IQ2_XS_R4 (#155)

* iq2_xs_r4: Zen4

* iq2_xs_r4: AVX2

* iq2_xs_r4: slightly better matrix x vector on AVX2

* iq2_xs_r4: NEON - not much better than iq2_xs

* iq2_xs_r4: slightly better NEON

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>

1 files changed, 7 insertions, 0 deletions

diff --git a/ggml/src/ggml-common.h b/ggml/src/ggml-common.h
index ad21dd50..2534b461 100644
--- a/ggml/src/ggml-common.h
+++ b/ggml/src/ggml-common.h
@@ -412,6 +412,13 @@ typedef struct {
 } block_iq2_xs;
 static_assert(sizeof(block_iq2_xs) == sizeof(ggml_half) + QK_K/8*sizeof(uint16_t) + QK_K/32, "wrong iq2_xs block size/padding");
 
+typedef struct {
+    ggml_half d[4];
+    uint16_t qs[QK_K/2];
+    uint8_t  scales[QK_K/8];
+} block_iq2_xs_r4;
+static_assert(sizeof(block_iq2_xs_r4) == 4*sizeof(block_iq2_xs), "wrong iq2_xs_r4 block size/padding");
+
 // 2.5625 bpw quants
 typedef struct {
     ggml_half d;