; k < bs4; ++k) {
-                    if (is_on_grid[k]) continue;
+                    //if (is_on_grid[k]) continue;
                     uint16_t u = 0;
                     for (int i = 0; i < 4; ++i) {
                         int l = nearest_int(0.5f*(id*xval[4*k+i]-1));
@@ -12004,7 +12049,7 @@ static void quantize_row_iq3_s_impl(int block_size, const float * restrict x, vo
         }
 
         float d = max_scale/31;
-        y[ibl].d = GGML_FP32_TO_FP16(d);
+        y[ibl].d = GGML_FP32_TO_FP16(d * 1.033f);
         float id = 1/d;
         for (int ib = 0; ib < QK_K/block_size; ib += 2) {
             int l1 = nearest_int(0.5f*(id*scales[ib+0]-1));