Merge branch 'master' into gg/flash-attn
This commit is contained in:
Georgi Gerganov
commit
3d03bcb7af
68 changed files with 6348 additions and 435 deletions
274
ggml-metal.metal
274
ggml-metal.metal
|
|
@ -2856,6 +2856,12 @@ typedef struct {
|
|||
} block_iq2_xs;
|
||||
// 74 bytes / block for QK_K = 256, so 2.3125 bpw
|
||||
|
||||
typedef struct {
|
||||
half d;
|
||||
uint8_t qs[3*QK_K/8];
|
||||
} block_iq3_xxs;
|
||||
// 98 bytes / block for QK_K = 256, so 3.0625 bpw
|
||||
|
||||
//====================================== dot products =========================
|
||||
|
||||
void kernel_mul_mv_q2_K_f32_impl(
|
||||
|
|
@ -4078,6 +4084,42 @@ constexpr constant static uint64_t iq2xs_grid[512] = {
|
|||
0x2b2b2b2b082b2b08, 0x2b2b2b2b082b2b2b, 0x2b2b2b2b2b190819, 0x2b2b2b2b2b2b2b2b,
|
||||
};
|
||||
|
||||
constexpr constant static uint32_t iq3xxs_grid[256] = {
|
||||
0x04040404, 0x04040414, 0x04040424, 0x04040c0c, 0x04040c1c, 0x04040c3e, 0x04041404, 0x04041414,
|
||||
0x04041c0c, 0x04042414, 0x04043e1c, 0x04043e2c, 0x040c040c, 0x040c041c, 0x040c0c04, 0x040c0c14,
|
||||
0x040c140c, 0x040c142c, 0x040c1c04, 0x040c1c14, 0x040c240c, 0x040c2c24, 0x040c3e04, 0x04140404,
|
||||
0x04140414, 0x04140424, 0x04140c0c, 0x04141404, 0x04141414, 0x04141c0c, 0x04141c1c, 0x04141c3e,
|
||||
0x04142c0c, 0x04142c3e, 0x04143e2c, 0x041c040c, 0x041c043e, 0x041c0c04, 0x041c0c14, 0x041c142c,
|
||||
0x041c3e04, 0x04240c1c, 0x04241c3e, 0x04242424, 0x04242c3e, 0x04243e1c, 0x04243e2c, 0x042c040c,
|
||||
0x042c043e, 0x042c1c14, 0x042c2c14, 0x04341c2c, 0x04343424, 0x043e0c04, 0x043e0c24, 0x043e0c34,
|
||||
0x043e241c, 0x043e340c, 0x0c04040c, 0x0c04041c, 0x0c040c04, 0x0c040c14, 0x0c04140c, 0x0c04141c,
|
||||
0x0c041c04, 0x0c041c14, 0x0c041c24, 0x0c04243e, 0x0c042c04, 0x0c0c0404, 0x0c0c0414, 0x0c0c0c0c,
|
||||
0x0c0c1404, 0x0c0c1414, 0x0c14040c, 0x0c14041c, 0x0c140c04, 0x0c140c14, 0x0c14140c, 0x0c141c04,
|
||||
0x0c143e14, 0x0c1c0404, 0x0c1c0414, 0x0c1c1404, 0x0c1c1c0c, 0x0c1c2434, 0x0c1c3434, 0x0c24040c,
|
||||
0x0c24042c, 0x0c242c04, 0x0c2c1404, 0x0c2c1424, 0x0c2c2434, 0x0c2c3e0c, 0x0c34042c, 0x0c3e1414,
|
||||
0x0c3e2404, 0x14040404, 0x14040414, 0x14040c0c, 0x14040c1c, 0x14041404, 0x14041414, 0x14041434,
|
||||
0x14041c0c, 0x14042414, 0x140c040c, 0x140c041c, 0x140c042c, 0x140c0c04, 0x140c0c14, 0x140c140c,
|
||||
0x140c1c04, 0x140c341c, 0x140c343e, 0x140c3e04, 0x14140404, 0x14140414, 0x14140c0c, 0x14140c3e,
|
||||
0x14141404, 0x14141414, 0x14141c3e, 0x14142404, 0x14142c2c, 0x141c040c, 0x141c0c04, 0x141c0c24,
|
||||
0x141c3e04, 0x141c3e24, 0x14241c2c, 0x14242c1c, 0x142c041c, 0x142c143e, 0x142c240c, 0x142c3e24,
|
||||
0x143e040c, 0x143e041c, 0x143e0c34, 0x143e242c, 0x1c04040c, 0x1c040c04, 0x1c040c14, 0x1c04140c,
|
||||
0x1c04141c, 0x1c042c04, 0x1c04342c, 0x1c043e14, 0x1c0c0404, 0x1c0c0414, 0x1c0c1404, 0x1c0c1c0c,
|
||||
0x1c0c2424, 0x1c0c2434, 0x1c14040c, 0x1c14041c, 0x1c140c04, 0x1c14142c, 0x1c142c14, 0x1c143e14,
|
||||
0x1c1c0c0c, 0x1c1c1c1c, 0x1c241c04, 0x1c24243e, 0x1c243e14, 0x1c2c0404, 0x1c2c0434, 0x1c2c1414,
|
||||
0x1c2c2c2c, 0x1c340c24, 0x1c341c34, 0x1c34341c, 0x1c3e1c1c, 0x1c3e3404, 0x24040424, 0x24040c3e,
|
||||
0x24041c2c, 0x24041c3e, 0x24042c1c, 0x24042c3e, 0x240c3e24, 0x24141404, 0x24141c3e, 0x24142404,
|
||||
0x24143404, 0x24143434, 0x241c043e, 0x241c242c, 0x24240424, 0x24242c0c, 0x24243424, 0x242c142c,
|
||||
0x242c241c, 0x242c3e04, 0x243e042c, 0x243e0c04, 0x243e0c14, 0x243e1c04, 0x2c040c14, 0x2c04240c,
|
||||
0x2c043e04, 0x2c0c0404, 0x2c0c0434, 0x2c0c1434, 0x2c0c2c2c, 0x2c140c24, 0x2c141c14, 0x2c143e14,
|
||||
0x2c1c0414, 0x2c1c2c1c, 0x2c240c04, 0x2c24141c, 0x2c24143e, 0x2c243e14, 0x2c2c0414, 0x2c2c1c0c,
|
||||
0x2c342c04, 0x2c3e1424, 0x2c3e2414, 0x34041424, 0x34042424, 0x34042434, 0x34043424, 0x340c140c,
|
||||
0x340c340c, 0x34140c3e, 0x34143424, 0x341c1c04, 0x341c1c34, 0x34242424, 0x342c042c, 0x342c2c14,
|
||||
0x34341c1c, 0x343e041c, 0x343e140c, 0x3e04041c, 0x3e04042c, 0x3e04043e, 0x3e040c04, 0x3e041c14,
|
||||
0x3e042c14, 0x3e0c1434, 0x3e0c2404, 0x3e140c14, 0x3e14242c, 0x3e142c14, 0x3e1c0404, 0x3e1c0c2c,
|
||||
0x3e1c1c1c, 0x3e1c3404, 0x3e24140c, 0x3e24240c, 0x3e2c0404, 0x3e2c0414, 0x3e2c1424, 0x3e341c04,
|
||||
};
|
||||
|
||||
|
||||
constexpr constant static uint8_t ksigns_iq2xs[128] = {
|
||||
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,
|
||||
|
|
@ -4367,6 +4409,143 @@ kernel void kernel_mul_mv_iq2_xs_f32(
|
|||
kernel_mul_mv_iq2_xs_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg);
|
||||
}
|
||||
|
||||
void kernel_mul_mv_iq3_xxs_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_iq3_xxs * x = (device const block_iq3_xxs *) 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 uint32_t * values = (threadgroup uint32_t *)shared_values;
|
||||
threadgroup uint8_t * shared_signs = (threadgroup uint8_t *)(values + 256);
|
||||
{
|
||||
int nval = 4;
|
||||
int pos = (32*sgitg + tiisg)*nval;
|
||||
for (int i = 0; i < nval; ++i) values[pos + i] = iq3xxs_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_iq3_xxs * xr = x + ibl;
|
||||
device const uint8_t * q3 = xr->qs + 8 * ib;
|
||||
device const uint16_t * gas = (device const uint16_t *)(xr->qs + QK_K/4) + 2 * ib;
|
||||
device const half * dh = &xr->d;
|
||||
|
||||
for (int row = 0; row < N_DST; row++) {
|
||||
|
||||
const float db = dh[0];
|
||||
const uint32_t aux32 = gas[0] | (gas[1] << 16);
|
||||
const float d = db * (0.5f + (aux32 >> 28));
|
||||
|
||||
float2 sum = {0};
|
||||
for (int l = 0; l < 4; ++l) {
|
||||
const threadgroup uint8_t * grid1 = (const threadgroup uint8_t *)(values + q3[2*l+0]);
|
||||
const threadgroup uint8_t * grid2 = (const threadgroup uint8_t *)(values + q3[2*l+1]);
|
||||
const uint8_t signs = shared_signs[(aux32 >> 7*l) & 127];
|
||||
for (int j = 0; j < 4; ++j) {
|
||||
sum[0] += yl[8*l + j + 0] * grid1[j] * (signs & kmask_iq2xs[j+0] ? -1.f : 1.f);
|
||||
sum[1] += yl[8*l + j + 4] * grid2[j] * (signs & kmask_iq2xs[j+4] ? -1.f : 1.f);
|
||||
}
|
||||
}
|
||||
sumf[row] += d * (sum[0] + sum[1]);
|
||||
|
||||
dh += nb*sizeof(block_iq3_xxs)/2;
|
||||
q3 += nb*sizeof(block_iq3_xxs);
|
||||
gas += nb*sizeof(block_iq3_xxs)/2;
|
||||
}
|
||||
|
||||
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 * 0.5f;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
[[host_name("kernel_mul_mv_iq3_xxs_f32")]]
|
||||
kernel void kernel_mul_mv_iq3_xxs_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_iq3_xxs_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg);
|
||||
}
|
||||
|
||||
|
||||
//============================= templates and their specializations =============================
|
||||
|
||||
// NOTE: this is not dequantizing - we are simply fitting the template
|
||||
|
|
@ -4684,6 +4863,33 @@ void dequantize_iq2_xs(device const block_iq2_xs * xb, short il, thread type4x4
|
|||
}
|
||||
}
|
||||
|
||||
template <typename type4x4>
|
||||
void dequantize_iq3_xxs(device const block_iq3_xxs * 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 uint8_t * q3 = xb->qs + 8*ib32;
|
||||
device const uint16_t * gas = (device const uint16_t *)(xb->qs + QK_K/4) + 2*ib32;
|
||||
const uint32_t aux32 = gas[0] | (gas[1] << 16);
|
||||
const float dl = d * (0.5f + (aux32 >> 28)) * 0.5f;
|
||||
constant uint8_t * grid1 = (constant uint8_t *)(iq3xxs_grid + q3[4*il+0]);
|
||||
constant uint8_t * grid2 = (constant uint8_t *)(iq3xxs_grid + q3[4*il+1]);
|
||||
uint8_t signs = ksigns_iq2xs[(aux32 >> 14*il) & 127];
|
||||
for (int i = 0; i < 4; ++i) {
|
||||
reg[0][i] = dl * grid1[i] * (signs & kmask_iq2xs[i+0] ? -1.f : 1.f);
|
||||
reg[1][i] = dl * grid2[i] * (signs & kmask_iq2xs[i+4] ? -1.f : 1.f);
|
||||
}
|
||||
grid1 = (constant uint8_t *)(iq3xxs_grid + q3[4*il+2]);
|
||||
grid2 = (constant uint8_t *)(iq3xxs_grid + q3[4*il+3]);
|
||||
signs = ksigns_iq2xs[(aux32 >> (14*il+7)) & 127];
|
||||
for (int i = 0; i < 4; ++i) {
|
||||
reg[2][i] = dl * grid1[i] * (signs & kmask_iq2xs[i+0] ? -1.f : 1.f);
|
||||
reg[3][i] = dl * grid2[i] * (signs & kmask_iq2xs[i+4] ? -1.f : 1.f);
|
||||
}
|
||||
}
|
||||
|
||||
template<typename block_q, short nl, void (*dequantize_func)(device const block_q *, short, thread float4x4 &)>
|
||||
kernel void kernel_get_rows(
|
||||
device const void * src0,
|
||||
|
|
@ -5225,6 +5431,7 @@ template [[host_name("kernel_get_rows_q5_K")]] kernel get_rows_t kernel_get_rows
|
|||
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_xs")]] kernel get_rows_t kernel_get_rows<block_iq2_xs, QK_NL, dequantize_iq2_xs>;
|
||||
template [[host_name("kernel_get_rows_iq3_xxs")]] kernel get_rows_t kernel_get_rows<block_iq3_xxs, QK_NL, dequantize_iq3_xxs>;
|
||||
|
||||
//
|
||||
// matrix-matrix multiplication
|
||||
|
|
@ -5263,6 +5470,7 @@ template [[host_name("kernel_mul_mm_q5_K_f32")]] kernel mat_mm_t kernel_mul_mm<b
|
|||
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_xs_f32")]] kernel mat_mm_t kernel_mul_mm<block_iq2_xs, QK_NL, dequantize_iq2_xs>;
|
||||
template [[host_name("kernel_mul_mm_iq3_xxs_f32")]] kernel mat_mm_t kernel_mul_mm<block_iq3_xxs, QK_NL, dequantize_iq3_xxs>;
|
||||
|
||||
//
|
||||
// indirect matrix-matrix multiplication
|
||||
|
|
@ -5313,6 +5521,7 @@ template [[host_name("kernel_mul_mm_id_q5_K_f32")]] kernel mat_mm_id_t kernel_mu
|
|||
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_xs_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_iq2_xs, QK_NL, dequantize_iq2_xs>;
|
||||
template [[host_name("kernel_mul_mm_id_iq3_xxs_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_iq3_xxs, QK_NL, dequantize_iq3_xxs>;
|
||||
|
||||
//
|
||||
// matrix-vector multiplication
|
||||
|
|
@ -6215,3 +6424,68 @@ kernel void kernel_mul_mv_id_iq2_xs_f32(
|
|||
tiisg,
|
||||
sgitg);
|
||||
}
|
||||
|
||||
[[host_name("kernel_mul_mv_id_iq3_xxs_f32")]]
|
||||
kernel void kernel_mul_mv_id_iq3_xxs_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_iq3_xxs_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);
|
||||
}
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue