ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-02-24 23:24:13 +00:00
Code Issues Packages Projects Releases Wiki Activity

Adding iq1_s to iqk_mul_mat (Zen4)

Browse Source
This commit is contained in:
Iwan Kawrakow
2025-02-19 20:04:53 +02:00
parent a0ebfdd661
commit 8bf80a09e2
1 changed files with 73 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

73
ggml/src/iqk/iqk_mul_mat.cpp
View File

@@ -3666,6 +3666,65 @@ static void mul_mat_iq1_s_r4_q8_1(int n, const void * vx, size_t bx, const DataI
}
}
// sum[ qy_i * ls_k * (qx_i - 1+/-delta_k)]
// = sum[qy_i * qx_i * ls_k] - 1/8*sum[qy_i * ls_k * (8+/-o_k)]
// = 1/8 * ( sum[qy_i * qx_i * 8*ls+k] - sum[qy_i * ls_k * (8+/-o_k)] )
template <int nrc_y>
static void mul_mat_iq1_s_q8_K(int n, const void * vx, size_t bx, const DataInfo& info, int nrc_x) {
GGML_ASSERT(n%QK_K == 0);
Q8<nrc_y, block_q8_K> q8(info);
__m256i qx[8];
__m256 acc[nrc_y] = {};
auto delta_mask = _mm_set1_epi16(-32768); // to avoid stupid overflow warnings when using 0x8000
__m256i shuffle0 = _mm256_set_epi64x(0x0302030203020302, 0x0100010001000100, 0x0302030203020302, 0x0100010001000100);
for (int ix = 0; ix < nrc_x; ++ix) {
auto iq1s = (const block_iq1_s *)((const char *)vx + ix*bx);
for (int ibl = 0; ibl < n/QK_K; ++ibl) {
float d = GGML_FP16_TO_FP32(iq1s[ibl].d);
auto qhb = _mm_loadu_si128((const __m128i *)iq1s[ibl].qh);
auto scales128 = _mm_and_si128(_mm_srli_epi16(qhb, 12), _mm_set1_epi16(7));
scales128 = _mm_add_epi16(_mm_slli_epi16(scales128, 1), _mm_set1_epi16(1));
auto mask = _mm_cmpeq_epi16_mask(_mm_and_si128(qhb, delta_mask), delta_mask);
auto deltas128 = _mm_mask_blend_epi16(mask, _mm_set1_epi16(-7), _mm_set1_epi16(-9));
deltas128 = _mm_mullo_epi16(scales128, deltas128);
scales128 = _mm_slli_epi16(scales128, 3);
auto deltas_l = _mm_unpacklo_epi16(deltas128, deltas128);
auto deltas_h = _mm_unpackhi_epi16(deltas128, deltas128);
auto deltas = MM256_SET_M128I(deltas_h, deltas_l); // blocks 0,0, 1,1, 2,2, ..., 7,7
auto scales = MM256_SET_M128I(scales128, scales128);
const uint8_t * qs = iq1s[ibl].qs;
const uint16_t * qh = iq1s[ibl].qh;
for (int ib = 0; ib < QK_K/32; ib += 2) {
qx[ib+0] = _mm256_set_epi64x(iq1s_grid_us[qs[3] | ((qh[ib+0] >> 1) & 0x700)], iq1s_grid_us[qs[2] | ((qh[ib+0] << 2) & 0x700)],
iq1s_grid_us[qs[1] | ((qh[ib+0] << 5) & 0x700)], iq1s_grid_us[qs[0] | ((qh[ib+0] << 8) & 0x700)]);
qx[ib+1] = _mm256_set_epi64x(iq1s_grid_us[qs[7] | ((qh[ib+1] >> 1) & 0x700)], iq1s_grid_us[qs[6] | ((qh[ib+1] << 2) & 0x700)],
iq1s_grid_us[qs[5] | ((qh[ib+1] << 5) & 0x700)], iq1s_grid_us[qs[4] | ((qh[ib+1] << 8) & 0x700)]);
qs += 8;
}
for (int iy = 0; iy < nrc_y; ++iy) {
auto bsums = q8.load_bsums(iy, ibl);
auto shuffle = shuffle0;
auto sumi = _mm256_setzero_si256();
for (int ib = 0; ib < QK_K/32; ib += 2) {
auto qy1 = q8.load_quants(iy, ibl, ib+0);
auto qy2 = q8.load_quants(iy, ibl, ib+1);
auto dot1 = _mm256_dpbusd_epi32(_mm256_setzero_si256(), qx[ib+0], qy1);
auto dot2 = _mm256_dpbusd_epi32(_mm256_setzero_si256(), qx[ib+1], qy2);
sumi = _mm256_dpwssd_epi32(sumi, _mm256_shuffle_epi8(scales, shuffle), _mm256_packs_epi32(dot1, dot2));
shuffle = _mm256_add_epi8(shuffle, _mm256_set1_epi8(4));
}
sumi = _mm256_dpwssd_epi32(sumi, bsums, deltas);
acc[iy] = _mm256_fmadd_ps(_mm256_set1_ps(d*q8.scale(iy, ibl)), _mm256_cvtepi32_ps(sumi), acc[iy]);
}
}
for (int iy = 0; iy < nrc_y; ++iy) {
info.store(ix, iy, 0.125f*hsum_float_8(acc[iy]));
acc[iy] = _mm256_setzero_ps();
}
}
}
template <int nrc_y>
static void mul_mat_iq1_m_r4_q8_0(int n, const void * vx, size_t bx, const DataInfo& info, int nrc_x) {
GGML_ASSERT(nrc_x%4 == 0);
@@ -9473,6 +9532,20 @@ bool MulMat::prepare(int typeA, int typeB, int ne00, MulMat& mm, int Ny) {
mm.funcs[7] = mul_mat_q8_0_r8_q8_1<8>;
expected_typeB = GGML_TYPE_Q8_1_X4;
break;
case GGML_TYPE_IQ1_S:
mm.funcs[0] = mul_mat_iq1_s_q8_K<1>;
mm.funcs[1] = mul_mat_iq1_s_q8_K<2>;
mm.funcs[2] = mul_mat_iq1_s_q8_K<3>;
mm.funcs[3] = mul_mat_iq1_s_q8_K<4>;
mm.funcs[4] = mul_mat_iq1_s_q8_K<5>;
mm.funcs[5] = mul_mat_iq1_s_q8_K<6>;
mm.funcs[6] = mul_mat_iq1_s_q8_K<7>;
mm.funcs[7] = mul_mat_iq1_s_q8_K<8>;
#ifdef HAVE_FANCY_SIMD
mm.func16 = mul_mat_iq1_s_q8_K<16>;
#endif
expected_typeB = GGML_TYPE_Q8_K;
break;
case GGML_TYPE_IQ1_S_R4:
assert (ne00 % QK4_NL == 0);
mm.funcs[0] = mul_mat_iq1_s_r4_q8_1<1>;