finish i2_s/i8_s vec_dot x86 simd
This commit is contained in:
parent
95dced07e4
commit
569a03ed97
3 changed files with 111 additions and 112 deletions
128
ggml-common.h
128
ggml-common.h
|
@ -1023,70 +1023,70 @@ GGML_TABLE_BEGIN(uint32_t, iq3s_grid, 512)
|
|||
GGML_TABLE_END()
|
||||
|
||||
GGML_TABLE_BEGIN(uint32_t, i2s_i8s, 256)
|
||||
0x00000000, 0x01000000, 0x00000000, 0xff000000,
|
||||
0x00010000, 0x01010000, 0x00010000, 0xff010000,
|
||||
0x00000000, 0x01000000, 0x00000000, 0xff000000,
|
||||
0x00ff0000, 0x01ff0000, 0x00ff0000, 0xffff0000,
|
||||
0x00000100, 0x01000100, 0x00000100, 0xff000100,
|
||||
0x00010100, 0x01010100, 0x00010100, 0xff010100,
|
||||
0x00000100, 0x01000100, 0x00000100, 0xff000100,
|
||||
0x00ff0100, 0x01ff0100, 0x00ff0100, 0xffff0100,
|
||||
0x00000000, 0x01000000, 0x00000000, 0xff000000,
|
||||
0x00010000, 0x01010000, 0x00010000, 0xff010000,
|
||||
0x00000000, 0x01000000, 0x00000000, 0xff000000,
|
||||
0x00ff0000, 0x01ff0000, 0x00ff0000, 0xffff0000,
|
||||
0x0000ff00, 0x0100ff00, 0x0000ff00, 0xff00ff00,
|
||||
0x0001ff00, 0x0101ff00, 0x0001ff00, 0xff01ff00,
|
||||
0x0000ff00, 0x0100ff00, 0x0000ff00, 0xff00ff00,
|
||||
0x00ffff00, 0x01ffff00, 0x00ffff00, 0xffffff00,
|
||||
0x00000001, 0x01000001, 0x00000001, 0xff000001,
|
||||
0x00010001, 0x01010001, 0x00010001, 0xff010001,
|
||||
0x00000001, 0x01000001, 0x00000001, 0xff000001,
|
||||
0x00ff0001, 0x01ff0001, 0x00ff0001, 0xffff0001,
|
||||
0x00000101, 0x01000101, 0x00000101, 0xff000101,
|
||||
0x00010101, 0x01010101, 0x00010101, 0xff010101,
|
||||
0x00000101, 0x01000101, 0x00000101, 0xff000101,
|
||||
0x00ff0101, 0x01ff0101, 0x00ff0101, 0xffff0101,
|
||||
0x00000001, 0x01000001, 0x00000001, 0xff000001,
|
||||
0x00010001, 0x01010001, 0x00010001, 0xff010001,
|
||||
0x00000001, 0x01000001, 0x00000001, 0xff000001,
|
||||
0x00ff0001, 0x01ff0001, 0x00ff0001, 0xffff0001,
|
||||
0x0000ff01, 0x0100ff01, 0x0000ff01, 0xff00ff01,
|
||||
0x0001ff01, 0x0101ff01, 0x0001ff01, 0xff01ff01,
|
||||
0x0000ff01, 0x0100ff01, 0x0000ff01, 0xff00ff01,
|
||||
0x00ffff01, 0x01ffff01, 0x00ffff01, 0xffffff01,
|
||||
0x00000000, 0x01000000, 0x00000000, 0xff000000,
|
||||
0x00010000, 0x01010000, 0x00010000, 0xff010000,
|
||||
0x00000000, 0x01000000, 0x00000000, 0xff000000,
|
||||
0x00ff0000, 0x01ff0000, 0x00ff0000, 0xffff0000,
|
||||
0x00000100, 0x01000100, 0x00000100, 0xff000100,
|
||||
0x00010100, 0x01010100, 0x00010100, 0xff010100,
|
||||
0x00000100, 0x01000100, 0x00000100, 0xff000100,
|
||||
0x00ff0100, 0x01ff0100, 0x00ff0100, 0xffff0100,
|
||||
0x00000000, 0x01000000, 0x00000000, 0xff000000,
|
||||
0x00010000, 0x01010000, 0x00010000, 0xff010000,
|
||||
0x00000000, 0x01000000, 0x00000000, 0xff000000,
|
||||
0x00ff0000, 0x01ff0000, 0x00ff0000, 0xffff0000,
|
||||
0x0000ff00, 0x0100ff00, 0x0000ff00, 0xff00ff00,
|
||||
0x0001ff00, 0x0101ff00, 0x0001ff00, 0xff01ff00,
|
||||
0x0000ff00, 0x0100ff00, 0x0000ff00, 0xff00ff00,
|
||||
0x00ffff00, 0x01ffff00, 0x00ffff00, 0xffffff00,
|
||||
0x000000ff, 0x010000ff, 0x000000ff, 0xff0000ff,
|
||||
0x000100ff, 0x010100ff, 0x000100ff, 0xff0100ff,
|
||||
0x000000ff, 0x010000ff, 0x000000ff, 0xff0000ff,
|
||||
0x00ff00ff, 0x01ff00ff, 0x00ff00ff, 0xffff00ff,
|
||||
0x000001ff, 0x010001ff, 0x000001ff, 0xff0001ff,
|
||||
0x000101ff, 0x010101ff, 0x000101ff, 0xff0101ff,
|
||||
0x000001ff, 0x010001ff, 0x000001ff, 0xff0001ff,
|
||||
0x00ff01ff, 0x01ff01ff, 0x00ff01ff, 0xffff01ff,
|
||||
0x000000ff, 0x010000ff, 0x000000ff, 0xff0000ff,
|
||||
0x000100ff, 0x010100ff, 0x000100ff, 0xff0100ff,
|
||||
0x000000ff, 0x010000ff, 0x000000ff, 0xff0000ff,
|
||||
0x00ff00ff, 0x01ff00ff, 0x00ff00ff, 0xffff00ff,
|
||||
0x0000ffff, 0x0100ffff, 0x0000ffff, 0xff00ffff,
|
||||
0x0001ffff, 0x0101ffff, 0x0001ffff, 0xff01ffff,
|
||||
0x0000ffff, 0x0100ffff, 0x0000ffff, 0xff00ffff,
|
||||
0x00ffffff, 0x01ffffff, 0x00ffffff, 0xffffffff,
|
||||
0x00000000, 0x01000000, 0x00000000, 0xff000000,
|
||||
0x00010000, 0x01010000, 0x00010000, 0xff010000,
|
||||
0x00000000, 0x01000000, 0x00000000, 0xff000000,
|
||||
0x00ff0000, 0x01ff0000, 0x00ff0000, 0xffff0000,
|
||||
0x00000100, 0x01000100, 0x00000100, 0xff000100,
|
||||
0x00010100, 0x01010100, 0x00010100, 0xff010100,
|
||||
0x00000100, 0x01000100, 0x00000100, 0xff000100,
|
||||
0x00ff0100, 0x01ff0100, 0x00ff0100, 0xffff0100,
|
||||
0x00000000, 0x01000000, 0x00000000, 0xff000000,
|
||||
0x00010000, 0x01010000, 0x00010000, 0xff010000,
|
||||
0x00000000, 0x01000000, 0x00000000, 0xff000000,
|
||||
0x00ff0000, 0x01ff0000, 0x00ff0000, 0xffff0000,
|
||||
0x0000ff00, 0x0100ff00, 0x0000ff00, 0xff00ff00,
|
||||
0x0001ff00, 0x0101ff00, 0x0001ff00, 0xff01ff00,
|
||||
0x0000ff00, 0x0100ff00, 0x0000ff00, 0xff00ff00,
|
||||
0x00ffff00, 0x01ffff00, 0x00ffff00, 0xffffff00,
|
||||
0x00000001, 0x01000001, 0x00000001, 0xff000001,
|
||||
0x00010001, 0x01010001, 0x00010001, 0xff010001,
|
||||
0x00000001, 0x01000001, 0x00000001, 0xff000001,
|
||||
0x00ff0001, 0x01ff0001, 0x00ff0001, 0xffff0001,
|
||||
0x00000101, 0x01000101, 0x00000101, 0xff000101,
