root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-24 18:34:36 +00:00
Code Issues Actions 7 Packages Projects Releases Wiki Activity

ggml : SOTA 2-bit quants (add IQ2_XS) (#4856)

Browse Source 
* iq2_xs: basics

* iq2_xs: this should have been in the basics

* iq2_xs: CUDA and scalar CPU works

* iq2_xs: WIP Metal

* iq2_xs: Metal now works

* iq2_xs: working, but dog slow, ARM_NEON dot product

* iq2_xs: better ARM_NEON dot product

We are now at 19.5 t/s for TG-128 and 61 t/s for PP-512 when
running on the CPU.

* iq2_xs: AVX2 dot product - 19.5 t/s

* iq2_xs: faster AVX2 dit product

21.4 t/s for TG-128, 59.2 t/s for PP-512.
The latter is 2x compared to the previous version.

* iq2_xs: had forgotten to delete iq2-data.h

* Add llama enum for IQ2_XS

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This commit is contained in:
Kawrakow 2024-01-11 20:39:39 +01:00 committed by GitHub
parent 3ba5b8ca8e
commit 49662cbed3
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
10 changed files with 1038 additions and 28 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

232
ggml-cuda.cu
View File

@ -486,6 +486,15 @@ typedef struct {
} block_iq2_xxs; } block_iq2_xxs;
static_assert(sizeof(block_iq2_xxs) == sizeof(ggml_fp16_t) + QK_K/8*sizeof(uint16_t), "wrong iq2_xxs block size/padding"); static_assert(sizeof(block_iq2_xxs) == sizeof(ggml_fp16_t) + QK_K/8*sizeof(uint16_t), "wrong iq2_xxs block size/padding");
#define QR2_XS 8
#define QI2_XS (QK_K / (4*QR2_XS))
typedef struct {
half d;
uint16_t qs[QK_K/8];
uint8_t scales[QK_K/32];
} block_iq2_xs;
static_assert(sizeof(block_iq2_xs) == sizeof(ggml_fp16_t) + QK_K/8*sizeof(uint16_t) + QK_K/32, "wrong iq2_xs block size/padding");
#define WARP_SIZE 32 #define WARP_SIZE 32
#define MATRIX_ROW_PADDING 512 // last row of quant. matrices is a multiple of this to avoid out-of-bounds memory accesses #define MATRIX_ROW_PADDING 512 // last row of quant. matrices is a multiple of this to avoid out-of-bounds memory accesses
@ -1328,7 +1337,7 @@ static __global__ void dequantize_block_q6_K(const void * __restrict__ vx, dst_t
#endif #endif
} }
static const __device__ uint64_t kgrid_iq2xxs[256] = { static const __device__ uint64_t iq2xxs_grid[256] = {
0x0808080808080808, 0x080808080808082b, 0x0808080808081919, 0x0808080808082b08, 0x0808080808080808, 0x080808080808082b, 0x0808080808081919, 0x0808080808082b08,
0x0808080808082b2b, 0x0808080808190819, 0x0808080808191908, 0x08080808082b0808, 0x0808080808082b2b, 0x0808080808190819, 0x0808080808191908, 0x08080808082b0808,
0x08080808082b082b, 0x08080808082b2b08, 0x08080808082b2b2b, 0x0808080819080819, 0x08080808082b082b, 0x08080808082b2b08, 0x08080808082b2b2b, 0x0808080819080819,
@ -1395,6 +1404,137 @@ static const __device__ uint64_t kgrid_iq2xxs[256] = {
0x2b2b082b08080808, 0x2b2b190808192b08, 0x2b2b2b0819190808, 0x2b2b2b1908081908, 0x2b2b082b08080808, 0x2b2b190808192b08, 0x2b2b2b0819190808, 0x2b2b2b1908081908,
}; };
static const __device__ uint64_t iq2xs_grid[512] = {
0x0808080808080808, 0x080808080808082b, 0x0808080808081919, 0x0808080808082b08,
0x0808080808082b2b, 0x0808080808190819, 0x0808080808191908, 0x080808080819192b,
0x0808080808192b19, 0x08080808082b0808, 0x08080808082b082b, 0x08080808082b1919,
0x08080808082b2b08, 0x0808080819080819, 0x0808080819081908, 0x080808081908192b,
0x0808080819082b19, 0x0808080819190808, 0x080808081919082b, 0x0808080819191919,
0x0808080819192b08, 0x08080808192b0819, 0x08080808192b1908, 0x080808082b080808,
0x080808082b08082b, 0x080808082b081919, 0x080808082b082b08, 0x080808082b190819,
0x080808082b191908, 0x080808082b192b19, 0x080808082b2b0808, 0x0808081908080819,
0x0808081908081908, 0x080808190808192b, 0x0808081908082b19, 0x0808081908190808,
0x080808190819082b, 0x0808081908191919, 0x0808081908192b08, 0x0808081908192b2b,
0x08080819082b0819, 0x08080819082b1908, 0x0808081919080808, 0x080808191908082b,
0x0808081919081919, 0x0808081919082b08, 0x0808081919190819, 0x0808081919191908,
0x08080819192b0808, 0x08080819192b2b08, 0x080808192b080819, 0x080808192b081908,
0x080808192b190808, 0x0808082b08080808, 0x0808082b0808082b, 0x0808082b08081919,
0x0808082b08082b08, 0x0808082b08190819, 0x0808082b08191908, 0x0808082b082b0808,
0x0808082b19080819, 0x0808082b19081908, 0x0808082b19190808, 0x0808082b19191919,
0x0808082b2b080808, 0x0808082b2b082b2b, 0x0808190808080819, 0x0808190808081908,
0x080819080808192b, 0x0808190808082b19, 0x0808190808190808, 0x080819080819082b,
0x0808190808191919, 0x0808190808192b08, 0x08081908082b0819, 0x08081908082b1908,
0x0808190819080808, 0x080819081908082b, 0x0808190819081919, 0x0808190819082b08,
0x0808190819190819, 0x0808190819191908, 0x080819081919192b, 0x08081908192b0808,
0x080819082b080819, 0x080819082b081908, 0x080819082b190808, 0x0808191908080808,
0x080819190808082b, 0x0808191908081919, 0x0808191908082b08, 0x0808191908190819,
0x0808191908191908, 0x08081919082b0808, 0x0808191919080819, 0x0808191919081908,
0x0808191919190808, 0x08081919192b0819, 0x080819192b080808, 0x0808192b08080819,
0x0808192b08081908, 0x0808192b08190808, 0x0808192b082b192b, 0x0808192b19080808,
0x0808192b1908082b, 0x0808192b2b081908, 0x08082b0808080808, 0x08082b080808082b,
0x08082b0808081919, 0x08082b0808082b08, 0x08082b0808082b2b, 0x08082b0808190819,
0x08082b0808191908, 0x08082b08082b0808, 0x08082b08082b1919, 0x08082b0819080819,
0x08082b0819081908, 0x08082b0819190808, 0x08082b0819192b08, 0x08082b082b080808,
0x08082b082b2b0808, 0x08082b082b2b2b2b, 0x08082b1908080819, 0x08082b1908081908,
0x08082b1908190808, 0x08082b1919080808, 0x08082b192b080819, 0x08082b192b082b19,
0x08082b2b08080808, 0x08082b2b082b0808, 0x08082b2b082b2b08, 0x08082b2b2b19192b,
0x08082b2b2b2b0808, 0x0819080808080819, 0x0819080808081908, 0x081908080808192b,
0x0819080808082b19, 0x0819080808190808, 0x081908080819082b, 0x0819080808191919,
0x0819080808192b08, 0x08190808082b0819, 0x08190808082b1908, 0x0819080819080808,
0x081908081908082b, 0x0819080819081919, 0x0819080819082b08, 0x0819080819190819,
0x0819080819191908, 0x08190808192b0808, 0x08190808192b2b2b, 0x081908082b080819,
0x081908082b081908, 0x081908082b190808, 0x0819081908080808, 0x081908190808082b,
0x0819081908081919, 0x0819081908082b08, 0x0819081908190819, 0x0819081908191908,
0x08190819082b0808, 0x0819081919080819, 0x0819081919081908, 0x0819081919190808,
0x081908192b080808, 0x081908192b191908, 0x081908192b19192b, 0x0819082b08080819,
0x0819082b08081908, 0x0819082b0808192b, 0x0819082b08190808, 0x0819082b19080808,
0x0819082b192b0808, 0x0819190808080808, 0x081919080808082b, 0x0819190808081919,
0x0819190808082b08, 0x0819190808190819, 0x0819190808191908, 0x08191908082b0808,
0x0819190819080819, 0x0819190819081908, 0x0819190819082b19, 0x0819190819190808,
0x08191908192b1908, 0x081919082b080808, 0x0819191908080819, 0x0819191908081908,
0x0819191908190808, 0x0819191919080808, 0x0819192b08080808, 0x0819192b08191908,
0x0819192b19082b19, 0x08192b0808080819, 0x08192b0808081908, 0x08192b0808190808,
0x08192b080819082b, 0x08192b0819080808, 0x08192b0819191908, 0x08192b082b08192b,
0x08192b1908080808, 0x08192b1908081919, 0x08192b19192b192b, 0x08192b2b19190819,
0x08192b2b2b2b2b19, 0x082b080808080808, 0x082b08080808082b, 0x082b080808081919,
0x082b080808082b08, 0x082b080808082b2b, 0x082b080808190819, 0x082b080808191908,
0x082b0808082b0808, 0x082b080819080819, 0x082b080819081908, 0x082b080819190808,
0x082b08082b080808, 0x082b08082b2b0808, 0x082b081908080819, 0x082b081908081908,
0x082b081908190808, 0x082b081919080808, 0x082b081919082b08, 0x082b0819192b1919,
0x082b082b08080808, 0x082b082b082b082b, 0x082b082b2b080808, 0x082b082b2b2b2b08,
0x082b190808080819, 0x082b190808081908, 0x082b190808190808, 0x082b1908082b2b19,
0x082b190819080808, 0x082b191908080808, 0x082b191919080819, 0x082b19191919082b,
0x082b19192b192b19, 0x082b192b08080819, 0x082b192b08192b2b, 0x082b192b2b2b192b,
0x082b2b0808080808, 0x082b2b0808082b08, 0x082b2b0808082b2b, 0x082b2b08082b0808,
0x082b2b0819191919, 0x082b2b082b082b08, 0x082b2b082b2b082b, 0x082b2b19192b2b08,
0x082b2b192b190808, 0x082b2b2b08082b08, 0x082b2b2b082b0808, 0x082b2b2b2b08082b,
0x082b2b2b2b082b08, 0x082b2b2b2b082b2b, 0x1908080808080819, 0x1908080808081908,
0x190808080808192b, 0x1908080808082b19, 0x1908080808190808, 0x190808080819082b,
0x1908080808191919, 0x1908080808192b08, 0x19080808082b0819, 0x19080808082b1908,
0x1908080819080808, 0x190808081908082b, 0x1908080819081919, 0x1908080819082b08,
0x1908080819082b2b, 0x1908080819190819, 0x1908080819191908, 0x19080808192b0808,
0x19080808192b1919, 0x190808082b080819, 0x190808082b081908, 0x190808082b190808,
0x1908081908080808, 0x190808190808082b, 0x1908081908081919, 0x1908081908082b08,
0x1908081908190819, 0x1908081908191908, 0x19080819082b0808, 0x1908081919080819,
0x1908081919081908, 0x1908081919190808, 0x190808192b080808, 0x190808192b081919,
0x190808192b2b082b, 0x1908082b08080819, 0x1908082b08081908, 0x1908082b08190808,
0x1908082b0819082b, 0x1908082b082b2b19, 0x1908082b19080808, 0x1908190808080808,
0x190819080808082b, 0x1908190808081919, 0x1908190808082b08, 0x1908190808190819,
0x1908190808191908, 0x1908190808192b19, 0x19081908082b0808, 0x1908190819080819,
0x1908190819081908, 0x1908190819190808, 0x190819082b080808, 0x190819082b191908,
0x1908191908080819, 0x1908191908081908, 0x1908191908190808, 0x19081919082b1908,
0x1908191919080808, 0x190819192b192b2b, 0x1908192b08080808, 0x1908192b08082b2b,
0x1908192b19081908, 0x1908192b19190808, 0x19082b0808080819, 0x19082b0808081908,
0x19082b0808190808, 0x19082b0819080808, 0x19082b0819081919, 0x19082b0819191908,
0x19082b08192b082b, 0x19082b1908080808, 0x19082b1908190819, 0x19082b1919081908,
0x19082b1919190808, 0x19082b19192b2b19, 0x19082b2b08081908, 0x1919080808080808,
0x191908080808082b, 0x1919080808081919, 0x1919080808082b08, 0x1919080808190819,
0x1919080808191908, 0x19190808082b0808, 0x19190808082b2b08, 0x1919080819080819,
0x1919080819081908, 0x1919080819190808, 0x191908082b080808, 0x1919081908080819,
0x1919081908081908, 0x1919081908190808, 0x1919081908191919, 0x1919081919080808,
0x191908191908082b, 0x1919082b08080808, 0x1919082b19081908, 0x1919082b2b2b2b2b,
0x1919190808080819, 0x1919190808081908, 0x1919190808190808, 0x19191908082b0819,
0x1919190819080808, 0x19191908192b0808, 0x191919082b080819, 0x191919082b2b0819,
0x1919191908080808, 0x1919191908082b08, 0x191919192b080808, 0x191919192b082b08,
0x1919192b082b0819, 0x1919192b192b2b08, 0x1919192b2b2b0819, 0x19192b0808080808,
0x19192b0808191908, 0x19192b0819080819, 0x19192b0819190808, 0x19192b082b192b19,
0x19192b1908192b2b, 0x19192b1919080808, 0x19192b191908082b, 0x19192b2b2b081919,
0x192b080808080819, 0x192b080808081908, 0x192b080808190808, 0x192b080819080808,
0x192b080819191908, 0x192b0808192b082b, 0x192b08082b08192b, 0x192b08082b2b2b19,
0x192b081908080808, 0x192b082b082b1908, 0x192b082b19082b2b, 0x192b082b2b19082b,
0x192b190808080808, 0x192b19080819192b, 0x192b191908190808, 0x192b191919080808,
0x192b191919081919, 0x192b19192b2b1908, 0x192b2b0808080819, 0x192b2b08192b2b2b,
0x192b2b19082b1919, 0x192b2b2b0808192b, 0x192b2b2b19191908, 0x192b2b2b192b082b,
0x2b08080808080808, 0x2b0808080808082b, 0x2b08080808081919, 0x2b08080808082b08,
0x2b08080808190819, 0x2b08080808191908, 0x2b080808082b0808, 0x2b080808082b2b2b,
0x2b08080819080819, 0x2b08080819081908, 0x2b08080819190808, 0x2b0808082b080808,
0x2b0808082b08082b, 0x2b0808082b2b2b08, 0x2b0808082b2b2b2b, 0x2b08081908080819,
0x2b08081908081908, 0x2b0808190808192b, 0x2b08081908190808, 0x2b08081919080808,
0x2b08081919190819, 0x2b08081919192b19, 0x2b08082b08080808, 0x2b08082b082b0808,
0x2b08082b2b080808, 0x2b08082b2b08082b, 0x2b08082b2b2b0808, 0x2b08082b2b2b2b08,
0x2b08190808080819, 0x2b08190808081908, 0x2b08190808190808, 0x2b0819080819082b,
0x2b08190808191919, 0x2b08190819080808, 0x2b081908192b0808, 0x2b0819082b082b19,
0x2b08191908080808, 0x2b08191919081908, 0x2b0819192b2b1919, 0x2b08192b08192b08,
0x2b08192b192b2b2b, 0x2b082b0808080808, 0x2b082b0808082b08, 0x2b082b08082b1919,
0x2b082b0819192b2b, 0x2b082b082b080808, 0x2b082b082b08082b, 0x2b082b082b2b2b08,
0x2b082b190808192b, 0x2b082b2b082b082b, 0x2b082b2b2b080808, 0x2b082b2b2b082b08,
0x2b082b2b2b19192b, 0x2b082b2b2b2b2b08, 0x2b19080808080819, 0x2b19080808081908,
0x2b19080808190808, 0x2b19080819080808, 0x2b1908081919192b, 0x2b1908082b081908,
0x2b19081908080808, 0x2b190819082b082b, 0x2b190819192b1908, 0x2b19082b1919192b,
0x2b19082b2b082b19, 0x2b19190808080808, 0x2b19190808081919, 0x2b19190819081908,
0x2b19190819190808, 0x2b19190819192b08, 0x2b191919082b2b19, 0x2b1919192b190808,
0x2b1919192b19082b, 0x2b19192b19080819, 0x2b192b0819190819, 0x2b192b082b2b192b,
0x2b192b1919082b19, 0x2b192b2b08191919, 0x2b192b2b192b0808, 0x2b2b080808080808,
0x2b2b08080808082b, 0x2b2b080808082b08, 0x2b2b080808082b2b, 0x2b2b0808082b0808,
0x2b2b0808082b2b2b, 0x2b2b08082b2b0808, 0x2b2b081919190819, 0x2b2b081919192b19,
0x2b2b08192b2b192b, 0x2b2b082b08080808, 0x2b2b082b0808082b, 0x2b2b082b08082b08,
0x2b2b082b082b2b2b, 0x2b2b082b2b080808, 0x2b2b082b2b2b0808, 0x2b2b190819080808,
0x2b2b19082b191919, 0x2b2b192b192b1919, 0x2b2b192b2b192b08, 0x2b2b2b0808082b2b,
0x2b2b2b08082b0808, 0x2b2b2b08082b082b, 0x2b2b2b08082b2b08, 0x2b2b2b082b2b0808,
0x2b2b2b082b2b2b08, 0x2b2b2b1908081908, 0x2b2b2b192b081908, 0x2b2b2b192b08192b,
0x2b2b2b2b082b2b08, 0x2b2b2b2b082b2b2b, 0x2b2b2b2b2b190819, 0x2b2b2b2b2b2b2b2b,
};
static const __device__ uint8_t ksigns_iq2xs[128] = { static const __device__ uint8_t ksigns_iq2xs[128] = {
0, 129, 130, 3, 132, 5, 6, 135, 136, 9, 10, 139, 12, 141, 142, 15, 0, 129, 130, 3, 132, 5, 6, 135, 136, 9, 10, 139, 12, 141, 142, 15,
144, 17, 18, 147, 20, 149, 150, 23, 24, 153, 154, 27, 156, 29, 30, 159, 144, 17, 18, 147, 20, 149, 150, 23, 24, 153, 154, 27, 156, 29, 30, 159,
@ -1439,7 +1579,7 @@ static __global__ void dequantize_block_iq2_xxs(const void * __restrict__ vx, ds
dst_t * y = yy + i*QK_K + 32*ib + 8*il; dst_t * y = yy + i*QK_K + 32*ib + 8*il;
const uint16_t * q2 = x[i].qs + 4*ib; const uint16_t * q2 = x[i].qs + 4*ib;
const uint8_t * aux8 = (const uint8_t *)q2; const uint8_t * aux8 = (const uint8_t *)q2;
const uint8_t * grid = (const uint8_t *)(kgrid_iq2xxs + aux8[il]); const uint8_t * grid = (const uint8_t *)(iq2xxs_grid + aux8[il]);
const uint32_t aux32 = q2[2] | (q2[3] << 16); const uint32_t aux32 = q2[2] | (q2[3] << 16);
const float d = (float)x[i].d * (0.5f + (aux32 >> 28)) * 0.25f; const float d = (float)x[i].d * (0.5f + (aux32 >> 28)) * 0.25f;
const uint8_t signs = ksigns_iq2xs[(aux32 >> 7*il) & 127]; const uint8_t signs = ksigns_iq2xs[(aux32 >> 7*il) & 127];
@ -1450,6 +1590,28 @@ static __global__ void dequantize_block_iq2_xxs(const void * __restrict__ vx, ds
} }
template<typename dst_t>
static __global__ void dequantize_block_iq2_xs(const void * __restrict__ vx, dst_t * __restrict__ yy) {
const int i = blockIdx.x;
const block_iq2_xs * x = (const block_iq2_xs *) vx;
const int tid = threadIdx.x;
#if QK_K == 256
const int il = tid/8; // 0...3
const int ib = tid%8; // 0...7
dst_t * y = yy + i*QK_K + 32*ib + 8*il;
const uint16_t * q2 = x[i].qs + 4*ib;
const uint8_t * grid = (const uint8_t *)(iq2xs_grid + (q2[il] & 511));
const float d = (float)x[i].d * (0.5f + ((x[i].scales[ib] >> 4*(il/2)) & 0xf)) * 0.25f;
const uint8_t signs = ksigns_iq2xs[q2[il] >> 9];
for (int j = 0; j < 8; ++j) y[j] = d * grid[j] * (signs & kmask_iq2xs[j] ? -1.f : 1.f);
#else
assert(false);
#endif
}
static __global__ void dequantize_mul_mat_vec_q2_k(const void * __restrict__ vx, const float * __restrict__ yy, float * __restrict__ dst, const int ncols, int nrows) { static __global__ void dequantize_mul_mat_vec_q2_k(const void * __restrict__ vx, const float * __restrict__ yy, float * __restrict__ dst, const int ncols, int nrows) {
static_assert(16%K_QUANTS_PER_ITERATION == 0, "16 must be divisible by K_QUANTS_PER_ITERATION"); static_assert(16%K_QUANTS_PER_ITERATION == 0, "16 must be divisible by K_QUANTS_PER_ITERATION");
@ -3996,7 +4158,7 @@ static __device__ __forceinline__ float vec_dot_iq2_xxs_q8_1(
uint32_t aux32 = q2[2] | (q2[3] << 16); uint32_t aux32 = q2[2] | (q2[3] << 16);
int sumi = 0; int sumi = 0;
for (int l = 0; l < 4; ++l) { for (int l = 0; l < 4; ++l) {
const uint8_t * grid = (const uint8_t *)(kgrid_iq2xxs + aux8[l]); const uint8_t * grid = (const uint8_t *)(iq2xxs_grid + aux8[l]);
const uint8_t signs = ksigns_iq2xs[aux32 & 127]; const uint8_t signs = ksigns_iq2xs[aux32 & 127];
for (int j = 0; j < 8; ++j) { for (int j = 0; j < 8; ++j) {
sumi += q8[j] * grid[j] * (signs & kmask_iq2xs[j] ? -1 : 1); sumi += q8[j] * grid[j] * (signs & kmask_iq2xs[j] ? -1 : 1);
@ -4012,8 +4174,8 @@ static __device__ __forceinline__ float vec_dot_iq2_xxs_q8_1(
const int il = iqs%2; const int il = iqs%2;
const uint16_t * q2 = bq2->qs + 4*ib32; const uint16_t * q2 = bq2->qs + 4*ib32;
const uint8_t * aux8 = (const uint8_t *)q2; const uint8_t * aux8 = (const uint8_t *)q2;
const uint8_t * grid1 = (const uint8_t *)(kgrid_iq2xxs + aux8[2*il+0]); const uint8_t * grid1 = (const uint8_t *)(iq2xxs_grid + aux8[2*il+0]);
const uint8_t * grid2 = (const uint8_t *)(kgrid_iq2xxs + aux8[2*il+1]); const uint8_t * grid2 = (const uint8_t *)(iq2xxs_grid + aux8[2*il+1]);
const uint32_t aux32 = q2[2] | (q2[3] << 16); const uint32_t aux32 = q2[2] | (q2[3] << 16);
const float d = (float)bq2->d * (0.5f + (aux32 >> 28)) * (float)bq8_1[ib32].ds.x * 0.25f; const float d = (float)bq2->d * (0.5f + (aux32 >> 28)) * (float)bq8_1[ib32].ds.x * 0.25f;
const uint8_t signs1 = ksigns_iq2xs[(aux32 >> 14*il) & 127]; const uint8_t signs1 = ksigns_iq2xs[(aux32 >> 14*il) & 127];
@ -4032,6 +4194,42 @@ static __device__ __forceinline__ float vec_dot_iq2_xxs_q8_1(
#endif #endif
} }
static __device__ __forceinline__ float vec_dot_iq2_xs_q8_1(
const void * __restrict__ vbq, const block_q8_1 * __restrict__ bq8_1, const int & iqs) {
#if QK_K == 256
const block_iq2_xs * bq2 = (const block_iq2_xs *) vbq;
const int ib32 = iqs;
const uint16_t * q2 = bq2->qs + 4*ib32;
const int8_t * q8 = bq8_1[ib32].qs;
const uint8_t ls1 = bq2->scales[ib32] & 0xf;
const uint8_t ls2 = bq2->scales[ib32] >> 4;
int sumi1 = 0;
for (int l = 0; l < 2; ++l) {
const uint8_t * grid = (const uint8_t *)(iq2xs_grid + (q2[l] & 511));
const uint8_t signs = ksigns_iq2xs[q2[l] >> 9];
for (int j = 0; j < 8; ++j) {
sumi1 += q8[j] * grid[j] * (signs & kmask_iq2xs[j] ? -1 : 1);
}
q8 += 8;
}
int sumi2 = 0;
for (int l = 2; l < 4; ++l) {
const uint8_t * grid = (const uint8_t *)(iq2xs_grid + (q2[l] & 511));
const uint8_t signs = ksigns_iq2xs[q2[l] >> 9];
for (int j = 0; j < 8; ++j) {
sumi2 += q8[j] * grid[j] * (signs & kmask_iq2xs[j] ? -1 : 1);
}
q8 += 8;
}
const float d = (float)bq2->d * (float)bq8_1[ib32].ds.x * 0.25f;
return d * ((0.5f + ls1) * sumi1 + (0.5f + ls2) * sumi2);
#else
assert(false);
return 0.f;
#endif
}
template <int qk, int qr, int qi, bool need_sum, typename block_q_t, int mmq_x, int mmq_y, int nwarps, template <int qk, int qr, int qi, bool need_sum, typename block_q_t, int mmq_x, int mmq_y, int nwarps,
allocate_tiles_cuda_t allocate_tiles, load_tiles_cuda_t load_tiles, int vdr, vec_dot_q_mul_mat_cuda_t vec_dot> allocate_tiles_cuda_t allocate_tiles, load_tiles_cuda_t load_tiles, int vdr, vec_dot_q_mul_mat_cuda_t vec_dot>
static __device__ __forceinline__ void mul_mat_q( static __device__ __forceinline__ void mul_mat_q(
@ -6035,6 +6233,12 @@ static void dequantize_row_iq2_xxs_cuda(const void * vx, dst_t * y, const int k,
dequantize_block_iq2_xxs<<<nb, 32, 0, stream>>>(vx, y); dequantize_block_iq2_xxs<<<nb, 32, 0, stream>>>(vx, y);
} }
template<typename dst_t>
static void dequantize_row_iq2_xs_cuda(const void * vx, dst_t * y, const int k, cudaStream_t stream) {
const int nb = k / QK_K;
dequantize_block_iq2_xs<<<nb, 32, 0, stream>>>(vx, y);
}
template <typename src_t, typename dst_t> template <typename src_t, typename dst_t>
static void convert_unary_cuda(const void * __restrict__ vx, dst_t * __restrict__ y, const int k, cudaStream_t stream) { static void convert_unary_cuda(const void * __restrict__ vx, dst_t * __restrict__ y, const int k, cudaStream_t stream) {
const int num_blocks = (k + CUDA_DEQUANTIZE_BLOCK_SIZE - 1) / CUDA_DEQUANTIZE_BLOCK_SIZE; const int num_blocks = (k + CUDA_DEQUANTIZE_BLOCK_SIZE - 1) / CUDA_DEQUANTIZE_BLOCK_SIZE;
@ -6065,6 +6269,8 @@ static to_fp16_cuda_t ggml_get_to_fp16_cuda(ggml_type type) {
return dequantize_row_q6_K_cuda; return dequantize_row_q6_K_cuda;
case GGML_TYPE_IQ2_XXS: case GGML_TYPE_IQ2_XXS:
return dequantize_row_iq2_xxs_cuda; return dequantize_row_iq2_xxs_cuda;
case GGML_TYPE_IQ2_XS:
return dequantize_row_iq2_xs_cuda;
case GGML_TYPE_F32: case GGML_TYPE_F32:
return convert_unary_cuda<float>; return convert_unary_cuda<float>;
default: default:
@ -6096,6 +6302,8 @@ static to_fp32_cuda_t ggml_get_to_fp32_cuda(ggml_type type) {
return dequantize_row_q6_K_cuda; return dequantize_row_q6_K_cuda;
case GGML_TYPE_IQ2_XXS: case GGML_TYPE_IQ2_XXS:
return dequantize_row_iq2_xxs_cuda; return dequantize_row_iq2_xxs_cuda;
case GGML_TYPE_IQ2_XS:
return dequantize_row_iq2_xs_cuda;
case GGML_TYPE_F16: case GGML_TYPE_F16:
return convert_unary_cuda<half>; return convert_unary_cuda<half>;
default: default:
@ -6299,6 +6507,15 @@ static void mul_mat_vec_iq2_xxs_q8_1_cuda(const void * vx, const void * vy, floa
<<<block_nums, block_dims, 0, stream>>>(vx, vy, dst, ncols, nrows); <<<block_nums, block_dims, 0, stream>>>(vx, vy, dst, ncols, nrows);
} }
static void mul_mat_vec_iq2_xs_q8_1_cuda(const void * vx, const void * vy, float * dst, const int ncols, const int nrows, cudaStream_t stream) {
GGML_ASSERT(ncols % QK_K == 0);
const int block_num_y = (nrows + GGML_CUDA_MMV_Y - 1) / GGML_CUDA_MMV_Y;
const dim3 block_nums(block_num_y, 1, 1);
const dim3 block_dims(WARP_SIZE, GGML_CUDA_MMV_Y, 1);
mul_mat_vec_q<QK_K, QI2_XS, block_iq2_xs, 1, vec_dot_iq2_xs_q8_1>
<<<block_nums, block_dims, 0, stream>>>(vx, vy, dst, ncols, nrows);
}
static void ggml_mul_mat_q4_0_q8_1_cuda( static void ggml_mul_mat_q4_0_q8_1_cuda(
const void * vx, const void * vy, float * dst, const int ncols_x, const int nrows_x, const void * vx, const void * vy, float * dst, const int ncols_x, const int nrows_x,
const int ncols_y, const int nrows_y, const int nrows_dst, cudaStream_t stream) { const int ncols_y, const int nrows_y, const int nrows_dst, cudaStream_t stream) {
@ -7871,6 +8088,7 @@ static int64_t get_row_rounding(ggml_type type) {
case GGML_TYPE_Q5_K: case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
case GGML_TYPE_IQ2_XXS: case GGML_TYPE_IQ2_XXS:
case GGML_TYPE_IQ2_XS:
return max_compute_capability >= CC_RDNA2 ? 128 : 64; return max_compute_capability >= CC_RDNA2 ? 128 : 64;
default: default:
GGML_ASSERT(false); GGML_ASSERT(false);
@ -7892,6 +8110,7 @@ static int64_t get_row_rounding(ggml_type type) {
case GGML_TYPE_Q4_K: case GGML_TYPE_Q4_K:
case GGML_TYPE_Q5_K: case GGML_TYPE_Q5_K:
case GGML_TYPE_IQ2_XXS: case GGML_TYPE_IQ2_XXS:
case GGML_TYPE_IQ2_XS:
return max_compute_capability >= CC_VOLTA ? 128 : 64; return max_compute_capability >= CC_VOLTA ? 128 : 64;
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
return 64; return 64;
@ -7945,6 +8164,9 @@ static void ggml_cuda_op_mul_mat_vec_q(
case GGML_TYPE_IQ2_XXS: case GGML_TYPE_IQ2_XXS:
mul_mat_vec_iq2_xxs_q8_1_cuda(src0_dd_i, src1_ddq_i, dst_dd_i, ne00, row_diff, stream); mul_mat_vec_iq2_xxs_q8_1_cuda(src0_dd_i, src1_ddq_i, dst_dd_i, ne00, row_diff, stream);
break; break;
case GGML_TYPE_IQ2_XS:
mul_mat_vec_iq2_xs_q8_1_cuda(src0_dd_i, src1_ddq_i, dst_dd_i, ne00, row_diff, stream);
break;
default: default:
GGML_ASSERT(false); GGML_ASSERT(false);
break; break;

42
ggml-metal.m
View File

@ -89,6 +89,7 @@ struct ggml_metal_context {
GGML_METAL_DECL_KERNEL(get_rows_q6_K); GGML_METAL_DECL_KERNEL(get_rows_q6_K);
GGML_METAL_DECL_KERNEL(get_rows_i32); GGML_METAL_DECL_KERNEL(get_rows_i32);
GGML_METAL_DECL_KERNEL(get_rows_iq2_xxs); GGML_METAL_DECL_KERNEL(get_rows_iq2_xxs);
GGML_METAL_DECL_KERNEL(get_rows_iq2_xs);
GGML_METAL_DECL_KERNEL(rms_norm); GGML_METAL_DECL_KERNEL(rms_norm);
GGML_METAL_DECL_KERNEL(group_norm); GGML_METAL_DECL_KERNEL(group_norm);
GGML_METAL_DECL_KERNEL(norm); GGML_METAL_DECL_KERNEL(norm);
@ -108,6 +109,7 @@ struct ggml_metal_context {
GGML_METAL_DECL_KERNEL(mul_mv_q5_K_f32); GGML_METAL_DECL_KERNEL(mul_mv_q5_K_f32);
GGML_METAL_DECL_KERNEL(mul_mv_q6_K_f32); GGML_METAL_DECL_KERNEL(mul_mv_q6_K_f32);
GGML_METAL_DECL_KERNEL(mul_mv_iq2_xxs_f32); GGML_METAL_DECL_KERNEL(mul_mv_iq2_xxs_f32);
GGML_METAL_DECL_KERNEL(mul_mv_iq2_xs_f32);
GGML_METAL_DECL_KERNEL(mul_mv_id_f32_f32); GGML_METAL_DECL_KERNEL(mul_mv_id_f32_f32);
//GGML_METAL_DECL_KERNEL(mul_mv_id_f16_f16); //GGML_METAL_DECL_KERNEL(mul_mv_id_f16_f16);
GGML_METAL_DECL_KERNEL(mul_mv_id_f16_f32); GGML_METAL_DECL_KERNEL(mul_mv_id_f16_f32);
@ -124,6 +126,7 @@ struct ggml_metal_context {
GGML_METAL_DECL_KERNEL(mul_mv_id_q5_K_f32); GGML_METAL_DECL_KERNEL(mul_mv_id_q5_K_f32);
GGML_METAL_DECL_KERNEL(mul_mv_id_q6_K_f32); GGML_METAL_DECL_KERNEL(mul_mv_id_q6_K_f32);
GGML_METAL_DECL_KERNEL(mul_mv_id_iq2_xxs_f32); GGML_METAL_DECL_KERNEL(mul_mv_id_iq2_xxs_f32);
GGML_METAL_DECL_KERNEL(mul_mv_id_iq2_xs_f32);
GGML_METAL_DECL_KERNEL(mul_mm_f32_f32); GGML_METAL_DECL_KERNEL(mul_mm_f32_f32);
GGML_METAL_DECL_KERNEL(mul_mm_f16_f32); GGML_METAL_DECL_KERNEL(mul_mm_f16_f32);
GGML_METAL_DECL_KERNEL(mul_mm_q4_0_f32); GGML_METAL_DECL_KERNEL(mul_mm_q4_0_f32);
@ -137,6 +140,7 @@ struct ggml_metal_context {
GGML_METAL_DECL_KERNEL(mul_mm_q5_K_f32); GGML_METAL_DECL_KERNEL(mul_mm_q5_K_f32);
GGML_METAL_DECL_KERNEL(mul_mm_q6_K_f32); GGML_METAL_DECL_KERNEL(mul_mm_q6_K_f32);
GGML_METAL_DECL_KERNEL(mul_mm_iq2_xxs_f32); GGML_METAL_DECL_KERNEL(mul_mm_iq2_xxs_f32);
GGML_METAL_DECL_KERNEL(mul_mm_iq2_xs_f32);
GGML_METAL_DECL_KERNEL(mul_mm_id_f32_f32); GGML_METAL_DECL_KERNEL(mul_mm_id_f32_f32);
GGML_METAL_DECL_KERNEL(mul_mm_id_f16_f32); GGML_METAL_DECL_KERNEL(mul_mm_id_f16_f32);
GGML_METAL_DECL_KERNEL(mul_mm_id_q4_0_f32); GGML_METAL_DECL_KERNEL(mul_mm_id_q4_0_f32);
@ -150,6 +154,7 @@ struct ggml_metal_context {
GGML_METAL_DECL_KERNEL(mul_mm_id_q5_K_f32); GGML_METAL_DECL_KERNEL(mul_mm_id_q5_K_f32);
GGML_METAL_DECL_KERNEL(mul_mm_id_q6_K_f32); GGML_METAL_DECL_KERNEL(mul_mm_id_q6_K_f32);
GGML_METAL_DECL_KERNEL(mul_mm_id_iq2_xxs_f32); GGML_METAL_DECL_KERNEL(mul_mm_id_iq2_xxs_f32);
GGML_METAL_DECL_KERNEL(mul_mm_id_iq2_xs_f32);
GGML_METAL_DECL_KERNEL(rope_f32); GGML_METAL_DECL_KERNEL(rope_f32);
GGML_METAL_DECL_KERNEL(rope_f16); GGML_METAL_DECL_KERNEL(rope_f16);
GGML_METAL_DECL_KERNEL(alibi_f32); GGML_METAL_DECL_KERNEL(alibi_f32);
@ -385,6 +390,7 @@ struct ggml_metal_context * ggml_metal_init(int n_cb) {
GGML_METAL_ADD_KERNEL(get_rows_q6_K); GGML_METAL_ADD_KERNEL(get_rows_q6_K);
GGML_METAL_ADD_KERNEL(get_rows_i32); GGML_METAL_ADD_KERNEL(get_rows_i32);
GGML_METAL_ADD_KERNEL(get_rows_iq2_xxs); GGML_METAL_ADD_KERNEL(get_rows_iq2_xxs);
GGML_METAL_ADD_KERNEL(get_rows_iq2_xs);
GGML_METAL_ADD_KERNEL(rms_norm); GGML_METAL_ADD_KERNEL(rms_norm);
GGML_METAL_ADD_KERNEL(group_norm); GGML_METAL_ADD_KERNEL(group_norm);
GGML_METAL_ADD_KERNEL(norm); GGML_METAL_ADD_KERNEL(norm);
@ -404,6 +410,7 @@ struct ggml_metal_context * ggml_metal_init(int n_cb) {
GGML_METAL_ADD_KERNEL(mul_mv_q5_K_f32); GGML_METAL_ADD_KERNEL(mul_mv_q5_K_f32);
GGML_METAL_ADD_KERNEL(mul_mv_q6_K_f32); GGML_METAL_ADD_KERNEL(mul_mv_q6_K_f32);
GGML_METAL_ADD_KERNEL(mul_mv_iq2_xxs_f32); GGML_METAL_ADD_KERNEL(mul_mv_iq2_xxs_f32);
GGML_METAL_ADD_KERNEL(mul_mv_iq2_xs_f32);
GGML_METAL_ADD_KERNEL(mul_mv_id_f32_f32); GGML_METAL_ADD_KERNEL(mul_mv_id_f32_f32);
//GGML_METAL_ADD_KERNEL(mul_mv_id_f16_f16); //GGML_METAL_ADD_KERNEL(mul_mv_id_f16_f16);
GGML_METAL_ADD_KERNEL(mul_mv_id_f16_f32); GGML_METAL_ADD_KERNEL(mul_mv_id_f16_f32);
@ -420,6 +427,7 @@ struct ggml_metal_context * ggml_metal_init(int n_cb) {
GGML_METAL_ADD_KERNEL(mul_mv_id_q5_K_f32); GGML_METAL_ADD_KERNEL(mul_mv_id_q5_K_f32);
GGML_METAL_ADD_KERNEL(mul_mv_id_q6_K_f32); GGML_METAL_ADD_KERNEL(mul_mv_id_q6_K_f32);
GGML_METAL_ADD_KERNEL(mul_mv_id_iq2_xxs_f32); GGML_METAL_ADD_KERNEL(mul_mv_id_iq2_xxs_f32);
GGML_METAL_ADD_KERNEL(mul_mv_id_iq2_xs_f32);
if ([ctx->device supportsFamily:MTLGPUFamilyApple7]) { if ([ctx->device supportsFamily:MTLGPUFamilyApple7]) {
GGML_METAL_ADD_KERNEL(mul_mm_f32_f32); GGML_METAL_ADD_KERNEL(mul_mm_f32_f32);
GGML_METAL_ADD_KERNEL(mul_mm_f16_f32); GGML_METAL_ADD_KERNEL(mul_mm_f16_f32);
@ -434,6 +442,7 @@ struct ggml_metal_context * ggml_metal_init(int n_cb) {
GGML_METAL_ADD_KERNEL(mul_mm_q5_K_f32); GGML_METAL_ADD_KERNEL(mul_mm_q5_K_f32);
GGML_METAL_ADD_KERNEL(mul_mm_q6_K_f32); GGML_METAL_ADD_KERNEL(mul_mm_q6_K_f32);
GGML_METAL_ADD_KERNEL(mul_mm_iq2_xxs_f32); GGML_METAL_ADD_KERNEL(mul_mm_iq2_xxs_f32);
GGML_METAL_ADD_KERNEL(mul_mm_iq2_xs_f32);
GGML_METAL_ADD_KERNEL(mul_mm_id_f32_f32); GGML_METAL_ADD_KERNEL(mul_mm_id_f32_f32);
GGML_METAL_ADD_KERNEL(mul_mm_id_f16_f32); GGML_METAL_ADD_KERNEL(mul_mm_id_f16_f32);
GGML_METAL_ADD_KERNEL(mul_mm_id_q4_0_f32); GGML_METAL_ADD_KERNEL(mul_mm_id_q4_0_f32);
@ -447,6 +456,7 @@ struct ggml_metal_context * ggml_metal_init(int n_cb) {
GGML_METAL_ADD_KERNEL(mul_mm_id_q5_K_f32); GGML_METAL_ADD_KERNEL(mul_mm_id_q5_K_f32);
GGML_METAL_ADD_KERNEL(mul_mm_id_q6_K_f32); GGML_METAL_ADD_KERNEL(mul_mm_id_q6_K_f32);
GGML_METAL_ADD_KERNEL(mul_mm_id_iq2_xxs_f32); GGML_METAL_ADD_KERNEL(mul_mm_id_iq2_xxs_f32);
GGML_METAL_ADD_KERNEL(mul_mm_id_iq2_xs_f32);
} }
GGML_METAL_ADD_KERNEL(rope_f32); GGML_METAL_ADD_KERNEL(rope_f32);
GGML_METAL_ADD_KERNEL(rope_f16); GGML_METAL_ADD_KERNEL(rope_f16);
@ -513,6 +523,7 @@ void ggml_metal_free(struct ggml_metal_context * ctx) {
GGML_METAL_DEL_KERNEL(get_rows_q6_K); GGML_METAL_DEL_KERNEL(get_rows_q6_K);
GGML_METAL_DEL_KERNEL(get_rows_i32); GGML_METAL_DEL_KERNEL(get_rows_i32);
GGML_METAL_DEL_KERNEL(get_rows_iq2_xxs); GGML_METAL_DEL_KERNEL(get_rows_iq2_xxs);
GGML_METAL_DEL_KERNEL(get_rows_iq2_xs);
GGML_METAL_DEL_KERNEL(rms_norm); GGML_METAL_DEL_KERNEL(rms_norm);
GGML_METAL_DEL_KERNEL(group_norm); GGML_METAL_DEL_KERNEL(group_norm);
GGML_METAL_DEL_KERNEL(norm); GGML_METAL_DEL_KERNEL(norm);
@ -532,6 +543,7 @@ void ggml_metal_free(struct ggml_metal_context * ctx) {
GGML_METAL_DEL_KERNEL(mul_mv_q5_K_f32); GGML_METAL_DEL_KERNEL(mul_mv_q5_K_f32);
GGML_METAL_DEL_KERNEL(mul_mv_q6_K_f32); GGML_METAL_DEL_KERNEL(mul_mv_q6_K_f32);
GGML_METAL_DEL_KERNEL(mul_mv_iq2_xxs_f32); GGML_METAL_DEL_KERNEL(mul_mv_iq2_xxs_f32);
GGML_METAL_DEL_KERNEL(mul_mv_iq2_xs_f32);
GGML_METAL_DEL_KERNEL(mul_mv_id_f32_f32); GGML_METAL_DEL_KERNEL(mul_mv_id_f32_f32);
//GGML_METAL_DEL_KERNEL(mul_mv_id_f16_f16); //GGML_METAL_DEL_KERNEL(mul_mv_id_f16_f16);
GGML_METAL_DEL_KERNEL(mul_mv_id_f16_f32); GGML_METAL_DEL_KERNEL(mul_mv_id_f16_f32);
@ -548,6 +560,7 @@ void ggml_metal_free(struct ggml_metal_context * ctx) {
GGML_METAL_DEL_KERNEL(mul_mv_id_q5_K_f32); GGML_METAL_DEL_KERNEL(mul_mv_id_q5_K_f32);
GGML_METAL_DEL_KERNEL(mul_mv_id_q6_K_f32); GGML_METAL_DEL_KERNEL(mul_mv_id_q6_K_f32);
GGML_METAL_DEL_KERNEL(mul_mv_id_iq2_xxs_f32); GGML_METAL_DEL_KERNEL(mul_mv_id_iq2_xxs_f32);
GGML_METAL_DEL_KERNEL(mul_mv_id_iq2_xs_f32);
if ([ctx->device supportsFamily:MTLGPUFamilyApple7]) { if ([ctx->device supportsFamily:MTLGPUFamilyApple7]) {
GGML_METAL_DEL_KERNEL(mul_mm_f32_f32); GGML_METAL_DEL_KERNEL(mul_mm_f32_f32);
GGML_METAL_DEL_KERNEL(mul_mm_f16_f32); GGML_METAL_DEL_KERNEL(mul_mm_f16_f32);
@ -562,6 +575,7 @@ void ggml_metal_free(struct ggml_metal_context * ctx) {
GGML_METAL_DEL_KERNEL(mul_mm_q5_K_f32); GGML_METAL_DEL_KERNEL(mul_mm_q5_K_f32);
GGML_METAL_DEL_KERNEL(mul_mm_q6_K_f32); GGML_METAL_DEL_KERNEL(mul_mm_q6_K_f32);
GGML_METAL_DEL_KERNEL(mul_mm_iq2_xxs_f32); GGML_METAL_DEL_KERNEL(mul_mm_iq2_xxs_f32);
GGML_METAL_DEL_KERNEL(mul_mm_iq2_xs_f32);
GGML_METAL_DEL_KERNEL(mul_mm_id_f32_f32); GGML_METAL_DEL_KERNEL(mul_mm_id_f32_f32);
GGML_METAL_DEL_KERNEL(mul_mm_id_f16_f32); GGML_METAL_DEL_KERNEL(mul_mm_id_f16_f32);
GGML_METAL_DEL_KERNEL(mul_mm_id_q4_0_f32); GGML_METAL_DEL_KERNEL(mul_mm_id_q4_0_f32);
@ -575,6 +589,7 @@ void ggml_metal_free(struct ggml_metal_context * ctx) {
GGML_METAL_DEL_KERNEL(mul_mm_id_q5_K_f32); GGML_METAL_DEL_KERNEL(mul_mm_id_q5_K_f32);
GGML_METAL_DEL_KERNEL(mul_mm_id_q6_K_f32); GGML_METAL_DEL_KERNEL(mul_mm_id_q6_K_f32);
GGML_METAL_DEL_KERNEL(mul_mm_id_iq2_xxs_f32); GGML_METAL_DEL_KERNEL(mul_mm_id_iq2_xxs_f32);
GGML_METAL_DEL_KERNEL(mul_mm_id_iq2_xs_f32);
} }
GGML_METAL_DEL_KERNEL(rope_f32); GGML_METAL_DEL_KERNEL(rope_f32);
GGML_METAL_DEL_KERNEL(rope_f16); GGML_METAL_DEL_KERNEL(rope_f16);
@ -1561,6 +1576,7 @@ bool ggml_metal_graph_compute(
case GGML_TYPE_Q5_K: [encoder setComputePipelineState:ctx->pipeline_mul_mm_q5_K_f32]; break; case GGML_TYPE_Q5_K: [encoder setComputePipelineState:ctx->pipeline_mul_mm_q5_K_f32]; break;
case GGML_TYPE_Q6_K: [encoder setComputePipelineState:ctx->pipeline_mul_mm_q6_K_f32]; break; case GGML_TYPE_Q6_K: [encoder setComputePipelineState:ctx->pipeline_mul_mm_q6_K_f32]; break;
case GGML_TYPE_IQ2_XXS: [encoder setComputePipelineState:ctx->pipeline_mul_mm_iq2_xxs_f32]; break; case GGML_TYPE_IQ2_XXS: [encoder setComputePipelineState:ctx->pipeline_mul_mm_iq2_xxs_f32]; break;
case GGML_TYPE_IQ2_XS : [encoder setComputePipelineState:ctx->pipeline_mul_mm_iq2_xs_f32]; break;
default: GGML_ASSERT(false && "MUL MAT-MAT not implemented"); default: GGML_ASSERT(false && "MUL MAT-MAT not implemented");
} }
[encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0];
@ -1679,6 +1695,12 @@ bool ggml_metal_graph_compute(
nth1 = 16; nth1 = 16;
[encoder setComputePipelineState:ctx->pipeline_mul_mv_iq2_xxs_f32]; [encoder setComputePipelineState:ctx->pipeline_mul_mv_iq2_xxs_f32];
} break; } break;
case GGML_TYPE_IQ2_XS:
{
nth0 = 4;
nth1 = 16;
[encoder setComputePipelineState:ctx->pipeline_mul_mv_iq2_xs_f32];
} break;
default: default:
{ {
GGML_METAL_LOG_ERROR("Asserting on type %d\n", (int)src0t); GGML_METAL_LOG_ERROR("Asserting on type %d\n", (int)src0t);
@ -1712,12 +1734,12 @@ bool ggml_metal_graph_compute(
if (src0t == GGML_TYPE_Q4_0 || src0t == GGML_TYPE_Q4_1 || if (src0t == GGML_TYPE_Q4_0 || src0t == GGML_TYPE_Q4_1 ||
src0t == GGML_TYPE_Q5_0 || src0t == GGML_TYPE_Q5_1 || src0t == GGML_TYPE_Q8_0 || src0t == GGML_TYPE_Q5_0 || src0t == GGML_TYPE_Q5_1 || src0t == GGML_TYPE_Q8_0 ||
//src0t == GGML_TYPE_IQ2_XXS ||
src0t == GGML_TYPE_Q2_K) { // || src0t == GGML_TYPE_Q4_K) { src0t == GGML_TYPE_Q2_K) { // || src0t == GGML_TYPE_Q4_K) {
[encoder dispatchThreadgroups:MTLSizeMake((ne01 + 7)/8, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 7)/8, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)];
} }
else if (src0t == GGML_TYPE_IQ2_XXS) { else if (src0t == GGML_TYPE_IQ2_XXS || src0t == GGML_TYPE_IQ2_XS) {
[encoder setThreadgroupMemoryLength:(256*8+128) atIndex:0]; const int mem_size = src0t == GGML_TYPE_IQ2_XXS ? 256*8+128 : 512*8+128;
[encoder setThreadgroupMemoryLength:mem_size atIndex:0];
[encoder dispatchThreadgroups:MTLSizeMake((ne01 + 7)/8, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 7)/8, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)];
} }
else if (src0t == GGML_TYPE_Q4_K) { else if (src0t == GGML_TYPE_Q4_K) {
@ -1810,6 +1832,7 @@ bool ggml_metal_graph_compute(
case GGML_TYPE_Q5_K: [encoder setComputePipelineState:ctx->pipeline_mul_mm_id_q5_K_f32]; break; case GGML_TYPE_Q5_K: [encoder setComputePipelineState:ctx->pipeline_mul_mm_id_q5_K_f32]; break;
case GGML_TYPE_Q6_K: [encoder setComputePipelineState:ctx->pipeline_mul_mm_id_q6_K_f32]; break; case GGML_TYPE_Q6_K: [encoder setComputePipelineState:ctx->pipeline_mul_mm_id_q6_K_f32]; break;
case GGML_TYPE_IQ2_XXS: [encoder setComputePipelineState:ctx->pipeline_mul_mm_id_iq2_xxs_f32]; break; case GGML_TYPE_IQ2_XXS: [encoder setComputePipelineState:ctx->pipeline_mul_mm_id_iq2_xxs_f32]; break;
case GGML_TYPE_IQ2_XS : [encoder setComputePipelineState:ctx->pipeline_mul_mm_id_iq2_xs_f32]; break;
default: GGML_ASSERT(false && "MUL_MAT_ID not implemented"); default: GGML_ASSERT(false && "MUL_MAT_ID not implemented");
} }
[encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0];
@ -1931,6 +1954,12 @@ bool ggml_metal_graph_compute(
nth1 = 16; nth1 = 16;
[encoder setComputePipelineState:ctx->pipeline_mul_mv_id_iq2_xxs_f32]; [encoder setComputePipelineState:ctx->pipeline_mul_mv_id_iq2_xxs_f32];
} break; } break;
case GGML_TYPE_IQ2_XS:
{
nth0 = 4;
nth1 = 16;
[encoder setComputePipelineState:ctx->pipeline_mul_mv_id_iq2_xs_f32];
} break;
default: default:
{ {
GGML_METAL_LOG_ERROR("Asserting on type %d\n", (int)src2t); GGML_METAL_LOG_ERROR("Asserting on type %d\n", (int)src2t);
@ -1980,12 +2009,12 @@ bool ggml_metal_graph_compute(
if (src2t == GGML_TYPE_Q4_0 || src2t == GGML_TYPE_Q4_1 || if (src2t == GGML_TYPE_Q4_0 || src2t == GGML_TYPE_Q4_1 ||
src2t == GGML_TYPE_Q5_0 || src2t == GGML_TYPE_Q5_1 || src2t == GGML_TYPE_Q8_0 || src2t == GGML_TYPE_Q5_0 || src2t == GGML_TYPE_Q5_1 || src2t == GGML_TYPE_Q8_0 ||
//src2t == GGML_TYPE_IQ2_XXS ||
src2t == GGML_TYPE_Q2_K) { // || src2t == GGML_TYPE_Q4_K) { src2t == GGML_TYPE_Q2_K) { // || src2t == GGML_TYPE_Q4_K) {
[encoder dispatchThreadgroups:MTLSizeMake((ne21 + 7)/8, _ne1, ne01*ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; [encoder dispatchThreadgroups:MTLSizeMake((ne21 + 7)/8, _ne1, ne01*ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)];
} }
else if (src2t == GGML_TYPE_IQ2_XXS) { else if (src2t == GGML_TYPE_IQ2_XXS || src2t == GGML_TYPE_IQ2_XS) {
[encoder setThreadgroupMemoryLength:(256*8+128) atIndex:0]; const int mem_size = src2t == GGML_TYPE_IQ2_XXS ? 256*8+128 : 512*8+128;
[encoder setThreadgroupMemoryLength:mem_size atIndex:0];
[encoder dispatchThreadgroups:MTLSizeMake((ne21 + 7)/8, _ne1, ne01*ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; [encoder dispatchThreadgroups:MTLSizeMake((ne21 + 7)/8, _ne1, ne01*ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)];
} }
else if (src2t == GGML_TYPE_Q4_K) { else if (src2t == GGML_TYPE_Q4_K) {
@ -2026,6 +2055,7 @@ bool ggml_metal_graph_compute(
case GGML_TYPE_Q6_K: [encoder setComputePipelineState:ctx->pipeline_get_rows_q6_K]; break; case GGML_TYPE_Q6_K: [encoder setComputePipelineState:ctx->pipeline_get_rows_q6_K]; break;
case GGML_TYPE_I32: [encoder setComputePipelineState:ctx->pipeline_get_rows_i32]; break; case GGML_TYPE_I32: [encoder setComputePipelineState:ctx->pipeline_get_rows_i32]; break;
case GGML_TYPE_IQ2_XXS: [encoder setComputePipelineState:ctx->pipeline_get_rows_iq2_xxs]; break; case GGML_TYPE_IQ2_XXS: [encoder setComputePipelineState:ctx->pipeline_get_rows_iq2_xxs]; break;
case GGML_TYPE_IQ2_XS : [encoder setComputePipelineState:ctx->pipeline_get_rows_iq2_xs]; break;
default: GGML_ASSERT(false && "not implemented"); default: GGML_ASSERT(false && "not implemented");
} }

378
ggml-metal.metal
View File

@ -2452,6 +2452,13 @@ typedef struct {
} block_iq2_xxs; } block_iq2_xxs;
// 66 bytes / block for QK_K = 256, so 2.0625 bpw // 66 bytes / block for QK_K = 256, so 2.0625 bpw
typedef struct {
half d;
uint16_t qs[QK_K/8];
uint8_t scales[QK_K/32];
} block_iq2_xs;
// 74 bytes / block for QK_K = 256, so 2.3125 bpw
//====================================== dot products ========================= //====================================== dot products =========================
void kernel_mul_mv_q2_K_f32_impl( void kernel_mul_mv_q2_K_f32_impl(
@ -3476,7 +3483,7 @@ kernel void kernel_mul_mv_q6_K_f32(
// ======================= "True" 2-bit // ======================= "True" 2-bit
constexpr constant static uint64_t kgrid_iq2xxs[256] = { constexpr constant static uint64_t iq2xxs_grid[256] = {
0x0808080808080808, 0x080808080808082b, 0x0808080808081919, 0x0808080808082b08, 0x0808080808080808, 0x080808080808082b, 0x0808080808081919, 0x0808080808082b08,
0x0808080808082b2b, 0x0808080808190819, 0x0808080808191908, 0x08080808082b0808, 0x0808080808082b2b, 0x0808080808190819, 0x0808080808191908, 0x08080808082b0808,
0x08080808082b082b, 0x08080808082b2b08, 0x08080808082b2b2b, 0x0808080819080819, 0x08080808082b082b, 0x08080808082b2b08, 0x08080808082b2b2b, 0x0808080819080819,
@ -3543,6 +3550,137 @@ constexpr constant static uint64_t kgrid_iq2xxs[256] = {
0x2b2b082b08080808, 0x2b2b190808192b08, 0x2b2b2b0819190808, 0x2b2b2b1908081908, 0x2b2b082b08080808, 0x2b2b190808192b08, 0x2b2b2b0819190808, 0x2b2b2b1908081908,
}; };
constexpr constant static uint64_t iq2xs_grid[512] = {
0x0808080808080808, 0x080808080808082b, 0x0808080808081919, 0x0808080808082b08,
0x0808080808082b2b, 0x0808080808190819, 0x0808080808191908, 0x080808080819192b,
0x0808080808192b19, 0x08080808082b0808, 0x08080808082b082b, 0x08080808082b1919,
0x08080808082b2b08, 0x0808080819080819, 0x0808080819081908, 0x080808081908192b,
0x0808080819082b19, 0x0808080819190808, 0x080808081919082b, 0x0808080819191919,
0x0808080819192b08, 0x08080808192b0819, 0x08080808192b1908, 0x080808082b080808,
0x080808082b08082b, 0x080808082b081919, 0x080808082b082b08, 0x080808082b190819,
0x080808082b191908, 0x080808082b192b19, 0x080808082b2b0808, 0x0808081908080819,
0x0808081908081908, 0x080808190808192b, 0x0808081908082b19, 0x0808081908190808,
0x080808190819082b, 0x0808081908191919, 0x0808081908192b08, 0x0808081908192b2b,
0x08080819082b0819, 0x08080819082b1908, 0x0808081919080808, 0x080808191908082b,
0x0808081919081919, 0x0808081919082b08, 0x0808081919190819, 0x0808081919191908,
0x08080819192b0808, 0x08080819192b2b08, 0x080808192b080819, 0x080808192b081908,
0x080808192b190808, 0x0808082b08080808, 0x0808082b0808082b, 0x0808082b08081919,
0x0808082b08082b08, 0x0808082b08190819, 0x0808082b08191908, 0x0808082b082b0808,
0x0808082b19080819, 0x0808082b19081908, 0x0808082b19190808, 0x0808082b19191919,
0x0808082b2b080808, 0x0808082b2b082b2b, 0x0808190808080819, 0x0808190808081908,
0x080819080808192b, 0x0808190808082b19, 0x0808190808190808, 0x080819080819082b,
0x0808190808191919, 0x0808190808192b08, 0x08081908082b0819, 0x08081908082b1908,
0x0808190819080808, 0x080819081908082b, 0x0808190819081919, 0x0808190819082b08,
0x0808190819190819, 0x0808190819191908, 0x080819081919192b, 0x08081908192b0808,
0x080819082b080819, 0x080819082b081908, 0x080819082b190808, 0x0808191908080808,
0x080819190808082b, 0x0808191908081919, 0x0808191908082b08, 0x0808191908190819,
0x0808191908191908, 0x08081919082b0808, 0x0808191919080819, 0x0808191919081908,
0x0808191919190808, 0x08081919192b0819, 0x080819192b080808, 0x0808192b08080819,
0x0808192b08081908, 0x0808192b08190808, 0x0808192b082b192b, 0x0808192b19080808,
0x0808192b1908082b, 0x0808192b2b081908, 0x08082b0808080808, 0x08082b080808082b,
0x08082b0808081919, 0x08082b0808082b08, 0x08082b0808082b2b, 0x08082b0808190819,
0x08082b0808191908, 0x08082b08082b0808, 0x08082b08082b1919, 0x08082b0819080819,
0x08082b0819081908, 0x08082b0819190808, 0x08082b0819192b08, 0x08082b082b080808,
0x08082b082b2b0808, 0x08082b082b2b2b2b, 0x08082b1908080819, 0x08082b1908081908,
0x08082b1908190808, 0x08082b1919080808, 0x08082b192b080819, 0x08082b192b082b19,
0x08082b2b08080808, 0x08082b2b082b0808, 0x08082b2b082b2b08, 0x08082b2b2b19192b,
0x08082b2b2b2b0808, 0x0819080808080819, 0x0819080808081908, 0x081908080808192b,
0x0819080808082b19, 0x0819080808190808, 0x081908080819082b, 0x0819080808191919,
0x0819080808192b08, 0x08190808082b0819, 0x08190808082b1908, 0x0819080819080808,
0x081908081908082b, 0x0819080819081919, 0x0819080819082b08, 0x0819080819190819,
0x0819080819191908, 0x08190808192b0808, 0x08190808192b2b2b, 0x081908082b080819,
0x081908082b081908, 0x081908082b190808, 0x0819081908080808, 0x081908190808082b,
0x0819081908081919, 0x0819081908082b08, 0x0819081908190819, 0x0819081908191908,
0x08190819082b0808, 0x0819081919080819, 0x0819081919081908, 0x0819081919190808,
0x081908192b080808, 0x081908192b191908, 0x081908192b19192b, 0x0819082b08080819,
0x0819082b08081908, 0x0819082b0808192b, 0x0819082b08190808, 0x0819082b19080808,
0x0819082b192b0808, 0x0819190808080808, 0x081919080808082b, 0x0819190808081919,
0x0819190808082b08, 0x0819190808190819, 0x0819190808191908, 0x08191908082b0808,
0x0819190819080819, 0x0819190819081908, 0x0819190819082b19, 0x0819190819190808,
0x08191908192b1908, 0x081919082b080808, 0x0819191908080819, 0x0819191908081908,
0x0819191908190808, 0x0819191919080808, 0x0819192b08080808, 0x0819192b08191908,
0x0819192b19082b19, 0x08192b0808080819, 0x08192b0808081908, 0x08192b0808190808,
0x08192b080819082b, 0x08192b0819080808, 0x08192b0819191908, 0x08192b082b08192b,
0x08192b1908080808, 0x08192b1908081919, 0x08192b19192b192b, 0x08192b2b19190819,
0x08192b2b2b2b2b19, 0x082b080808080808, 0x082b08080808082b, 0x082b080808081919,
0x082b080808082b08, 0x082b080808082b2b, 0x082b080808190819, 0x082b080808191908,
0x082b0808082b0808, 0x082b080819080819, 0x082b080819081908, 0x082b080819190808,
0x082b08082b080808, 0x082b08082b2b0808, 0x082b081908080819, 0x082b081908081908,
0x082b081908190808, 0x082b081919080808, 0x082b081919082b08, 0x082b0819192b1919,
0x082b082b08080808, 0x082b082b082b082b, 0x082b082b2b080808, 0x082b082b2b2b2b08,
0x082b190808080819, 0x082b190808081908, 0x082b190808190808, 0x082b1908082b2b19,
0x082b190819080808, 0x082b191908080808, 0x082b191919080819, 0x082b19191919082b,
0x082b19192b192b19, 0x082b192b08080819, 0x082b192b08192b2b, 0x082b192b2b2b192b,
0x082b2b0808080808, 0x082b2b0808082b08, 0x082b2b0808082b2b, 0x082b2b08082b0808,
0x082b2b0819191919, 0x082b2b082b082b08, 0x082b2b082b2b082b, 0x082b2b19192b2b08,
0x082b2b192b190808, 0x082b2b2b08082b08, 0x082b2b2b082b0808, 0x082b2b2b2b08082b,
0x082b2b2b2b082b08, 0x082b2b2b2b082b2b, 0x1908080808080819, 0x1908080808081908,
0x190808080808192b, 0x1908080808082b19, 0x1908080808190808, 0x190808080819082b,
0x1908080808191919, 0x1908080808192b08, 0x19080808082b0819, 0x19080808082b1908,
0x1908080819080808, 0x190808081908082b, 0x1908080819081919, 0x1908080819082b08,
0x1908080819082b2b, 0x1908080819190819, 0x1908080819191908, 0x19080808192b0808,
0x19080808192b1919, 0x190808082b080819, 0x190808082b081908, 0x190808082b190808,
0x1908081908080808, 0x190808190808082b, 0x1908081908081919, 0x1908081908082b08,
0x1908081908190819, 0x1908081908191908, 0x19080819082b0808, 0x1908081919080819,
0x1908081919081908, 0x1908081919190808, 0x190808192b080808, 0x190808192b081919,
0x190808192b2b082b, 0x1908082b08080819, 0x1908082b08081908, 0x1908082b08190808,
0x1908082b0819082b, 0x1908082b082b2b19, 0x1908082b19080808, 0x1908190808080808,
0x190819080808082b, 0x1908190808081919, 0x1908190808082b08, 0x1908190808190819,
0x1908190808191908, 0x1908190808192b19, 0x19081908082b0808, 0x1908190819080819,
0x1908190819081908, 0x1908190819190808, 0x190819082b080808, 0x190819082b191908,
0x1908191908080819, 0x1908191908081908, 0x1908191908190808, 0x19081919082b1908,
0x1908191919080808, 0x190819192b192b2b, 0x1908192b08080808, 0x1908192b08082b2b,
0x1908192b19081908, 0x1908192b19190808, 0x19082b0808080819, 0x19082b0808081908,
0x19082b0808190808, 0x19082b0819080808, 0x19082b0819081919, 0x19082b0819191908,
0x19082b08192b082b, 0x19082b1908080808, 0x19082b1908190819, 0x19082b1919081908,
0x19082b1919190808, 0x19082b19192b2b19, 0x19082b2b08081908, 0x1919080808080808,
0x191908080808082b, 0x1919080808081919, 0x1919080808082b08, 0x1919080808190819,
0x1919080808191908, 0x19190808082b0808, 0x19190808082b2b08, 0x1919080819080819,
0x1919080819081908, 0x1919080819190808, 0x191908082b080808, 0x1919081908080819,
0x1919081908081908, 0x1919081908190808, 0x1919081908191919, 0x1919081919080808,
0x191908191908082b, 0x1919082b08080808, 0x1919082b19081908, 0x1919082b2b2b2b2b,
0x1919190808080819, 0x1919190808081908, 0x1919190808190808, 0x19191908082b0819,
0x1919190819080808, 0x19191908192b0808, 0x191919082b080819, 0x191919082b2b0819,
0x1919191908080808, 0x1919191908082b08, 0x191919192b080808, 0x191919192b082b08,
0x1919192b082b0819, 0x1919192b192b2b08, 0x1919192b2b2b0819, 0x19192b0808080808,
0x19192b0808191908, 0x19192b0819080819, 0x19192b0819190808, 0x19192b082b192b19,
0x19192b1908192b2b, 0x19192b1919080808, 0x19192b191908082b, 0x19192b2b2b081919,
0x192b080808080819, 0x192b080808081908, 0x192b080808190808, 0x192b080819080808,
0x192b080819191908, 0x192b0808192b082b, 0x192b08082b08192b, 0x192b08082b2b2b19,
0x192b081908080808, 0x192b082b082b1908, 0x192b082b19082b2b, 0x192b082b2b19082b,
0x192b190808080808, 0x192b19080819192b, 0x192b191908190808, 0x192b191919080808,
0x192b191919081919, 0x192b19192b2b1908, 0x192b2b0808080819, 0x192b2b08192b2b2b,
0x192b2b19082b1919, 0x192b2b2b0808192b, 0x192b2b2b19191908, 0x192b2b2b192b082b,
0x2b08080808080808, 0x2b0808080808082b, 0x2b08080808081919, 0x2b08080808082b08,
0x2b08080808190819, 0x2b08080808191908, 0x2b080808082b0808, 0x2b080808082b2b2b,
0x2b08080819080819, 0x2b08080819081908, 0x2b08080819190808, 0x2b0808082b080808,
0x2b0808082b08082b, 0x2b0808082b2b2b08, 0x2b0808082b2b2b2b, 0x2b08081908080819,
0x2b08081908081908, 0x2b0808190808192b, 0x2b08081908190808, 0x2b08081919080808,
0x2b08081919190819, 0x2b08081919192b19, 0x2b08082b08080808, 0x2b08082b082b0808,
0x2b08082b2b080808, 0x2b08082b2b08082b, 0x2b08082b2b2b0808, 0x2b08082b2b2b2b08,
0x2b08190808080819, 0x2b08190808081908, 0x2b08190808190808, 0x2b0819080819082b,
0x2b08190808191919, 0x2b08190819080808, 0x2b081908192b0808, 0x2b0819082b082b19,
0x2b08191908080808, 0x2b08191919081908, 0x2b0819192b2b1919, 0x2b08192b08192b08,
0x2b08192b192b2b2b, 0x2b082b0808080808, 0x2b082b0808082b08, 0x2b082b08082b1919,
0x2b082b0819192b2b, 0x2b082b082b080808, 0x2b082b082b08082b, 0x2b082b082b2b2b08,
0x2b082b190808192b, 0x2b082b2b082b082b, 0x2b082b2b2b080808, 0x2b082b2b2b082b08,
0x2b082b2b2b19192b, 0x2b082b2b2b2b2b08, 0x2b19080808080819, 0x2b19080808081908,
0x2b19080808190808, 0x2b19080819080808, 0x2b1908081919192b, 0x2b1908082b081908,
0x2b19081908080808, 0x2b190819082b082b, 0x2b190819192b1908, 0x2b19082b1919192b,
0x2b19082b2b082b19, 0x2b19190808080808, 0x2b19190808081919, 0x2b19190819081908,
0x2b19190819190808, 0x2b19190819192b08, 0x2b191919082b2b19, 0x2b1919192b190808,
0x2b1919192b19082b, 0x2b19192b19080819, 0x2b192b0819190819, 0x2b192b082b2b192b,
0x2b192b1919082b19, 0x2b192b2b08191919, 0x2b192b2b192b0808, 0x2b2b080808080808,
0x2b2b08080808082b, 0x2b2b080808082b08, 0x2b2b080808082b2b, 0x2b2b0808082b0808,
0x2b2b0808082b2b2b, 0x2b2b08082b2b0808, 0x2b2b081919190819, 0x2b2b081919192b19,
0x2b2b08192b2b192b, 0x2b2b082b08080808, 0x2b2b082b0808082b, 0x2b2b082b08082b08,
0x2b2b082b082b2b2b, 0x2b2b082b2b080808, 0x2b2b082b2b2b0808, 0x2b2b190819080808,
0x2b2b19082b191919, 0x2b2b192b192b1919, 0x2b2b192b2b192b08, 0x2b2b2b0808082b2b,
0x2b2b2b08082b0808, 0x2b2b2b08082b082b, 0x2b2b2b08082b2b08, 0x2b2b2b082b2b0808,
0x2b2b2b082b2b2b08, 0x2b2b2b1908081908, 0x2b2b2b192b081908, 0x2b2b2b192b08192b,
0x2b2b2b2b082b2b08, 0x2b2b2b2b082b2b2b, 0x2b2b2b2b2b190819, 0x2b2b2b2b2b2b2b2b,
};
constexpr constant static uint8_t ksigns_iq2xs[128] = { constexpr constant static uint8_t ksigns_iq2xs[128] = {
0, 129, 130, 3, 132, 5, 6, 135, 136, 9, 10, 139, 12, 141, 142, 15, 0, 129, 130, 3, 132, 5, 6, 135, 136, 9, 10, 139, 12, 141, 142, 15,
144, 17, 18, 147, 20, 149, 150, 23, 24, 153, 154, 27, 156, 29, 30, 159, 144, 17, 18, 147, 20, 149, 150, 23, 24, 153, 154, 27, 156, 29, 30, 159,
@ -3600,7 +3738,7 @@ void kernel_mul_mv_iq2_xxs_f32_impl(
{ {
int nval = 4; int nval = 4;
int pos = (32*sgitg + tiisg)*nval; int pos = (32*sgitg + tiisg)*nval;
for (int i = 0; i < nval; ++i) values[pos + i] = kgrid_iq2xxs[pos + i]; for (int i = 0; i < nval; ++i) values[pos + i] = iq2xxs_grid[pos + i];
nval = 2; nval = 2;
pos = (32*sgitg + tiisg)*nval; pos = (32*sgitg + tiisg)*nval;
for (int i = 0; i < nval; ++i) shared_signs[pos+i] = ksigns_iq2xs[pos+i]; for (int i = 0; i < nval; ++i) shared_signs[pos+i] = ksigns_iq2xs[pos+i];
@ -3689,6 +3827,149 @@ kernel void kernel_mul_mv_iq2_xxs_f32(
kernel_mul_mv_iq2_xxs_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg); kernel_mul_mv_iq2_xxs_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg);
} }
void kernel_mul_mv_iq2_xs_f32_impl(
device const void * src0,
device const float * src1,
device float * dst,
constant int64_t & ne00,
constant int64_t & ne01,
constant int64_t & ne02,
constant int64_t & ne10,
constant int64_t & ne12,
constant int64_t & ne0,
constant int64_t & ne1,
constant uint & r2,
constant uint & r3,
threadgroup int8_t * shared_values [[threadgroup(0)]],
uint3 tgpig[[threadgroup_position_in_grid]],
uint tiisg[[thread_index_in_simdgroup]],
uint sgitg[[simdgroup_index_in_threadgroup]]) {
const int nb = ne00/QK_K;
const int r0 = tgpig.x;
const int r1 = tgpig.y;
const int im = tgpig.z;
const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST;
const int ib_row = first_row * nb;
const uint i12 = im%ne12;
const uint i13 = im/ne12;
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
device const block_iq2_xs * x = (device const block_iq2_xs *) src0 + ib_row + offset0;
device const float * y = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
float yl[32];
float sumf[N_DST]={0.f}, all_sum;
const int nb32 = nb * (QK_K / 32);
threadgroup uint64_t * values = (threadgroup uint64_t *)shared_values;
threadgroup uint8_t * shared_signs = (threadgroup uint8_t *)(values + 512);
{
int nval = 8;
int pos = (32*sgitg + tiisg)*nval;
for (int i = 0; i < nval; ++i) values[pos + i] = iq2xs_grid[pos + i];
nval = 2;
pos = (32*sgitg + tiisg)*nval;
for (int i = 0; i < nval; ++i) shared_signs[pos+i] = ksigns_iq2xs[pos+i];
threadgroup_barrier(mem_flags::mem_threadgroup);
}
#if QK_K == 256
const int ix = tiisg;
device const float * y4 = y + 32 * ix;
for (int ib32 = ix; ib32 < nb32; ib32 += 32) {
for (int i = 0; i < 32; ++i) {
yl[i] = y4[i];
}
const int ibl = ib32 / (QK_K / 32);
const int ib = ib32 % (QK_K / 32);
device const block_iq2_xs * xr = x + ibl;
device const uint16_t * q2 = xr->qs + 4 * ib;
device const uint8_t * sc = xr->scales + ib;
device const half * dh = &xr->d;
for (int row = 0; row < N_DST; row++) {
const float db = dh[0];
const uint8_t ls1 = sc[0] & 0xf;
const uint8_t ls2 = sc[0] >> 4;
const float d1 = db * (0.5f + ls1);
const float d2 = db * (0.5f + ls2);
float sum1 = 0, sum2 = 0;
for (int l = 0; l < 2; ++l) {
const threadgroup uint8_t * grid = (const threadgroup uint8_t *)(values + (q2[l] & 511));
const uint8_t signs = shared_signs[(q2[l] >> 9)];
for (int j = 0; j < 8; ++j) {
sum1 += yl[8*l + j] * grid[j] * (signs & kmask_iq2xs[j] ? -1.f : 1.f);
}
}
for (int l = 2; l < 4; ++l) {
const threadgroup uint8_t * grid = (const threadgroup uint8_t *)(values + (q2[l] & 511));
const uint8_t signs = shared_signs[(q2[l] >> 9)];
for (int j = 0; j < 8; ++j) {
sum2 += yl[8*l + j] * grid[j] * (signs & kmask_iq2xs[j] ? -1.f : 1.f);
}
}
sumf[row] += d1 * sum1 + d2 * sum2;
dh += nb*sizeof(block_iq2_xs)/2;
q2 += nb*sizeof(block_iq2_xs)/2;
sc += nb*sizeof(block_iq2_xs);
}
y4 += 32 * 32;
}
#else
// TODO
#endif
for (int row = 0; row < N_DST; ++row) {
all_sum = simd_sum(sumf[row]);
if (tiisg == 0) {
dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum * 0.25f;
}
}
}
[[host_name("kernel_mul_mv_iq2_xs_f32")]]
kernel void kernel_mul_mv_iq2_xs_f32(
device const void * src0,
device const float * src1,
device float * dst,
constant int64_t & ne00,
constant int64_t & ne01,
constant int64_t & ne02,
constant uint64_t & nb00,
constant uint64_t & nb01,
constant uint64_t & nb02,
constant int64_t & ne10,
constant int64_t & ne11,
constant int64_t & ne12,
constant uint64_t & nb10,
constant uint64_t & nb11,
constant uint64_t & nb12,
constant int64_t & ne0,
constant int64_t & ne1,
constant uint & r2,
constant uint & r3,
threadgroup int8_t * shared_values [[threadgroup(0)]],
uint3 tgpig[[threadgroup_position_in_grid]],
uint tiisg[[thread_index_in_simdgroup]],
uint sgitg[[simdgroup_index_in_threadgroup]]) {
kernel_mul_mv_iq2_xs_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg);
}
//============================= templates and their specializations ============================= //============================= templates and their specializations =============================
// NOTE: this is not dequantizing - we are simply fitting the template // NOTE: this is not dequantizing - we are simply fitting the template
@ -3973,18 +4254,39 @@ void dequantize_iq2_xxs(device const block_iq2_xxs * xb, short il, thread type4x
const uint32_t aux32_s = q2[2] | (q2[3] << 16); const uint32_t aux32_s = q2[2] | (q2[3] << 16);
thread const uint8_t * aux8 = (thread const uint8_t *)&aux32_g; thread const uint8_t * aux8 = (thread const uint8_t *)&aux32_g;
const float dl = d * (0.5f + (aux32_s >> 28)) * 0.25f; const float dl = d * (0.5f + (aux32_s >> 28)) * 0.25f;
constant uint8_t * grid = (constant uint8_t *)(kgrid_iq2xxs + aux8[2*il+0]); constant uint8_t * grid = (constant uint8_t *)(iq2xxs_grid + aux8[2*il+0]);
uint8_t signs = ksigns_iq2xs[(aux32_s >> 14*il) & 127]; uint8_t signs = ksigns_iq2xs[(aux32_s >> 14*il) & 127];
for (int i = 0; i < 8; ++i) { for (int i = 0; i < 8; ++i) {
reg[i/4][i%4] = dl * grid[i] * (signs & kmask_iq2xs[i] ? -1.f : 1.f); reg[i/4][i%4] = dl * grid[i] * (signs & kmask_iq2xs[i] ? -1.f : 1.f);
} }
grid = (constant uint8_t *)(kgrid_iq2xxs + aux8[2*il+1]); grid = (constant uint8_t *)(iq2xxs_grid + aux8[2*il+1]);
signs = ksigns_iq2xs[(aux32_s >> (14*il+7)) & 127]; signs = ksigns_iq2xs[(aux32_s >> (14*il+7)) & 127];
for (int i = 0; i < 8; ++i) { for (int i = 0; i < 8; ++i) {
reg[2+i/4][i%4] = dl * grid[i] * (signs & kmask_iq2xs[i] ? -1.f : 1.f); reg[2+i/4][i%4] = dl * grid[i] * (signs & kmask_iq2xs[i] ? -1.f : 1.f);
} }
} }
template <typename type4x4>
void dequantize_iq2_xs(device const block_iq2_xs * xb, short il, thread type4x4 & reg) {
// il is 0...15 for QK_K = 256 => index of block of 32 is il/2
const float d = xb->d;
const int ib32 = il/2;
il = il%2;
// il = 0 or 1. il = 0 processes the first 16 quants in a block of 32, il = 1 the second 16
device const uint16_t * q2 = xb->qs + 4*ib32;
const float dl = d * (0.5f + ((xb->scales[ib32] >> 4*il) & 0xf)) * 0.25f;
constant uint8_t * grid = (constant uint8_t *)(iq2xs_grid + (q2[2*il+0] & 511));
uint8_t signs = ksigns_iq2xs[q2[2*il+0] >> 9];
for (int i = 0; i < 8; ++i) {
reg[i/4][i%4] = dl * grid[i] * (signs & kmask_iq2xs[i] ? -1.f : 1.f);
}
grid = (constant uint8_t *)(iq2xs_grid + (q2[2*il+1] & 511));
signs = ksigns_iq2xs[q2[2*il+1] >> 9];
for (int i = 0; i < 8; ++i) {
reg[2+i/4][i%4] = dl * grid[i] * (signs & kmask_iq2xs[i] ? -1.f : 1.f);
}
}
template<typename block_q, short nl, void (*dequantize_func)(device const block_q *, short, thread float4x4 &)> template<typename block_q, short nl, void (*dequantize_func)(device const block_q *, short, thread float4x4 &)>
kernel void kernel_get_rows( kernel void kernel_get_rows(
device const void * src0, device const void * src0,
@ -4525,6 +4827,7 @@ template [[host_name("kernel_get_rows_q4_K")]] kernel get_rows_t kernel_get_rows
template [[host_name("kernel_get_rows_q5_K")]] kernel get_rows_t kernel_get_rows<block_q5_K, QK_NL, dequantize_q5_K>; template [[host_name("kernel_get_rows_q5_K")]] kernel get_rows_t kernel_get_rows<block_q5_K, QK_NL, dequantize_q5_K>;
template [[host_name("kernel_get_rows_q6_K")]] kernel get_rows_t kernel_get_rows<block_q6_K, QK_NL, dequantize_q6_K>; template [[host_name("kernel_get_rows_q6_K")]] kernel get_rows_t kernel_get_rows<block_q6_K, QK_NL, dequantize_q6_K>;
template [[host_name("kernel_get_rows_iq2_xxs")]] kernel get_rows_t kernel_get_rows<block_iq2_xxs, QK_NL, dequantize_iq2_xxs>; template [[host_name("kernel_get_rows_iq2_xxs")]] kernel get_rows_t kernel_get_rows<block_iq2_xxs, QK_NL, dequantize_iq2_xxs>;
template [[host_name("kernel_get_rows_iq2_xs")]] kernel get_rows_t kernel_get_rows<block_iq2_xs, QK_NL, dequantize_iq2_xs>;
// //
// matrix-matrix multiplication // matrix-matrix multiplication
@ -4562,6 +4865,7 @@ template [[host_name("kernel_mul_mm_q4_K_f32")]] kernel mat_mm_t kernel_mul_mm<b
template [[host_name("kernel_mul_mm_q5_K_f32")]] kernel mat_mm_t kernel_mul_mm<block_q5_K, QK_NL, dequantize_q5_K>; template [[host_name("kernel_mul_mm_q5_K_f32")]] kernel mat_mm_t kernel_mul_mm<block_q5_K, QK_NL, dequantize_q5_K>;
template [[host_name("kernel_mul_mm_q6_K_f32")]] kernel mat_mm_t kernel_mul_mm<block_q6_K, QK_NL, dequantize_q6_K>; template [[host_name("kernel_mul_mm_q6_K_f32")]] kernel mat_mm_t kernel_mul_mm<block_q6_K, QK_NL, dequantize_q6_K>;
template [[host_name("kernel_mul_mm_iq2_xxs_f32")]] kernel mat_mm_t kernel_mul_mm<block_iq2_xxs, QK_NL, dequantize_iq2_xxs>; template [[host_name("kernel_mul_mm_iq2_xxs_f32")]] kernel mat_mm_t kernel_mul_mm<block_iq2_xxs, QK_NL, dequantize_iq2_xxs>;
template [[host_name("kernel_mul_mm_iq2_xs_f32")]] kernel mat_mm_t kernel_mul_mm<block_iq2_xs, QK_NL, dequantize_iq2_xs>;
// //
// indirect matrix-matrix multiplication // indirect matrix-matrix multiplication
@ -4611,6 +4915,7 @@ template [[host_name("kernel_mul_mm_id_q4_K_f32")]] kernel mat_mm_id_t kernel_mu
template [[host_name("kernel_mul_mm_id_q5_K_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_q5_K, QK_NL, dequantize_q5_K>; template [[host_name("kernel_mul_mm_id_q5_K_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_q5_K, QK_NL, dequantize_q5_K>;
template [[host_name("kernel_mul_mm_id_q6_K_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_q6_K, QK_NL, dequantize_q6_K>; template [[host_name("kernel_mul_mm_id_q6_K_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_q6_K, QK_NL, dequantize_q6_K>;
template [[host_name("kernel_mul_mm_id_iq2_xxs_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_iq2_xxs, QK_NL, dequantize_iq2_xxs>; template [[host_name("kernel_mul_mm_id_iq2_xxs_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_iq2_xxs, QK_NL, dequantize_iq2_xxs>;
template [[host_name("kernel_mul_mm_id_iq2_xs_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_iq2_xs, QK_NL, dequantize_iq2_xs>;
// //
// matrix-vector multiplication // matrix-vector multiplication
@ -5448,3 +5753,68 @@ kernel void kernel_mul_mv_id_iq2_xxs_f32(
tiisg, tiisg,
sgitg); sgitg);
} }
[[host_name("kernel_mul_mv_id_iq2_xs_f32")]]
kernel void kernel_mul_mv_id_iq2_xs_f32(
device const char * ids,
device const char * src1,
device float * dst,
constant uint64_t & nbi1,
constant int64_t & ne00,
constant int64_t & ne01,
constant int64_t & ne02,
constant uint64_t & nb00,
constant uint64_t & nb01,
constant uint64_t & nb02,
constant int64_t & ne10,
constant int64_t & ne11,
constant int64_t & ne12,
constant int64_t & ne13,
constant uint64_t & nb10,
constant uint64_t & nb11,
constant uint64_t & nb12,
constant int64_t & ne0,
constant int64_t & ne1,
constant uint64_t & nb1,
constant uint & r2,
constant uint & r3,
constant int & idx,
device const char * src00,
device const char * src01,
device const char * src02,
device const char * src03,
device const char * src04,
device const char * src05,
device const char * src06,
device const char * src07,
threadgroup int8_t * shared_values [[threadgroup(0)]],
uint3 tgpig[[threadgroup_position_in_grid]],
uint tiitg[[thread_index_in_threadgroup]],
uint tiisg[[thread_index_in_simdgroup]],
uint sgitg[[simdgroup_index_in_threadgroup]]) {
device const char * src0[8] = {src00, src01, src02, src03, src04, src05, src06, src07};
const int64_t bid = tgpig.z/(ne12*ne13);
tgpig.z = tgpig.z%(ne12*ne13);
const int32_t id = ((device int32_t *) (ids + bid*nbi1))[idx];
kernel_mul_mv_iq2_xs_f32_impl(
src0[id],
(device const float *) (src1 + bid*nb11),
dst + bid*ne0,
ne00,
ne01,
ne02,
ne10,
ne12,
ne0,
ne1,
r2,
r3,
shared_values,
tgpig,
tiisg,
sgitg);
}

360
ggml-quants.c
View File

@ -2342,15 +2342,7 @@ size_t ggml_quantize_q6_K(const float * src, void * dst, int n, int k, int64_t *
// ====================== "True" 2-bit (de)-quantization // ====================== "True" 2-bit (de)-quantization
void quantize_row_iq2_xxs_reference(const float * restrict x, block_iq2_xxs * restrict y, int k) { static const uint64_t iq2xxs_grid[256] = {
(void)x;
(void)y;
(void)k;
assert(k % QK_K == 0);
//fprintf(stderr, "=========================== %s: not implemented\n", __func__);
}
static const uint64_t iq2xxs_grid[256] = {
0x0808080808080808, 0x080808080808082b, 0x0808080808081919, 0x0808080808082b08, 0x0808080808080808, 0x080808080808082b, 0x0808080808081919, 0x0808080808082b08,
0x0808080808082b2b, 0x0808080808190819, 0x0808080808191908, 0x08080808082b0808, 0x0808080808082b2b, 0x0808080808190819, 0x0808080808191908, 0x08080808082b0808,
0x08080808082b082b, 0x08080808082b2b08, 0x08080808082b2b2b, 0x0808080819080819, 0x08080808082b082b, 0x08080808082b2b08, 0x08080808082b2b2b, 0x0808080819080819,
@ -2417,6 +2409,137 @@ static const uint64_t iq2xxs_grid[256] = {
0x2b2b082b08080808, 0x2b2b190808192b08, 0x2b2b2b0819190808, 0x2b2b2b1908081908, 0x2b2b082b08080808, 0x2b2b190808192b08, 0x2b2b2b0819190808, 0x2b2b2b1908081908,
}; };
static const uint64_t iq2xs_grid[512] = {
0x0808080808080808, 0x080808080808082b, 0x0808080808081919, 0x0808080808082b08,
0x0808080808082b2b, 0x0808080808190819, 0x0808080808191908, 0x080808080819192b,
0x0808080808192b19, 0x08080808082b0808, 0x08080808082b082b, 0x08080808082b1919,
0x08080808082b2b08, 0x0808080819080819, 0x0808080819081908, 0x080808081908192b,
0x0808080819082b19, 0x0808080819190808, 0x080808081919082b, 0x0808080819191919,
0x0808080819192b08, 0x08080808192b0819, 0x08080808192b1908, 0x080808082b080808,
0x080808082b08082b, 0x080808082b081919, 0x080808082b082b08, 0x080808082b190819,
0x080808082b191908, 0x080808082b192b19, 0x080808082b2b0808, 0x0808081908080819,
0x0808081908081908, 0x080808190808192b, 0x0808081908082b19, 0x0808081908190808,
0x080808190819082b, 0x0808081908191919, 0x0808081908192b08, 0x0808081908192b2b,
0x08080819082b0819, 0x08080819082b1908, 0x0808081919080808, 0x080808191908082b,
0x0808081919081919, 0x0808081919082b08, 0x0808081919190819, 0x0808081919191908,
0x08080819192b0808, 0x08080819192b2b08, 0x080808192b080819, 0x080808192b081908,
0x080808192b190808, 0x0808082b08080808, 0x0808082b0808082b, 0x0808082b08081919,
0x0808082b08082b08, 0x0808082b08190819, 0x0808082b08191908, 0x0808082b082b0808,
0x0808082b19080819, 0x0808082b19081908, 0x0808082b19190808, 0x0808082b19191919,
0x0808082b2b080808, 0x0808082b2b082b2b, 0x0808190808080819, 0x0808190808081908,
0x080819080808192b, 0x0808190808082b19, 0x0808190808190808, 0x080819080819082b,
0x0808190808191919, 0x0808190808192b08, 0x08081908082b0819, 0x08081908082b1908,
0x0808190819080808, 0x080819081908082b, 0x0808190819081919, 0x0808190819082b08,
0x0808190819190819, 0x0808190819191908, 0x080819081919192b, 0x08081908192b0808,
0x080819082b080819, 0x080819082b081908, 0x080819082b190808, 0x0808191908080808,
0x080819190808082b, 0x0808191908081919, 0x0808191908082b08, 0x0808191908190819,
0x0808191908191908, 0x08081919082b0808, 0x0808191919080819, 0x0808191919081908,
0x0808191919190808, 0x08081919192b0819, 0x080819192b080808, 0x0808192b08080819,
0x0808192b08081908, 0x0808192b08190808, 0x0808192b082b192b, 0x0808192b19080808,
0x0808192b1908082b, 0x0808192b2b081908, 0x08082b0808080808, 0x08082b080808082b,
0x08082b0808081919, 0x08082b0808082b08, 0x08082b0808082b2b, 0x08082b0808190819,
0x08082b0808191908, 0x08082b08082b0808, 0x08082b08082b1919, 0x08082b0819080819,
0x08082b0819081908, 0x08082b0819190808, 0x08082b0819192b08, 0x08082b082b080808,
0x08082b082b2b0808, 0x08082b082b2b2b2b, 0x08082b1908080819, 0x08082b1908081908,
0x08082b1908190808, 0x08082b1919080808, 0x08082b192b080819, 0x08082b192b082b19,
0x08082b2b08080808, 0x08082b2b082b0808, 0x08082b2b082b2b08, 0x08082b2b2b19192b,
0x08082b2b2b2b0808, 0x0819080808080819, 0x0819080808081908, 0x081908080808192b,
0x0819080808082b19, 0x0819080808190808, 0x081908080819082b, 0x0819080808191919,
0x0819080808192b08, 0x08190808082b0819, 0x08190808082b1908, 0x0819080819080808,
0x081908081908082b, 0x0819080819081919, 0x0819080819082b08, 0x0819080819190819,
0x0819080819191908, 0x08190808192b0808, 0x08190808192b2b2b, 0x081908082b080819,
0x081908082b081908, 0x081908082b190808, 0x0819081908080808, 0x081908190808082b,
0x0819081908081919, 0x0819081908082b08, 0x0819081908190819, 0x0819081908191908,
0x08190819082b0808, 0x0819081919080819, 0x0819081919081908, 0x0819081919190808,
0x081908192b080808, 0x081908192b191908, 0x081908192b19192b, 0x0819082b08080819,
0x0819082b08081908, 0x0819082b0808192b, 0x0819082b08190808, 0x0819082b19080808,
0x0819082b192b0808, 0x0819190808080808, 0x081919080808082b, 0x0819190808081919,
0x0819190808082b08, 0x0819190808190819, 0x0819190808191908, 0x08191908082b0808,
0x0819190819080819, 0x0819190819081908, 0x0819190819082b19, 0x0819190819190808,
0x08191908192b1908, 0x081919082b080808, 0x0819191908080819, 0x0819191908081908,
0x0819191908190808, 0x0819191919080808, 0x0819192b08080808, 0x0819192b08191908,
0x0819192b19082b19, 0x08192b0808080819, 0x08192b0808081908, 0x08192b0808190808,
0x08192b080819082b, 0x08192b0819080808, 0x08192b0819191908, 0x08192b082b08192b,
0x08192b1908080808, 0x08192b1908081919, 0x08192b19192b192b, 0x08192b2b19190819,
0x08192b2b2b2b2b19, 0x082b080808080808, 0x082b08080808082b, 0x082b080808081919,
0x082b080808082b08, 0x082b080808082b2b, 0x082b080808190819, 0x082b080808191908,
0x082b0808082b0808, 0x082b080819080819, 0x082b080819081908, 0x082b080819190808,
0x082b08082b080808, 0x082b08082b2b0808, 0x082b081908080819, 0x082b081908081908,
0x082b081908190808, 0x082b081919080808, 0x082b081919082b08, 0x082b0819192b1919,
0x082b082b08080808, 0x082b082b082b082b, 0x082b082b2b080808, 0x082b082b2b2b2b08,
0x082b190808080819, 0x082b190808081908, 0x082b190808190808, 0x082b1908082b2b19,
0x082b190819080808, 0x082b191908080808, 0x082b191919080819, 0x082b19191919082b,
0x082b19192b192b19, 0x082b192b08080819, 0x082b192b08192b2b, 0x082b192b2b2b192b,
0x082b2b0808080808, 0x082b2b0808082b08, 0x082b2b0808082b2b, 0x082b2b08082b0808,
0x082b2b0819191919, 0x082b2b082b082b08, 0x082b2b082b2b082b, 0x082b2b19192b2b08,
0x082b2b192b190808, 0x082b2b2b08082b08, 0x082b2b2b082b0808, 0x082b2b2b2b08082b,
0x082b2b2b2b082b08, 0x082b2b2b2b082b2b, 0x1908080808080819, 0x1908080808081908,
0x190808080808192b, 0x1908080808082b19, 0x1908080808190808, 0x190808080819082b,
0x1908080808191919, 0x1908080808192b08, 0x19080808082b0819, 0x19080808082b1908,
0x1908080819080808, 0x190808081908082b, 0x1908080819081919, 0x1908080819082b08,
0x1908080819082b2b, 0x1908080819190819, 0x1908080819191908, 0x19080808192b0808,
0x19080808192b1919, 0x190808082b080819, 0x190808082b081908, 0x190808082b190808,
0x1908081908080808, 0x190808190808082b, 0x1908081908081919, 0x1908081908082b08,
0x1908081908190819, 0x1908081908191908, 0x19080819082b0808, 0x1908081919080819,
0x1908081919081908, 0x1908081919190808, 0x190808192b080808, 0x190808192b081919,
0x190808192b2b082b, 0x1908082b08080819, 0x1908082b08081908, 0x1908082b08190808,
0x1908082b0819082b, 0x1908082b082b2b19, 0x1908082b19080808, 0x1908190808080808,
0x190819080808082b, 0x1908190808081919, 0x1908190808082b08, 0x1908190808190819,
0x1908190808191908, 0x1908190808192b19, 0x19081908082b0808, 0x1908190819080819,
0x1908190819081908, 0x1908190819190808, 0x190819082b080808, 0x190819082b191908,
0x1908191908080819, 0x1908191908081908, 0x1908191908190808, 0x19081919082b1908,
0x1908191919080808, 0x190819192b192b2b, 0x1908192b08080808, 0x1908192b08082b2b,
0x1908192b19081908, 0x1908192b19190808, 0x19082b0808080819, 0x19082b0808081908,
0x19082b0808190808, 0x19082b0819080808, 0x19082b0819081919, 0x19082b0819191908,
0x19082b08192b082b, 0x19082b1908080808, 0x19082b1908190819, 0x19082b1919081908,
0x19082b1919190808, 0x19082b19192b2b19, 0x19082b2b08081908, 0x1919080808080808,
0x191908080808082b, 0x1919080808081919, 0x1919080808082b08, 0x1919080808190819,
0x1919080808191908, 0x19190808082b0808, 0x19190808082b2b08, 0x1919080819080819,
0x1919080819081908, 0x1919080819190808, 0x191908082b080808, 0x1919081908080819,
0x1919081908081908, 0x1919081908190808, 0x1919081908191919, 0x1919081919080808,
0x191908191908082b, 0x1919082b08080808, 0x1919082b19081908, 0x1919082b2b2b2b2b,
0x1919190808080819, 0x1919190808081908, 0x1919190808190808, 0x19191908082b0819,
0x1919190819080808, 0x19191908192b0808, 0x191919082b080819, 0x191919082b2b0819,
0x1919191908080808, 0x1919191908082b08, 0x191919192b080808, 0x191919192b082b08,
0x1919192b082b0819, 0x1919192b192b2b08, 0x1919192b2b2b0819, 0x19192b0808080808,
0x19192b0808191908, 0x19192b0819080819, 0x19192b0819190808, 0x19192b082b192b19,
0x19192b1908192b2b, 0x19192b1919080808, 0x19192b191908082b, 0x19192b2b2b081919,
0x192b080808080819, 0x192b080808081908, 0x192b080808190808, 0x192b080819080808,
0x192b080819191908, 0x192b0808192b082b, 0x192b08082b08192b, 0x192b08082b2b2b19,
0x192b081908080808, 0x192b082b082b1908, 0x192b082b19082b2b, 0x192b082b2b19082b,
0x192b190808080808, 0x192b19080819192b, 0x192b191908190808, 0x192b191919080808,
0x192b191919081919, 0x192b19192b2b1908, 0x192b2b0808080819, 0x192b2b08192b2b2b,
0x192b2b19082b1919, 0x192b2b2b0808192b, 0x192b2b2b19191908, 0x192b2b2b192b082b,
0x2b08080808080808, 0x2b0808080808082b, 0x2b08080808081919, 0x2b08080808082b08,
0x2b08080808190819, 0x2b08080808191908, 0x2b080808082b0808, 0x2b080808082b2b2b,
0x2b08080819080819, 0x2b08080819081908, 0x2b08080819190808, 0x2b0808082b080808,
0x2b0808082b08082b, 0x2b0808082b2b2b08, 0x2b0808082b2b2b2b, 0x2b08081908080819,
0x2b08081908081908, 0x2b0808190808192b, 0x2b08081908190808, 0x2b08081919080808,
0x2b08081919190819, 0x2b08081919192b19, 0x2b08082b08080808, 0x2b08082b082b0808,
0x2b08082b2b080808, 0x2b08082b2b08082b, 0x2b08082b2b2b0808, 0x2b08082b2b2b2b08,
0x2b08190808080819, 0x2b08190808081908, 0x2b08190808190808, 0x2b0819080819082b,
0x2b08190808191919, 0x2b08190819080808, 0x2b081908192b0808, 0x2b0819082b082b19,
0x2b08191908080808, 0x2b08191919081908, 0x2b0819192b2b1919, 0x2b08192b08192b08,
0x2b08192b192b2b2b, 0x2b082b0808080808, 0x2b082b0808082b08, 0x2b082b08082b1919,
0x2b082b0819192b2b, 0x2b082b082b080808, 0x2b082b082b08082b, 0x2b082b082b2b2b08,
0x2b082b190808192b, 0x2b082b2b082b082b, 0x2b082b2b2b080808, 0x2b082b2b2b082b08,
0x2b082b2b2b19192b, 0x2b082b2b2b2b2b08, 0x2b19080808080819, 0x2b19080808081908,
0x2b19080808190808, 0x2b19080819080808, 0x2b1908081919192b, 0x2b1908082b081908,
0x2b19081908080808, 0x2b190819082b082b, 0x2b190819192b1908, 0x2b19082b1919192b,
0x2b19082b2b082b19, 0x2b19190808080808, 0x2b19190808081919, 0x2b19190819081908,
0x2b19190819190808, 0x2b19190819192b08, 0x2b191919082b2b19, 0x2b1919192b190808,
0x2b1919192b19082b, 0x2b19192b19080819, 0x2b192b0819190819, 0x2b192b082b2b192b,
0x2b192b1919082b19, 0x2b192b2b08191919, 0x2b192b2b192b0808, 0x2b2b080808080808,
0x2b2b08080808082b, 0x2b2b080808082b08, 0x2b2b080808082b2b, 0x2b2b0808082b0808,
0x2b2b0808082b2b2b, 0x2b2b08082b2b0808, 0x2b2b081919190819, 0x2b2b081919192b19,
0x2b2b08192b2b192b, 0x2b2b082b08080808, 0x2b2b082b0808082b, 0x2b2b082b08082b08,
0x2b2b082b082b2b2b, 0x2b2b082b2b080808, 0x2b2b082b2b2b0808, 0x2b2b190819080808,
0x2b2b19082b191919, 0x2b2b192b192b1919, 0x2b2b192b2b192b08, 0x2b2b2b0808082b2b,
0x2b2b2b08082b0808, 0x2b2b2b08082b082b, 0x2b2b2b08082b2b08, 0x2b2b2b082b2b0808,
0x2b2b2b082b2b2b08, 0x2b2b2b1908081908, 0x2b2b2b192b081908, 0x2b2b2b192b08192b,
0x2b2b2b2b082b2b08, 0x2b2b2b2b082b2b2b, 0x2b2b2b2b2b190819, 0x2b2b2b2b2b2b2b2b,
};
static const uint8_t ksigns_iq2xs[128] = { static const uint8_t ksigns_iq2xs[128] = {
0, 129, 130, 3, 132, 5, 6, 135, 136, 9, 10, 139, 12, 141, 142, 15, 0, 129, 130, 3, 132, 5, 6, 135, 136, 9, 10, 139, 12, 141, 142, 15,
144, 17, 18, 147, 20, 149, 150, 23, 24, 153, 154, 27, 156, 29, 30, 159, 144, 17, 18, 147, 20, 149, 150, 23, 24, 153, 154, 27, 156, 29, 30, 159,
@ -2427,8 +2550,17 @@ static const uint8_t ksigns_iq2xs[128] = {
96, 225, 226, 99, 228, 101, 102, 231, 232, 105, 106, 235, 108, 237, 238, 111, 96, 225, 226, 99, 228, 101, 102, 231, 232, 105, 106, 235, 108, 237, 238, 111,
240, 113, 114, 243, 116, 245, 246, 119, 120, 249, 250, 123, 252, 125, 126, 255, 240, 113, 114, 243, 116, 245, 246, 119, 120, 249, 250, 123, 252, 125, 126, 255,
}; };
static const uint8_t kmask_iq2xs[8] = {1, 2, 4, 8, 16, 32, 64, 128}; static const uint8_t kmask_iq2xs[8] = {1, 2, 4, 8, 16, 32, 64, 128};
void quantize_row_iq2_xxs_reference(const float * restrict x, block_iq2_xxs * restrict y, int k) {
(void)x;
(void)y;
(void)k;
assert(k % QK_K == 0);
//fprintf(stderr, "=========================== %s: not implemented\n", __func__);
}
void dequantize_row_iq2_xxs(const block_iq2_xxs * restrict x, float * restrict y, int k) { void dequantize_row_iq2_xxs(const block_iq2_xxs * restrict x, float * restrict y, int k) {
assert(k % QK_K == 0); assert(k % QK_K == 0);
const int nb = k / QK_K; const int nb = k / QK_K;
@ -2472,6 +2604,58 @@ size_t ggml_quantize_iq2_xxs(const float * src, void * dst, int n, int k, int64_
return (n/QK_K*sizeof(block_iq2_xxs)); return (n/QK_K*sizeof(block_iq2_xxs));
} }
// ====================== 2.3125 bpw (de)-quantization
void quantize_row_iq2_xs_reference(const float * restrict x, block_iq2_xs * restrict y, int k) {
(void)x;
(void)y;
(void)k;
assert(k % QK_K == 0);
//fprintf(stderr, "=========================== %s: not implemented\n", __func__);
}
void dequantize_row_iq2_xs(const block_iq2_xs * restrict x, float * restrict y, int k) {
assert(k % QK_K == 0);
const int nb = k / QK_K;
float db[2];
for (int i = 0; i < nb; i++) {
const float d = GGML_FP16_TO_FP32(x[i].d);
for (int ib32 = 0; ib32 < QK_K/32; ++ib32) {
db[0] = d * (0.5f + (x[i].scales[ib32] & 0xf)) * 0.25f;
db[1] = d * (0.5f + (x[i].scales[ib32] >> 4)) * 0.25f;
for (int l = 0; l < 4; ++l) {
const uint8_t * grid = (const uint8_t *)(iq2xs_grid + (x[i].qs[4*ib32 + l] & 511));
const uint8_t signs = ksigns_iq2xs[x[i].qs[4*ib32 + l] >> 9];
for (int j = 0; j < 8; ++j) {
y[j] = db[l/2] * grid[j] * (signs & kmask_iq2xs[j] ? -1.f : 1.f);
}
y += 8;
}
}
}
}
void quantize_row_iq2_xs(const float * restrict x, void * restrict vy, int k) {
assert(k % QK_K == 0);
block_iq2_xs * restrict y = vy;
quantize_row_iq2_xs_reference(x, y, k);
}
size_t ggml_quantize_iq2_xs(const float * src, void * dst, int n, int k, int64_t * hist) {
assert(k % QK_K == 0);
(void)hist; // TODO: collect histograms
for (int j = 0; j < n; j += k) {
block_iq2_xs * restrict y = (block_iq2_xs *)dst + j/QK_K;
quantize_row_iq2_xs_reference(src + j, y, k);
}
return (n/QK_K*sizeof(block_iq2_xs));
}
//===================================== Q8_K ============================================== //===================================== Q8_K ==============================================
void quantize_row_q8_K_reference(const float * restrict x, block_q8_K * restrict y, int k) { void quantize_row_q8_K_reference(const float * restrict x, block_q8_K * restrict y, int k) {
@ -7357,3 +7541,161 @@ void ggml_vec_dot_iq2_xxs_q8_K(const int n, float * restrict s, const void * res
*s = 0.125f * sumf; *s = 0.125f * sumf;
#endif #endif
} }
void ggml_vec_dot_iq2_xs_q8_K(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) {
assert(n % QK_K == 0);
const block_iq2_xs * restrict x = vx;
const block_q8_K * restrict y = vy;
const int nb = n / QK_K;
#if defined(__ARM_NEON)
const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs;
int8x16x4_t q2u;
int8x16x4_t q2s;
int8x16x4_t q8b;
int32x4x4_t scales32;
float sumf = 0;
for (int i = 0; i < nb; ++i) {
const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d;
const uint16_t * restrict q2 = x[i].qs;
const int8_t * restrict q8 = y[i].qs;
const uint8x8_t scales8 = vld1_u8(x[i].scales);
const uint8x8_t scales_l = vand_u8(scales8, vdup_n_u8(0xf));
const uint8x8_t scales_h = vshr_n_u8(scales8, 4);
uint8x16_t scales = vcombine_u8(vzip1_u8(scales_l, scales_h), vzip2_u8(scales_l, scales_h));
scales = vaddq_u8(vshlq_n_u8(scales, 1), vdupq_n_u8(1));
const uint16x8_t scales1 = vmovl_u8(vget_low_u8(scales));
const uint16x8_t scales2 = vmovl_u8(vget_high_u8(scales));
scales32.val[0] = vreinterpretq_s32_u32(vmovl_u16(vget_low_u16(scales1)));
scales32.val[1] = vreinterpretq_s32_u32(vmovl_u16(vget_high_u16(scales1)));
scales32.val[2] = vreinterpretq_s32_u32(vmovl_u16(vget_low_u16(scales2)));
scales32.val[3] = vreinterpretq_s32_u32(vmovl_u16(vget_high_u16(scales2)));
int32x4_t sumi = vdupq_n_s32(0);
for (int ib64 = 0; ib64 < QK_K/64; ++ib64) {
q8b = vld1q_s8_x4(q8); q8 += 64;
q2u.val[0] = vcombine_s8(vld1_s8((const void *)(iq2xs_grid + (q2[0] & 511))), vld1_s8((const void *)(iq2xs_grid + (q2[1] & 511))));
q2u.val[1] = vcombine_s8(vld1_s8((const void *)(iq2xs_grid + (q2[2] & 511))), vld1_s8((const void *)(iq2xs_grid + (q2[3] & 511))));
q2u.val[2] = vcombine_s8(vld1_s8((const void *)(iq2xs_grid + (q2[4] & 511))), vld1_s8((const void *)(iq2xs_grid + (q2[5] & 511))));
q2u.val[3] = vcombine_s8(vld1_s8((const void *)(iq2xs_grid + (q2[6] & 511))), vld1_s8((const void *)(iq2xs_grid + (q2[7] & 511))));
q2s.val[0] = vcombine_s8(vld1_s8((const void *)(signs64 + (q2[0] >> 9))), vld1_s8((const void *)(signs64 + (q2[1] >> 9))));
q2s.val[1] = vcombine_s8(vld1_s8((const void *)(signs64 + (q2[2] >> 9))), vld1_s8((const void *)(signs64 + (q2[3] >> 9))));
q2s.val[2] = vcombine_s8(vld1_s8((const void *)(signs64 + (q2[4] >> 9))), vld1_s8((const void *)(signs64 + (q2[5] >> 9))));
q2s.val[3] = vcombine_s8(vld1_s8((const void *)(signs64 + (q2[6] >> 9))), vld1_s8((const void *)(signs64 + (q2[7] >> 9))));
q2u.val[0] = vmulq_s8(q2u.val[0], q2s.val[0]);
q2u.val[1] = vmulq_s8(q2u.val[1], q2s.val[1]);
q2u.val[2] = vmulq_s8(q2u.val[2], q2s.val[2]);
q2u.val[3] = vmulq_s8(q2u.val[3], q2s.val[3]);
const int32x4_t p1 = ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[0], q8b.val[0]);
const int32x4_t p2 = ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[1], q8b.val[1]);
const int32x4_t p3 = ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[2], q8b.val[2]);
const int32x4_t p4 = ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[3], q8b.val[3]);
const int32x4_t p = vpaddq_s32(vpaddq_s32(p1, p2), vpaddq_s32(p3, p4));
sumi = vmlaq_s32(sumi, p, scales32.val[ib64]);
q2 += 8;
}
sumf += d*vaddvq_s32(sumi);
}
*s = 0.125f * sumf;
#elif defined(__AVX2__)
const __m128i m4 = _mm_set1_epi8(0xf);
const __m128i m1 = _mm_set1_epi8(1);
const __m128i m511 = _mm_set1_epi16(511);
const __m128i m127 = _mm_set1_epi16(127);
const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs;
uint64_t aux64;
// somewhat hacky, but gives a significant boost in performance
__m128i aux_gindex, aux_sindex;
const uint16_t * gindex = (const uint16_t *)&aux_gindex;
const uint16_t * sindex = (const uint16_t *)&aux_sindex;
__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;
const uint16_t * restrict q2 = x[i].qs;
const int8_t * restrict q8 = y[i].qs;
memcpy(&aux64, x[i].scales, 8);
__m128i stmp = _mm_set1_epi64x(aux64);
stmp = _mm_unpacklo_epi8(_mm_and_si128(stmp, m4), _mm_and_si128(_mm_srli_epi16(stmp, 4), m4));
const __m128i scales = _mm_add_epi8(_mm_slli_epi16(stmp, 1), m1);
__m256i sumi1 = _mm256_setzero_si256();
__m256i sumi2 = _mm256_setzero_si256();
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 __m128i q2_data = _mm_loadu_si128((const __m128i*)q2); q2 += 8;
aux_gindex = _mm_and_si128(q2_data, m511);
aux_sindex = _mm_and_si128(_mm_srli_epi16(q2_data, 9), m127);
const __m256i q2_1 = _mm256_set_epi64x(iq2xs_grid[gindex[3]], iq2xs_grid[gindex[2]], iq2xs_grid[gindex[1]], iq2xs_grid[gindex[0]]);
const __m256i q2_2 = _mm256_set_epi64x(iq2xs_grid[gindex[7]], iq2xs_grid[gindex[6]], iq2xs_grid[gindex[5]], iq2xs_grid[gindex[4]]);
const __m256i s2_1 = _mm256_set_epi64x(signs64[sindex[3]], signs64[sindex[2]], signs64[sindex[1]], signs64[sindex[0]]);
const __m256i s2_2 = _mm256_set_epi64x(signs64[sindex[7]], signs64[sindex[6]], signs64[sindex[5]], signs64[sindex[4]]);
const __m256i q8s_1 = _mm256_sign_epi8(q8_1, s2_1);
const __m256i q8s_2 = _mm256_sign_epi8(q8_2, s2_2);
const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1);
const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2);
const __m256i sc1 = _mm256_cvtepi8_epi16(_mm_shuffle_epi8(scales, get_scale_shuffle(ib32+0)));
const __m256i sc2 = _mm256_cvtepi8_epi16(_mm_shuffle_epi8(scales, get_scale_shuffle(ib32+1)));
sumi1 = _mm256_add_epi32(sumi1, _mm256_madd_epi16(dot1, sc1));
sumi2 = _mm256_add_epi32(sumi2, _mm256_madd_epi16(dot2, sc2));
}
accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accumf);
}
*s = 0.125f * hsum_float_8(accumf);
#else
float sumf = 0.f;
for (int i = 0; i < nb; ++i) {
const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d;
const uint16_t * restrict q2 = x[i].qs;
const uint8_t * restrict sc = x[i].scales;
const int8_t * restrict q8 = y[i].qs;
int32_t bsum = 0;
for (int ib32 = 0; ib32 < QK_K/32; ++ib32) {
const uint16_t ls1 = 2*(sc[ib32] & 0xf) + 1;
const uint16_t ls2 = 2*(sc[ib32] >> 4) + 1;
int32_t sumi = 0;
for (int l = 0; l < 2; ++l) {
const uint8_t * grid = (const uint8_t *)(iq2xs_grid + (q2[l] & 511));
const uint8_t signs = ksigns_iq2xs[q2[l] >> 9];
for (int j = 0; j < 8; ++j) {
sumi += grid[j] * q8[j] * (signs & kmask_iq2xs[j] ? -1 : 1);
}
q8 += 8;
}
bsum += sumi * ls1;
sumi = 0;
for (int l = 2; l < 4; ++l) {
const uint8_t * grid = (const uint8_t *)(iq2xs_grid + (q2[l] & 511));
const uint8_t signs = ksigns_iq2xs[q2[l] >> 9];
for (int j = 0; j < 8; ++j) {
sumi += grid[j] * q8[j] * (signs & kmask_iq2xs[j] ? -1 : 1);
}
q8 += 8;
}
bsum += sumi * ls2;
q2 += 4;
}
sumf += d * bsum;
}
*s = 0.125f * sumf;
#endif
}

12
ggml-quants.h
View File

@ -174,6 +174,14 @@ typedef struct {
} block_iq2_xxs; } block_iq2_xxs;
static_assert(sizeof(block_iq2_xxs) == sizeof(ggml_fp16_t) + QK_K/8*sizeof(uint16_t), "wrong iq2_xxs block size/padding"); static_assert(sizeof(block_iq2_xxs) == sizeof(ggml_fp16_t) + QK_K/8*sizeof(uint16_t), "wrong iq2_xxs block size/padding");
// 2.3125 bpw quants
typedef struct {
ggml_fp16_t d;
uint16_t qs[QK_K/8];
uint8_t scales[QK_K/32];
} block_iq2_xs;
static_assert(sizeof(block_iq2_xs) == sizeof(ggml_fp16_t) + QK_K/8*sizeof(uint16_t) + QK_K/32, "wrong iq2_xs block size/padding");
// Quantization // Quantization
void quantize_row_q4_0_reference(const float * restrict x, block_q4_0 * restrict y, int k); void quantize_row_q4_0_reference(const float * restrict x, block_q4_0 * restrict y, int k);
void quantize_row_q4_1_reference(const float * restrict x, block_q4_1 * restrict y, int k); void quantize_row_q4_1_reference(const float * restrict x, block_q4_1 * restrict y, int k);
@ -189,6 +197,7 @@ void quantize_row_q5_K_reference(const float * restrict x, block_q5_K * restrict
void quantize_row_q6_K_reference(const float * restrict x, block_q6_K * restrict y, int k); void quantize_row_q6_K_reference(const float * restrict x, block_q6_K * restrict y, int k);
void quantize_row_q8_K_reference(const float * restrict x, block_q8_K * restrict y, int k); void quantize_row_q8_K_reference(const float * restrict x, block_q8_K * restrict y, int k);
void quantize_row_iq2_xxs_reference(const float * restrict x, block_iq2_xxs * restrict y, int k); void quantize_row_iq2_xxs_reference(const float * restrict x, block_iq2_xxs * restrict y, int k);
void quantize_row_iq2_xs_reference (const float * restrict x, block_iq2_xs * restrict y, int k);
void quantize_row_q4_0(const float * restrict x, void * restrict y, int k); void quantize_row_q4_0(const float * restrict x, void * restrict y, int k);
void quantize_row_q4_1(const float * restrict x, void * restrict y, int k); void quantize_row_q4_1(const float * restrict x, void * restrict y, int k);
@ -204,6 +213,7 @@ void quantize_row_q5_K(const float * restrict x, void * restrict y, int k);
void quantize_row_q6_K(const float * restrict x, void * restrict y, int k); void quantize_row_q6_K(const float * restrict x, void * restrict y, int k);
void quantize_row_q8_K(const float * restrict x, void * restrict y, int k); void quantize_row_q8_K(const float * restrict x, void * restrict y, int k);
void quantize_row_iq2_xxs(const float * restrict x, void * restrict y, int k); void quantize_row_iq2_xxs(const float * restrict x, void * restrict y, int k);
void quantize_row_iq2_xs (const float * restrict x, void * restrict y, int k);
// Dequantization // Dequantization
void dequantize_row_q4_0(const block_q4_0 * restrict x, float * restrict y, int k); void dequantize_row_q4_0(const block_q4_0 * restrict x, float * restrict y, int k);
@ -220,6 +230,7 @@ void dequantize_row_q5_K(const block_q5_K * restrict x, float * restrict y, int
void dequantize_row_q6_K(const block_q6_K * restrict x, float * restrict y, int k); void dequantize_row_q6_K(const block_q6_K * restrict x, float * restrict y, int k);
void dequantize_row_q8_K(const block_q8_K * restrict x, float * restrict y, int k); void dequantize_row_q8_K(const block_q8_K * restrict x, float * restrict y, int k);
void dequantize_row_iq2_xxs(const block_iq2_xxs * restrict x, float * restrict y, int k); void dequantize_row_iq2_xxs(const block_iq2_xxs * restrict x, float * restrict y, int k);
void dequantize_row_iq2_xs (const block_iq2_xs * restrict x, float * restrict y, int k);
// Dot product // Dot product
void ggml_vec_dot_q4_0_q8_0(int n, float * restrict s, const void * restrict vx, const void * restrict vy); void ggml_vec_dot_q4_0_q8_0(int n, float * restrict s, const void * restrict vx, const void * restrict vy);
@ -234,3 +245,4 @@ void ggml_vec_dot_q4_K_q8_K(int n, float * restrict s, const void * restrict vx,
void ggml_vec_dot_q5_K_q8_K(int n, float * restrict s, const void * restrict vx, const void * restrict vy); void ggml_vec_dot_q5_K_q8_K(int n, float * restrict s, const void * restrict vx, const void * restrict vy);
void ggml_vec_dot_q6_K_q8_K(int n, float * restrict s, const void * restrict vx, const void * restrict vy); void ggml_vec_dot_q6_K_q8_K(int n, float * restrict s, const void * restrict vx, const void * restrict vy);
void ggml_vec_dot_iq2_xxs_q8_K(int n, float * restrict s, const void * restrict vx, const void * restrict vy); void ggml_vec_dot_iq2_xxs_q8_K(int n, float * restrict s, const void * restrict vx, const void * restrict vy);
void ggml_vec_dot_iq2_xs_q8_K (int n, float * restrict s, const void * restrict vx, const void * restrict vy);

30
ggml.c
View File

@ -584,6 +584,17 @@ static const ggml_type_traits_t type_traits[GGML_TYPE_COUNT] = {
.vec_dot = ggml_vec_dot_iq2_xxs_q8_K, .vec_dot = ggml_vec_dot_iq2_xxs_q8_K,
.vec_dot_type = GGML_TYPE_Q8_K, .vec_dot_type = GGML_TYPE_Q8_K,
}, },
[GGML_TYPE_IQ2_XS] = {
.type_name = "iq2_xs",
.blck_size = QK_K,
.type_size = sizeof(block_iq2_xs),
.is_quantized = true,
.to_float = (ggml_to_float_t) dequantize_row_iq2_xs,
.from_float = quantize_row_iq2_xs,
.from_float_reference = (ggml_from_float_t) quantize_row_iq2_xs_reference,
.vec_dot = ggml_vec_dot_iq2_xs_q8_K,
.vec_dot_type = GGML_TYPE_Q8_K,
},
[GGML_TYPE_Q8_K] = { [GGML_TYPE_Q8_K] = {
.type_name = "q8_K", .type_name = "q8_K",
.blck_size = QK_K, .blck_size = QK_K,
@ -2123,6 +2134,7 @@ enum ggml_type ggml_ftype_to_ggml_type(enum ggml_ftype ftype) {
case GGML_FTYPE_MOSTLY_Q5_K: wtype = GGML_TYPE_Q5_K; break; case GGML_FTYPE_MOSTLY_Q5_K: wtype = GGML_TYPE_Q5_K; break;
case GGML_FTYPE_MOSTLY_Q6_K: wtype = GGML_TYPE_Q6_K; break; case GGML_FTYPE_MOSTLY_Q6_K: wtype = GGML_TYPE_Q6_K; break;
case GGML_FTYPE_MOSTLY_IQ2_XXS: wtype = GGML_TYPE_IQ2_XXS; break; case GGML_FTYPE_MOSTLY_IQ2_XXS: wtype = GGML_TYPE_IQ2_XXS; break;
case GGML_FTYPE_MOSTLY_IQ2_XS: wtype = GGML_TYPE_IQ2_XS; break;
case GGML_FTYPE_UNKNOWN: wtype = GGML_TYPE_COUNT; break; case GGML_FTYPE_UNKNOWN: wtype = GGML_TYPE_COUNT; break;
case GGML_FTYPE_MOSTLY_Q4_1_SOME_F16: wtype = GGML_TYPE_COUNT; break; case GGML_FTYPE_MOSTLY_Q4_1_SOME_F16: wtype = GGML_TYPE_COUNT; break;
} }
@ -7435,6 +7447,7 @@ static void ggml_compute_forward_add(
case GGML_TYPE_Q5_K: case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
case GGML_TYPE_IQ2_XXS: case GGML_TYPE_IQ2_XXS:
case GGML_TYPE_IQ2_XS:
{ {
ggml_compute_forward_add_q_f32(params, src0, src1, dst); ggml_compute_forward_add_q_f32(params, src0, src1, dst);
} break; } break;
@ -7700,6 +7713,7 @@ static void ggml_compute_forward_add1(
case GGML_TYPE_Q5_K: case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
case GGML_TYPE_IQ2_XXS: case GGML_TYPE_IQ2_XXS:
case GGML_TYPE_IQ2_XS:
{ {
ggml_compute_forward_add1_q_f32(params, src0, src1, dst); ggml_compute_forward_add1_q_f32(params, src0, src1, dst);
} break; } break;
@ -7815,6 +7829,7 @@ static void ggml_compute_forward_acc(
case GGML_TYPE_Q5_K: case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
case GGML_TYPE_IQ2_XXS: case GGML_TYPE_IQ2_XXS:
case GGML_TYPE_IQ2_XS:
default: default:
{ {
GGML_ASSERT(false); GGML_ASSERT(false);
@ -10457,6 +10472,7 @@ static void ggml_compute_forward_out_prod(
case GGML_TYPE_Q5_K: case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
case GGML_TYPE_IQ2_XXS: case GGML_TYPE_IQ2_XXS:
case GGML_TYPE_IQ2_XS:
{ {
ggml_compute_forward_out_prod_q_f32(params, src0, src1, dst); ggml_compute_forward_out_prod_q_f32(params, src0, src1, dst);
} break; } break;
@ -10632,6 +10648,7 @@ static void ggml_compute_forward_set(
case GGML_TYPE_Q5_K: case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
case GGML_TYPE_IQ2_XXS: case GGML_TYPE_IQ2_XXS:
case GGML_TYPE_IQ2_XS:
default: default:
{ {
GGML_ASSERT(false); GGML_ASSERT(false);
@ -10827,6 +10844,7 @@ static void ggml_compute_forward_get_rows(
case GGML_TYPE_Q5_K: case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
case GGML_TYPE_IQ2_XXS: case GGML_TYPE_IQ2_XXS:
case GGML_TYPE_IQ2_XS:
{ {
ggml_compute_forward_get_rows_q(params, src0, src1, dst); ggml_compute_forward_get_rows_q(params, src0, src1, dst);
} break; } break;
@ -11464,6 +11482,7 @@ static void ggml_compute_forward_alibi(
case GGML_TYPE_Q5_K: case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
case GGML_TYPE_IQ2_XXS: case GGML_TYPE_IQ2_XXS:
case GGML_TYPE_IQ2_XS:
case GGML_TYPE_Q8_K: case GGML_TYPE_Q8_K:
case GGML_TYPE_I8: case GGML_TYPE_I8:
case GGML_TYPE_I16: case GGML_TYPE_I16:
@ -11539,6 +11558,7 @@ static void ggml_compute_forward_clamp(
case GGML_TYPE_Q5_K: case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
case GGML_TYPE_IQ2_XXS: case GGML_TYPE_IQ2_XXS:
case GGML_TYPE_IQ2_XS:
case GGML_TYPE_Q8_K: case GGML_TYPE_Q8_K:
case GGML_TYPE_I8: case GGML_TYPE_I8:
case GGML_TYPE_I16: case GGML_TYPE_I16:
@ -18660,6 +18680,12 @@ size_t ggml_quantize_chunk(enum ggml_type type, const float * src, void * dst, i
block_iq2_xxs * block = (block_iq2_xxs*)dst + start / QK_K; block_iq2_xxs * block = (block_iq2_xxs*)dst + start / QK_K;
result = ggml_quantize_iq2_xxs(src + start, block, n, n, hist); result = ggml_quantize_iq2_xxs(src + start, block, n, n, hist);
} break; } break;
case GGML_TYPE_IQ2_XS:
{
GGML_ASSERT(start % QK_K == 0);
block_iq2_xs * block = (block_iq2_xs*)dst + start / QK_K;
result = ggml_quantize_iq2_xs(src + start, block, n, n, hist);
} break;
case GGML_TYPE_F16: case GGML_TYPE_F16:
{ {
int elemsize = sizeof(ggml_fp16_t); int elemsize = sizeof(ggml_fp16_t);
@ -19015,8 +19041,8 @@ struct gguf_context * gguf_init_from_file(const char * fname, struct gguf_init_p
(int64_t) info->ne[3]; (int64_t) info->ne[3];
if (ne % ggml_blck_size(info->type) != 0) { if (ne % ggml_blck_size(info->type) != 0) {
fprintf(stderr, "%s: tensor '%s' number of elements (%" PRId64 ") is not a multiple of block size (%d)\n", fprintf(stderr, "%s: tensor '%s' of type %d (%s) number of elements (%" PRId64 ") is not a multiple of block size (%d)\n",
__func__, info->name.data, ne, ggml_blck_size(info->type)); __func__, info->name.data, (int)info->type, ggml_type_name(info->type), ne, ggml_blck_size(info->type));
fclose(file); fclose(file);
gguf_free(ctx); gguf_free(ctx);
return NULL; return NULL;

3
ggml.h
View File

@ -342,6 +342,7 @@ extern "C" {
GGML_TYPE_Q6_K = 14, GGML_TYPE_Q6_K = 14,
GGML_TYPE_Q8_K = 15, GGML_TYPE_Q8_K = 15,
GGML_TYPE_IQ2_XXS = 16, GGML_TYPE_IQ2_XXS = 16,
GGML_TYPE_IQ2_XS = 17,
GGML_TYPE_I8, GGML_TYPE_I8,
GGML_TYPE_I16, GGML_TYPE_I16,
GGML_TYPE_I32, GGML_TYPE_I32,
@ -377,6 +378,7 @@ extern "C" {
GGML_FTYPE_MOSTLY_Q5_K = 13, // except 1d tensors GGML_FTYPE_MOSTLY_Q5_K = 13, // except 1d tensors
GGML_FTYPE_MOSTLY_Q6_K = 14, // except 1d tensors GGML_FTYPE_MOSTLY_Q6_K = 14, // except 1d tensors
GGML_FTYPE_MOSTLY_IQ2_XXS = 15, // except 1d tensors GGML_FTYPE_MOSTLY_IQ2_XXS = 15, // except 1d tensors
GGML_FTYPE_MOSTLY_IQ2_XS = 16, // except 1d tensors
}; };
// available tensor operations: // available tensor operations:
@ -2061,6 +2063,7 @@ extern "C" {
GGML_API size_t ggml_quantize_q5_K(const float * src, void * dst, int n, int k, int64_t * hist); GGML_API size_t ggml_quantize_q5_K(const float * src, void * dst, int n, int k, int64_t * hist);
GGML_API size_t ggml_quantize_q6_K(const float * src, void * dst, int n, int k, int64_t * hist); GGML_API size_t ggml_quantize_q6_K(const float * src, void * dst, int n, int k, int64_t * hist);
GGML_API size_t ggml_quantize_iq2_xxs(const float * src, void * dst, int n, int k, int64_t * hist); GGML_API size_t ggml_quantize_iq2_xxs(const float * src, void * dst, int n, int k, int64_t * hist);
GGML_API size_t ggml_quantize_iq2_xs (const float * src, void * dst, int n, int k, int64_t * hist);
GGML_API size_t ggml_quantize_chunk(enum ggml_type type, const float * src, void * dst, int start, int n, int64_t * hist); GGML_API size_t ggml_quantize_chunk(enum ggml_type type, const float * src, void * dst, int start, int n, int64_t * hist);

3
llama.cpp
View File

@ -2223,6 +2223,7 @@ struct llama_model_loader {
case GGML_TYPE_Q5_K: ftype = LLAMA_FTYPE_MOSTLY_Q5_K_M; break; case GGML_TYPE_Q5_K: ftype = LLAMA_FTYPE_MOSTLY_Q5_K_M; break;
case GGML_TYPE_Q6_K: ftype = LLAMA_FTYPE_MOSTLY_Q6_K; break; case GGML_TYPE_Q6_K: ftype = LLAMA_FTYPE_MOSTLY_Q6_K; break;
case GGML_TYPE_IQ2_XXS: ftype = LLAMA_FTYPE_MOSTLY_IQ2_XXS; break; case GGML_TYPE_IQ2_XXS: ftype = LLAMA_FTYPE_MOSTLY_IQ2_XXS; break;
case GGML_TYPE_IQ2_XS: ftype = LLAMA_FTYPE_MOSTLY_IQ2_XS; break;
default: default:
{ {
LLAMA_LOG_WARN("%s: unknown type %s\n", __func__, ggml_type_name(type_max)); LLAMA_LOG_WARN("%s: unknown type %s\n", __func__, ggml_type_name(type_max));
@ -2595,6 +2596,7 @@ static std::string llama_model_ftype_name(llama_ftype ftype) {
case LLAMA_FTYPE_MOSTLY_Q5_K_M: return "Q5_K - Medium"; case LLAMA_FTYPE_MOSTLY_Q5_K_M: return "Q5_K - Medium";
case LLAMA_FTYPE_MOSTLY_Q6_K: return "Q6_K"; case LLAMA_FTYPE_MOSTLY_Q6_K: return "Q6_K";
case LLAMA_FTYPE_MOSTLY_IQ2_XXS:return "IQ2_XSS - 2.0625 bpw"; case LLAMA_FTYPE_MOSTLY_IQ2_XXS:return "IQ2_XSS - 2.0625 bpw";
case LLAMA_FTYPE_MOSTLY_IQ2_XS: return "IQ2_XS - 2.3125 bpw";
default: return "unknown, may not work"; default: return "unknown, may not work";
} }
@ -9050,6 +9052,7 @@ static void llama_model_quantize_internal(const std::string & fname_inp, const s
case LLAMA_FTYPE_MOSTLY_Q5_K_M: quantized_type = GGML_TYPE_Q5_K; break; case LLAMA_FTYPE_MOSTLY_Q5_K_M: quantized_type = GGML_TYPE_Q5_K; break;
case LLAMA_FTYPE_MOSTLY_Q6_K: quantized_type = GGML_TYPE_Q6_K; break; case LLAMA_FTYPE_MOSTLY_Q6_K: quantized_type = GGML_TYPE_Q6_K; break;
case LLAMA_FTYPE_MOSTLY_IQ2_XXS:quantized_type = GGML_TYPE_IQ2_XXS; break; case LLAMA_FTYPE_MOSTLY_IQ2_XXS:quantized_type = GGML_TYPE_IQ2_XXS; break;
case LLAMA_FTYPE_MOSTLY_IQ2_XS :quantized_type = GGML_TYPE_IQ2_XS; break;
default: throw std::runtime_error(format("invalid output file type %d\n", ftype)); default: throw std::runtime_error(format("invalid output file type %d\n", ftype));
} }

1
llama.h
View File

@ -104,6 +104,7 @@ extern "C" {
LLAMA_FTYPE_MOSTLY_Q5_K_M = 17, // except 1d tensors LLAMA_FTYPE_MOSTLY_Q5_K_M = 17, // except 1d tensors
LLAMA_FTYPE_MOSTLY_Q6_K = 18, // except 1d tensors LLAMA_FTYPE_MOSTLY_Q6_K = 18, // except 1d tensors
LLAMA_FTYPE_MOSTLY_IQ2_XXS = 19, // except 1d tensors LLAMA_FTYPE_MOSTLY_IQ2_XXS = 19, // except 1d tensors
LLAMA_FTYPE_MOSTLY_IQ2_XS = 20, // except 1d tensors
LLAMA_FTYPE_GUESSED = 1024, // not specified in the model file LLAMA_FTYPE_GUESSED = 1024, // not specified in the model file
}; };

5
tests/test-quantize-fns.cpp
View File

@ -134,8 +134,9 @@ int main(int argc, char * argv[]) {
continue; continue;
} }
if ((ggml_type)i == GGML_TYPE_IQ2_XXS) { const ggml_type ei = (ggml_type)i;
printf("Skip %s due to missing quantization functionality\n", ggml_type_name((ggml_type) i)); if (ei == GGML_TYPE_IQ2_XXS || ei == GGML_TYPE_IQ2_XS) {
printf("Skip %s due to missing quantization functionality\n", ggml_type_name(ei));
continue; continue;
} }