|
||||
0x00010101, 0x01010101, 0x00010101, 0xff010101,
|
||||
0x00000101, 0x01000101, 0x00000101, 0xff000101,
|
||||
0x00ff0101, 0x01ff0101, 0x00ff0101, 0xffff0101,
|
||||
0x00000001, 0x01000001, 0x00000001, 0xff000001,
|
||||
0x00010001, 0x01010001, 0x00010001, 0xff010001,
|
||||
0x00000001, 0x01000001, 0x00000001, 0xff000001,
|
||||
0x00ff0001, 0x01ff0001, 0x00ff0001, 0xffff0001,
|
||||
0x0000ff01, 0x0100ff01, 0x0000ff01, 0xff00ff01,
|
||||
0x0001ff01, 0x0101ff01, 0x0001ff01, 0xff01ff01,
|
||||
0x0000ff01, 0x0100ff01, 0x0000ff01, 0xff00ff01,
|
||||
0x00ffff01, 0x01ffff01, 0x00ffff01, 0xffffff01,
|
||||
0x00000000, 0x01000000, 0x00000000, 0xff000000,
|
||||
0x00010000, 0x01010000, 0x00010000, 0xff010000,
|
||||
0x00000000, 0x01000000, 0x00000000, 0xff000000,
|
||||
0x00ff0000, 0x01ff0000, 0x00ff0000, 0xffff0000,
|
||||
0x00000100, 0x01000100, 0x00000100, 0xff000100,
|
||||
0x00010100, 0x01010100, 0x00010100, 0xff010100,
|
||||
0x00000100, 0x01000100, 0x00000100, 0xff000100,
|
||||
0x00ff0100, 0x01ff0100, 0x00ff0100, 0xffff0100,
|
||||
0x00000000, 0x01000000, 0x00000000, 0xff000000,
|
||||
0x00010000, 0x01010000, 0x00010000, 0xff010000,
|
||||
0x00000000, 0x01000000, 0x00000000, 0xff000000,
|
||||
0x00ff0000, 0x01ff0000, 0x00ff0000, 0xffff0000,
|
||||
0x0000ff00, 0x0100ff00, 0x0000ff00, 0xff00ff00,
|
||||
0x0001ff00, 0x0101ff00, 0x0001ff00, 0xff01ff00,
|
||||
0x0000ff00, 0x0100ff00, 0x0000ff00, 0xff00ff00,
|
||||
0x00ffff00, 0x01ffff00, 0x00ffff00, 0xffffff00,
|
||||
0x000000ff, 0x010000ff, 0x000000ff, 0xff0000ff,
|
||||
0x000100ff, 0x010100ff, 0x000100ff, 0xff0100ff,
|
||||
0x000000ff, 0x010000ff, 0x000000ff, 0xff0000ff,
|
||||
0x00ff00ff, 0x01ff00ff, 0x00ff00ff, 0xffff00ff,
|
||||
0x000001ff, 0x010001ff, 0x000001ff, 0xff0001ff,
|
||||
0x000101ff, 0x010101ff, 0x000101ff, 0xff0101ff,
|
||||
0x000001ff, 0x010001ff, 0x000001ff, 0xff0001ff,
|
||||
0x00ff01ff, 0x01ff01ff, 0x00ff01ff, 0xffff01ff,
|
||||
0x000000ff, 0x010000ff, 0x000000ff, 0xff0000ff,
|
||||
0x000100ff, 0x010100ff, 0x000100ff, 0xff0100ff,
|
||||
0x000000ff, 0x010000ff, 0x000000ff, 0xff0000ff,
|
||||
0x00ff00ff, 0x01ff00ff, 0x00ff00ff, 0xffff00ff,
|
||||
0x0000ffff, 0x0100ffff, 0x0000ffff, 0xff00ffff,
|
||||
0x0001ffff, 0x0101ffff, 0x0001ffff, 0xff01ffff,
|
||||
0x0000ffff, 0x0100ffff, 0x0000ffff, 0xff00ffff,
|
||||
0x00ffffff, 0x01ffffff, 0x00ffffff, 0xffffffff,
|
||||
GGML_TABLE_END()
|
||||
|
||||
#define NGRID_IQ1S 2048
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue