diff options
| -rw-r--r-- | examples/quantize/quantize.cpp | 1 | ||||
| -rw-r--r-- | ggml/include/ggml.h | 2 | ||||
| -rw-r--r-- | ggml/src/ggml-common.h | 6 | ||||
| -rw-r--r-- | ggml/src/ggml-quants.c | 1 | ||||
| -rw-r--r-- | ggml/src/ggml.c | 26 | ||||
| -rw-r--r-- | ggml/src/iqk/iqk_mul_mat.cpp | 201 | ||||
| -rw-r--r-- | ggml/src/iqk/iqk_quantize.cpp | 101 | ||||
| -rw-r--r-- | ggml/src/iqk/iqk_quantize.h | 6 | ||||
| -rw-r--r-- | include/llama.h | 1 | ||||
| -rw-r--r-- | src/llama.cpp | 21 |
10 files changed, 364 insertions, 2 deletions
diff --git a/examples/quantize/quantize.cpp b/examples/quantize/quantize.cpp index d505f493..cefe2735 100644 --- a/examples/quantize/quantize.cpp +++ b/examples/quantize/quantize.cpp @@ -51,6 +51,7 @@ static const std::vector<struct quant_option> QUANT_OPTIONS = { { "Q8_0_R4", LLAMA_FTYPE_MOSTLY_Q8_0_R4, " 8.50 bpw quantization", }, { "IQ4_XS", LLAMA_FTYPE_MOSTLY_IQ4_XS, " 4.25 bpw non-linear quantization", }, { "IQ4_KS", LLAMA_FTYPE_MOSTLY_IQ4_KS, " 4.25 bpw non-linear quantization", }, + { "IQ4_KS_R4",LLAMA_FTYPE_MOSTLY_IQ4_KS_R4,"IQ4_KS repacked", }, { "IQ4_KSS", LLAMA_FTYPE_MOSTLY_IQ4_KSS, " 4.0 bpw non-linear quantization", }, { "IQ2_K", LLAMA_FTYPE_MOSTLY_IQ2_K, " 2.375 bpw non-linear quantization",}, { "IQ2_K_R4", LLAMA_FTYPE_MOSTLY_IQ2_K_R4, "IQ2_K repacked",}, diff --git a/ggml/include/ggml.h b/ggml/include/ggml.h index 07142692..1a4e7f19 100644 --- a/ggml/include/ggml.h +++ b/ggml/include/ggml.h @@ -427,6 +427,7 @@ extern "C" { GGML_TYPE_IQ3_K_R4 = 338, GGML_TYPE_IQ4_K_R4 = 339, GGML_TYPE_IQ5_K_R4 = 340, + GGML_TYPE_IQ4_KS_R4 = 344, GGML_TYPE_Q8_K_R8 = 399, GGML_TYPE_COUNT, }; @@ -504,6 +505,7 @@ extern "C" { GGML_FTYPE_MOSTLY_IQ3_K_R4 = 331, // except 1d tensors GGML_FTYPE_MOSTLY_IQ4_K_R4 = 332, // except 1d tensors GGML_FTYPE_MOSTLY_IQ5_K_R4 = 333, // except 1d tensors + GGML_FTYPE_MOSTLY_IQ4_KS_R4 = 337, // except 1d tensors GGML_FTYPE_MOSTLY_Q8_K_R8 = 399, // except 1d tensors }; diff --git a/ggml/src/ggml-common.h b/ggml/src/ggml-common.h index 0af461c7..6e87dbaa 100644 --- a/ggml/src/ggml-common.h +++ b/ggml/src/ggml-common.h @@ -509,6 +509,12 @@ typedef struct { static_assert(sizeof(block_iq4_ks) == QK_K/32 + QK_K/2, "wrong iq4_ks block size/padding"); typedef struct { + uint8_t scales[QK_K/8]; + uint8_t qs[QK_K*2]; +} block_iq4_ks_r4; +static_assert(sizeof(block_iq4_ks_r4) == 4*sizeof(block_iq4_ks), "wrong iq4_ks_r4 block size/padding"); + +typedef struct { uint32_t qs[QK_K/8]; } block_iq4_kss; static_assert(sizeof(block_iq4_kss) == QK_K/8*sizeof(uint32_t), "wrong iq4_kss block size/padding"); diff --git a/ggml/src/ggml-quants.c b/ggml/src/ggml-quants.c index d0de3d0f..2b59808b 100644 --- a/ggml/src/ggml-quants.c +++ b/ggml/src/ggml-quants.c @@ -15211,6 +15211,7 @@ bool ggml_validate_row_data(enum ggml_type type, const void * data, size_t nbyte case GGML_TYPE_IQ3_K_R4: break; case GGML_TYPE_IQ4_K_R4: break; case GGML_TYPE_IQ5_K_R4: break; + case GGML_TYPE_IQ4_KS_R4: break; case GGML_TYPE_Q8_K_R8: break; case GGML_TYPE_BF16_R16: break; case GGML_TYPE_Q4_0_4_4: diff --git a/ggml/src/ggml.c b/ggml/src/ggml.c index 526b1139..758ee018 100644 --- a/ggml/src/ggml.c +++ b/ggml/src/ggml.c @@ -1165,6 +1165,23 @@ static const ggml_type_traits_t type_traits[GGML_TYPE_COUNT] = { .nrows = 1, .row_meta_size = 4, }, + [GGML_TYPE_IQ4_KS_R4] = { + .type_name = "iq4_ks_r4", + .blck_size = QK_K, + .type_size = sizeof(block_iq4_ks), + .is_quantized = true, + .to_float = (ggml_to_float_t) dequantize_row_iq4_ks_r4, + .from_float = quantize_row_iq4_ks_r4, + .from_float_ref = (ggml_from_float_t)quantize_row_iq4_ks_r4_ref, + .vec_dot = vec_dot_iq4_ks_r4_q8_k, +#if defined __AVX2__ + .vec_dot_type = GGML_TYPE_Q8_K32, +#else + .vec_dot_type = GGML_TYPE_Q8_K, +#endif + .nrows = 1, + .row_meta_size = 4, + }, [GGML_TYPE_IQ4_KSS] = { .type_name = "iq4_kss", .blck_size = QK_K, @@ -4197,6 +4214,7 @@ enum ggml_type ggml_ftype_to_ggml_type(enum ggml_ftype ftype) { case GGML_FTYPE_MOSTLY_Q8_0_R4: wtype = GGML_TYPE_Q8_0_R4; break; case GGML_FTYPE_MOSTLY_IQ4_XS: wtype = GGML_TYPE_IQ4_XS; break; case GGML_FTYPE_MOSTLY_IQ4_KS: wtype = GGML_TYPE_IQ4_KS; break; + case GGML_FTYPE_MOSTLY_IQ4_KS_R4: wtype = GGML_TYPE_IQ4_KS_R4;break; case GGML_FTYPE_MOSTLY_IQ4_KSS: wtype = GGML_TYPE_IQ4_KSS; break; case GGML_FTYPE_MOSTLY_IQ2_K: wtype = GGML_TYPE_IQ2_K; break; case GGML_FTYPE_MOSTLY_IQ2_K_R4: wtype = GGML_TYPE_IQ2_K_R4; break; @@ -10737,6 +10755,7 @@ static void ggml_compute_forward_add( case GGML_TYPE_Q8_0_R4: case GGML_TYPE_IQ4_XS: case GGML_TYPE_IQ4_KS: + case GGML_TYPE_IQ4_KS_R4: case GGML_TYPE_IQ4_KSS: case GGML_TYPE_IQ2_K: case GGML_TYPE_IQ2_K_R4: @@ -11196,6 +11215,7 @@ static void ggml_compute_forward_add1( case GGML_TYPE_Q8_0_R4: case GGML_TYPE_IQ4_XS: case GGML_TYPE_IQ4_KS: + case GGML_TYPE_IQ4_KS_R4: case GGML_TYPE_IQ4_KSS: case GGML_TYPE_IQ2_K: case GGML_TYPE_IQ2_K_R4: @@ -11352,6 +11372,7 @@ static void ggml_compute_forward_acc( case GGML_TYPE_Q8_0_R4: case GGML_TYPE_IQ4_XS: case GGML_TYPE_IQ4_KS: + case GGML_TYPE_IQ4_KS_R4: case GGML_TYPE_IQ4_KSS: case GGML_TYPE_IQ2_K: case GGML_TYPE_IQ2_K_R4: @@ -14554,6 +14575,7 @@ static void ggml_compute_forward_out_prod( case GGML_TYPE_Q8_0_R4: case GGML_TYPE_IQ4_XS: case GGML_TYPE_IQ4_KS: + case GGML_TYPE_IQ4_KS_R4: case GGML_TYPE_IQ4_KSS: case GGML_TYPE_IQ2_K: case GGML_TYPE_IQ2_K_R4: @@ -14950,6 +14972,7 @@ static void ggml_compute_forward_set( case GGML_TYPE_Q8_0_R4: case GGML_TYPE_IQ4_XS: case GGML_TYPE_IQ4_KS: + case GGML_TYPE_IQ4_KS_R4: case GGML_TYPE_IQ4_KSS: case GGML_TYPE_IQ2_K: case GGML_TYPE_IQ2_K_R4: @@ -15240,6 +15263,7 @@ static void ggml_compute_forward_get_rows( case GGML_TYPE_Q8_0_R4: case GGML_TYPE_IQ4_XS: case GGML_TYPE_IQ4_KS: + case GGML_TYPE_IQ4_KS_R4: case GGML_TYPE_IQ4_KSS: case GGML_TYPE_IQ2_K: case GGML_TYPE_IQ2_K_R4: @@ -15859,6 +15883,7 @@ static void ggml_compute_forward_clamp( case GGML_TYPE_Q8_0_R4: case GGML_TYPE_IQ4_XS: case GGML_TYPE_IQ4_KS: + case GGML_TYPE_IQ4_KS_R4: case GGML_TYPE_IQ4_KSS: case GGML_TYPE_IQ2_K: case GGML_TYPE_IQ2_K_R4: @@ -22706,6 +22731,7 @@ size_t ggml_quantize_chunk( case GGML_TYPE_Q8_0_R4: result = quantize_q8_0_r4(src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break; case GGML_TYPE_IQ4_XS: result = quantize_iq4_xs (src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break; case GGML_TYPE_IQ4_KS: result = quantize_iq4_ks (src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break; + case GGML_TYPE_IQ4_KS_R4:result = quantize_iq4_ks_r4(src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break; case GGML_TYPE_IQ4_KSS: result = quantize_iq4_kss(src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break; case GGML_TYPE_IQ2_K: result = quantize_iq2_k (src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break; case GGML_TYPE_IQ2_K_R4:result = quantize_iq2_k_r4(src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break; diff --git a/ggml/src/iqk/iqk_mul_mat.cpp b/ggml/src/iqk/iqk_mul_mat.cpp index 3e8cbf06..f4576319 100644 --- a/ggml/src/iqk/iqk_mul_mat.cpp +++ b/ggml/src/iqk/iqk_mul_mat.cpp @@ -186,6 +186,7 @@ struct MulMat { case GGML_TYPE_IQ3_K_R4: case GGML_TYPE_IQ4_K_R4: case GGML_TYPE_IQ5_K_R4: + case GGML_TYPE_IQ4_KS_R4: case GGML_TYPE_IQ2_BN_R4: return 4; case GGML_TYPE_Q8_K_R8: return 8; case GGML_TYPE_BF16_R16: return 16; @@ -3103,6 +3104,115 @@ static void mul_mat_iq4_xs_r4_q8_k(int n, const void * vx, size_t bx, const Data #endif template <int nrc_y> +static void mul_mat_iq4_ks_r4_q8_k(int n, const void * vx, size_t bx, const DataInfo& info, int nrc_x) { + GGML_ASSERT(nrc_x%4 == 0); + Q8<nrc_y, block_q8_K> q8(info); + auto m4 = _mm256_set1_epi8(0xf); +#ifndef HAVE_FANCY_SIMD + auto s_shuffle = _mm256_set_epi64x(0x0f0e0f0e0d0c0d0c, 0x0b0a0b0a09080908, 0x0706070605040504, 0x0302030201000100); + auto values128 = _mm_loadu_si128((const __m128i *)iq4k_values); + auto values = MM256_SET_M128I(values128, values128); +#else + auto values = load_iq4nl_values_256(); +#endif + int nbl = n / QK_K; + using helper_t = union { __m256i vec; uint32_t val[8]; }; +#ifndef HAVE_FANCY_SIMD + helper_t h, h_shift; +#else + using helper512_t = union { __m512i vec; uint64_t val[8]; }; + helper_t h; + helper512_t h_shift; +#endif + __m256 acc[nrc_y] = {}; + __m256i isum[nrc_y] = {}; + __m256i qx[4]; + for (int ix = 0; ix < nrc_x; ix += 4) { + auto dptr = (const float *)((const char *)vx + (ix+0)*bx); + const block_iq4_ks_r4 * iq4 = (const block_iq4_ks_r4 *)(dptr + 4); + auto d4 = _mm_loadu_ps(dptr); + for (int ibl = 0; ibl < nbl; ++ibl) { // Block of 256 + auto scales = _mm256_loadu_si256((const __m256i *)iq4[ibl].scales); + h.vec = _mm256_sub_epi8(_mm256_and_si256(scales, _mm256_set1_epi8(-2)), _mm256_set1_epi8(127)); +#ifndef HAVE_FANCY_SIMD + h_shift.vec = _mm256_slli_epi16(_mm256_and_si256(scales, _mm256_set1_epi8(1)), 2); + { + __m256 v1 = _mm256_mul_ps(_mm256_cvtepi32_ps(MM256_SET_M128I(_mm_cvtepi8_epi32(_mm_set1_epi32(h.val[4])), _mm_cvtepi8_epi32(_mm_set1_epi32(h.val[0])))), + _mm256_cvtepi32_ps(MM256_SET_M128I(_mm_cvtepi8_epi32(_mm_set1_epi32(h_shift.val[4])), _mm_cvtepi8_epi32(_mm_set1_epi32(h_shift.val[0]))))); + __m256 v2 = _mm256_mul_ps(_mm256_cvtepi32_ps(MM256_SET_M128I(_mm_cvtepi8_epi32(_mm_set1_epi32(h.val[5])), _mm_cvtepi8_epi32(_mm_set1_epi32(h.val[1])))), + _mm256_cvtepi32_ps(MM256_SET_M128I(_mm_cvtepi8_epi32(_mm_set1_epi32(h_shift.val[5])), _mm_cvtepi8_epi32(_mm_set1_epi32(h_shift.val[1]))))); + __m256 v3 = _mm256_mul_ps(_mm256_cvtepi32_ps(MM256_SET_M128I(_mm_cvtepi8_epi32(_mm_set1_epi32(h.val[6])), _mm_cvtepi8_epi32(_mm_set1_epi32(h.val[2])))), + _mm256_cvtepi32_ps(MM256_SET_M128I(_mm_cvtepi8_epi32(_mm_set1_epi32(h_shift.val[6])), _mm_cvtepi8_epi32(_mm_set1_epi32(h_shift.val[2]))))); + __m256 v4 = _mm256_mul_ps(_mm256_cvtepi32_ps(MM256_SET_M128I(_mm_cvtepi8_epi32(_mm_set1_epi32(h.val[7])), _mm_cvtepi8_epi32(_mm_set1_epi32(h.val[3])))), + _mm256_cvtepi32_ps(MM256_SET_M128I(_mm_cvtepi8_epi32(_mm_set1_epi32(h_shift.val[7])), _mm_cvtepi8_epi32(_mm_set1_epi32(h_shift.val[3]))))); + for (int iy = 0; iy < nrc_y; ++iy) { + auto m8 = _mm256_loadu_ps((const float *)q8.y[iy][ibl].bsums); + acc[iy] = _mm256_fmadd_ps(v1, _mm256_shuffle_ps(m8, m8, 0x00), acc[iy]); + acc[iy] = _mm256_fmadd_ps(v2, _mm256_shuffle_ps(m8, m8, 0x55), acc[iy]); + acc[iy] = _mm256_fmadd_ps(v3, _mm256_shuffle_ps(m8, m8, 0xaa), acc[iy]); + acc[iy] = _mm256_fmadd_ps(v4, _mm256_shuffle_ps(m8, m8, 0xff), acc[iy]); + } + } +#else + auto shift = _mm256_add_epi8(_mm256_set1_epi8(-64), _mm256_slli_epi16(_mm256_and_si256(scales, _mm256_set1_epi8(1)), 1)); + h_shift.vec = _mm512_mullo_epi16(_mm512_cvtepi8_epi16(shift), _mm512_cvtepi8_epi16(h.vec)); +#endif + for (int ib = 0; ib < QK_K/32; ++ib) { +#ifdef HAVE_FANCY_SIMD + auto iscales = _mm256_cvtepi8_epi32(_mm_set1_epi32(h.val[ib])); + auto ishifts = _mm256_cvtepi16_epi32(_mm_set1_epi64x(h_shift.val[ib])); + auto scales_m = _mm256_cvtepi32_ps(ishifts); + for (int iy = 0; iy < nrc_y; ++iy) { + float m8 = ((const float *)q8.y[iy][ibl].bsums)[ib]; + acc[iy] = _mm256_fmadd_ps(scales_m, _mm256_set1_ps(m8), acc[iy]); + } +#endif + auto bits1 = _mm256_loadu_si256((const __m256i *)iq4[ibl].qs+2*ib+0); + auto bits2 = _mm256_loadu_si256((const __m256i *)iq4[ibl].qs+2*ib+1); + qx[0] = _mm256_shuffle_epi8(values, _mm256_and_si256(bits1, m4)); + qx[1] = _mm256_shuffle_epi8(values, _mm256_and_si256(bits2, m4)); + qx[2] = _mm256_shuffle_epi8(values, _mm256_and_si256(_mm256_srli_epi16(bits1, 4), m4)); + qx[3] = _mm256_shuffle_epi8(values, _mm256_and_si256(_mm256_srli_epi16(bits2, 4), m4)); +#ifndef HAVE_FANCY_SIMD + auto iscales = _mm256_shuffle_epi8(_mm256_cvtepi8_epi16(_mm_set1_epi32(h.val[ib])), s_shuffle); + auto s1 = _mm256_sign_epi8(qx[0], qx[0]); + auto s2 = _mm256_sign_epi8(qx[1], qx[1]); + auto s3 = _mm256_sign_epi8(qx[2], qx[2]); + auto s4 = _mm256_sign_epi8(qx[3], qx[3]); +#endif + for (int iy = 0; iy < nrc_y; ++iy) { + auto y = _mm256_loadu_si256((const __m256i*)q8.y[iy][ibl].qs+ib); +#ifdef HAVE_FANCY_SIMD + auto sumi = _mm256_setzero_si256(); + sumi = _mm256_dpbusd_epi32(sumi, qx[0], _mm256_shuffle_epi32(y, 0x00)); + sumi = _mm256_dpbusd_epi32(sumi, qx[1], _mm256_shuffle_epi32(y, 0x55)); + sumi = _mm256_dpbusd_epi32(sumi, qx[2], _mm256_shuffle_epi32(y, 0xaa)); + sumi = _mm256_dpbusd_epi32(sumi, qx[3], _mm256_shuffle_epi32(y, 0xff)); + isum[iy] = _mm256_add_epi32(isum[iy], _mm256_mullo_epi32(iscales, sumi)); +#else + auto sumi1 = _mm256_maddubs_epi16(s1, _mm256_sign_epi8(_mm256_shuffle_epi32(y, 0x00), qx[0])); + auto sumi2 = _mm256_maddubs_epi16(s2, _mm256_sign_epi8(_mm256_shuffle_epi32(y, 0x55), qx[1])); + auto sumi3 = _mm256_maddubs_epi16(s3, _mm256_sign_epi8(_mm256_shuffle_epi32(y, 0xaa), qx[2])); + auto sumi4 = _mm256_maddubs_epi16(s4, _mm256_sign_epi8(_mm256_shuffle_epi32(y, 0xff), qx[3])); + isum[iy] = _mm256_add_epi32(isum[iy], _mm256_add_epi32(_mm256_madd_epi16(iscales, sumi1), _mm256_madd_epi16(iscales, sumi2))); + isum[iy] = _mm256_add_epi32(isum[iy], _mm256_add_epi32(_mm256_madd_epi16(iscales, sumi3), _mm256_madd_epi16(iscales, sumi4))); +#endif + } + } + for (int iy = 0; iy < nrc_y; ++iy) { + acc[iy] = _mm256_fmadd_ps(_mm256_set1_ps(q8.scale(iy, ibl)), _mm256_cvtepi32_ps(isum[iy]), acc[iy]); + isum[iy] = _mm256_setzero_si256(); + } + } + for (int iy = 0; iy < nrc_y; ++iy) { + auto sum = _mm_add_ps(_mm256_castps256_ps128(acc[iy]), _mm256_extractf128_ps(acc[iy], 1)); + acc[iy] = _mm256_setzero_ps(); + info.store(ix+0, iy, _mm_mul_ps(d4, sum)); + } + } +} + +template <int nrc_y> static void mul_mat_q4_k_r4_q8_k_avx2(int n, const void * vx, size_t bx, const DataInfo& info, int nrc_x) { GGML_ASSERT(nrc_x%4 == 0); Q8<nrc_y, block_q8_K> q8(info); @@ -6482,6 +6592,18 @@ bool MulMat::prepare(int typeA, int typeB, int ne00, MulMat& mm, int Ny) { mm.funcs[7] = mul_mat_iq4_xs_r4_q8_k<8>; expected_typeB = GGML_TYPE_Q8_K32; break; + case GGML_TYPE_IQ4_KS_R4: + assert (ne00 % QK_K == 0); + mm.funcs[0] = mul_mat_iq4_ks_r4_q8_k<1>; + mm.funcs[1] = mul_mat_iq4_ks_r4_q8_k<2>; + mm.funcs[2] = mul_mat_iq4_ks_r4_q8_k<3>; + mm.funcs[3] = mul_mat_iq4_ks_r4_q8_k<4>; + mm.funcs[4] = mul_mat_iq4_ks_r4_q8_k<5>; + mm.funcs[5] = mul_mat_iq4_ks_r4_q8_k<6>; + mm.funcs[6] = mul_mat_iq4_ks_r4_q8_k<7>; + mm.funcs[7] = mul_mat_iq4_ks_r4_q8_k<8>; + expected_typeB = GGML_TYPE_Q8_K32; + break; case GGML_TYPE_Q2_K_R4: assert (ne00 % QK_K == 0); mm.funcs[0] = mul_mat_q2_k_r4_q8_k<1>; @@ -9278,6 +9400,81 @@ void mul_mat_iq4_xs_r4_q8_k(int n, const void * vx, size_t bx, const DataInfo& i } } +template <int nrc_y> +void mul_mat_iq4_ks_r4_q8_k(int n, const void * vx, size_t bx, const DataInfo& info, int nrc_x) { + GGML_ASSERT(nrc_x%4 == 0); + Q8<nrc_y, block_q8_K> q8(info); + auto m4 = vdupq_n_u8(0xf); + auto values = vld1q_s8(iq4k_values); + int nbl = n / QK_K; + int8x16_t qx[8]; + int16x8x4_t iscales; + int32x4x4_t scales; + float32x4_t acc[nrc_y] = {}; + int32x4_t isum[nrc_y] = {}; + for (int ix = 0; ix < nrc_x; ix += 4) { + auto dptr = (const float *)((const char *)vx + ix*bx); + auto d4 = vld1q_f32(dptr); + const block_iq4_ks_r4 * iq4 = (const block_iq4_ks_r4 *)(dptr + 4); + for (int ibl = 0; ibl < nbl; ++ibl) { + auto sas = vld1q_u8_x2(iq4[ibl].scales); + auto scale = vandq_u8(sas.val[0], vdupq_n_u8(254)); + iscales.val[0] = vaddq_s16(vreinterpretq_s16_u16(vmovl_u8(vget_low_u8 (scale))), vdupq_n_s16(-127)); + iscales.val[1] = vaddq_s16(vreinterpretq_s16_u16(vmovl_u8(vget_high_u8(scale))), vdupq_n_s16(-127)); + scale = vandq_u8(sas.val[1], vdupq_n_u8(254)); + iscales.val[2] = vaddq_s16(vreinterpretq_s16_u16(vmovl_u8(vget_low_u8 (scale))), vdupq_n_s16(-127)); + iscales.val[3] = vaddq_s16(vreinterpretq_s16_u16(vmovl_u8(vget_high_u8(scale))), vdupq_n_s16(-127)); + // Adding the block shifts costs us ~9% in performance drop. + // Is there a better way? + sas.val[0] = vshlq_n_u8(vandq_u8(sas.val[0], vdupq_n_u8(1)), 2); + sas.val[1] = vshlq_n_u8(vandq_u8(sas.val[1], vdupq_n_u8(1)), 2); + { + auto s16_1 = vmulq_s16(iscales.val[0], vmovl_u8(vget_low_u8 (sas.val[0]))); + auto s16_2 = vmulq_s16(iscales.val[1], vmovl_u8(vget_high_u8(sas.val[0]))); + auto s16_3 = vmulq_s16(iscales.val[2], vmovl_u8(vget_low_u8 (sas.val[1]))); + auto s16_4 = vmulq_s16(iscales.val[3], vmovl_u8(vget_high_u8(sas.val[1]))); + for (int iy = 0; iy < nrc_y; ++iy) { + auto bsums = vld1q_s16_x2(q8.y[iy][ibl].bsums); + auto bs = vpaddq_s16(bsums.val[0], bsums.val[1]); + auto b8 = vget_low_s16(bs); + isum[iy] = vmlal_lane_s16(isum[iy], vget_low_s16 (s16_1), b8, 0); + isum[iy] = vmlal_lane_s16(isum[iy], vget_high_s16(s16_1), b8, 1); + isum[iy] = vmlal_lane_s16(isum[iy], vget_low_s16 (s16_2), b8, 2); + isum[iy] = vmlal_lane_s16(isum[iy], vget_high_s16(s16_2), b8, 3); + b8 = vget_high_s16(bs); + isum[iy] = vmlal_lane_s16(isum[iy], vget_low_s16 (s16_3), b8, 0); + isum[iy] = vmlal_lane_s16(isum[iy], vget_high_s16(s16_3), b8, 1); + isum[iy] = vmlal_lane_s16(isum[iy], vget_low_s16 (s16_4), b8, 2); + isum[iy] = vmlal_lane_s16(isum[iy], vget_high_s16(s16_4), b8, 3); + } + } + for (int is = 0; is < 2; ++is) { + scales.val[0] = vmovl_s16(vget_low_s16 (iscales.val[2*is+0])); + scales.val[1] = vmovl_s16(vget_high_s16(iscales.val[2*is+0])); + scales.val[2] = vmovl_s16(vget_low_s16 (iscales.val[2*is+1])); + scales.val[3] = vmovl_s16(vget_high_s16(iscales.val[2*is+1])); + for (int ib = 0; ib < 4; ++ib) { + auto bits = vld1q_u8_x4(iq4[ibl].qs + 256*is + 64*ib); + prepare_iq4_nl_quants(values, m4, bits, qx); + for (int iy = 0; iy < nrc_y; ++iy) { + auto y = vld1q_s8_x2(q8.y[iy][ibl].qs+128*is+32*ib); + auto sumi = interleaved_dotq(qx, y); + isum[iy] = vmlaq_s32(isum[iy], scales.val[ib], sumi); + } + } + } + for (int iy = 0; iy < nrc_y; ++iy) { + acc[iy] = vfmaq_f32(acc[iy], vdupq_n_f32(q8.scale(iy, ibl)), vcvtq_f32_s32(isum[iy])); + isum[iy] = vdupq_n_s32(0); + } + } + for (int iy = 0; iy < nrc_y; ++iy) { + info.store(ix, iy, vmulq_f32(d4, acc[iy])); + acc[iy] = vdupq_n_f32(0.f); + } + } +} + template <int nrc_y, int k_shift> inline void iq3_4_add_shift(int ibl, const Q8<nrc_y, block_q8_K>& q8, const int8x16x4_t& i8scales, uint8x16_t extra, int32x4_t * isum) { @@ -10571,6 +10768,10 @@ bool MulMat::prepare(int typeA, int typeB, int ne00, MulMat& m, int /*Ny*/) { SET_MUL_MAT_FUNCTIONS(m, mul_mat_iq4_xs_r4_q8_k); expected_Btype = GGML_TYPE_Q8_K32; break; + case GGML_TYPE_IQ4_KS_R4: + SET_MUL_MAT_FUNCTIONS(m, mul_mat_iq4_ks_r4_q8_k); + expected_Btype = GGML_TYPE_Q8_K; + break; case GGML_TYPE_Q2_K_R4: SET_MUL_MAT_FUNCTIONS(m, mul_mat_q2_k_r4_q8_k); expected_Btype = GGML_TYPE_Q8_K; diff --git a/ggml/src/iqk/iqk_quantize.cpp b/ggml/src/iqk/iqk_quantize.cpp index 0007dc04..874c6d55 100644 --- a/ggml/src/iqk/iqk_quantize.cpp +++ b/ggml/src/iqk/iqk_quantize.cpp @@ -3839,6 +3839,107 @@ void vec_dot_iq4_xs_r4_q8_k(int n, float * s, size_t bs, const void * vx, size_t } // +// ========================================= iq4_ks_r4 +// + +void quantize_row_iq4_ks_r4_ref(const float * x, block_iq4_ks_r4 * y, int64_t k) { + quantize_iq4_ks_r4(x, (void *)y, 4, k/4, nullptr); +} + +void quantize_row_iq4_ks_r4(const float * x, void * y, int64_t k) { + quantize_iq4_ks_r4(x, y, 4, k/4, nullptr); +} + +static void repack_iq4_ks(int nrows, int n_per_row, const block_iq4_ks * x, block_iq4_ks_r4 * y) { + GGML_ASSERT(nrows%4 == 0); + GGML_ASSERT(n_per_row%QK_K == 0); + auto row_size = ggml_row_size(GGML_TYPE_IQ4_KS, n_per_row); + int nblock = n_per_row/QK_K; + char * cy = (char *)y; + const char * cx = (const char *)x; + const block_iq4_ks * x4[4]; + for (int row = 0; row < nrows; row += 4) { + float * dptr = (float *)cy; + block_iq4_ks_r4 * y = (block_iq4_ks_r4 *)(dptr + 4); + for (int k = 0; k < 4; ++k) { + auto dk = (const float *)(cx + k*row_size); + dptr[k] = dk[0]; + x4[k] = (const block_iq4_ks *)(dk + 1); + } + for (int ibl = 0; ibl < nblock; ++ibl) { + for (int k = 0; k < 4; ++k) { + for (int ib = 0; ib < QK_K/32; ++ib) { + y[ibl].scales[4*ib+k] = x4[k][ibl].scales[ib]; + for (int i = 0; i < 4; ++i) { + y[ibl].qs[64*ib+4*k+i+ 0] = (x4[k][ibl].qs[16*ib+i+0] & 0xf) | ((x4[k][ibl].qs[16*ib+i+ 8] & 0x0f) << 4); // 0....3 + 8...11 from each row + y[ibl].qs[64*ib+4*k+i+16] = (x4[k][ibl].qs[16*ib+i+0] >> 4) | ((x4[k][ibl].qs[16*ib+i+ 8] & 0xf0)); // 16...19 + 24...27 from each row + y[ibl].qs[64*ib+4*k+i+32] = (x4[k][ibl].qs[16*ib+i+4] & 0xf) | ((x4[k][ibl].qs[16*ib+i+12] & 0x0f) << 4); // 4....7 + 12...15 from each row + y[ibl].qs[64*ib+4*k+i+48] = (x4[k][ibl].qs[16*ib+i+4] >> 4) | ((x4[k][ibl].qs[16*ib+i+12] & 0xf0)); // 20...23 + 28...31 from each row + } + } + } + } + cx += 4*row_size; + cy += 4*row_size; + } +} + +size_t quantize_iq4_ks_r4(const float * src, void * dst, int64_t nrows, int64_t n_per_row, const float * imatrix) { + GGML_ASSERT(nrows%4 == 0); + GGML_ASSERT(n_per_row%QK_K == 0); + char * qcur = (char *)dst; + auto row_size = ggml_row_size(GGML_TYPE_IQ4_KS, n_per_row); + std::vector<char> qtmp(4*row_size); + for (int row = 0; row < nrows; row += 4) { + quantize_iq4_ks(src, (void *)qtmp.data(), 4, n_per_row, imatrix); + repack_iq4_ks(4, n_per_row, (const block_iq4_ks *)qtmp.data(), (block_iq4_ks_r4 *)qcur); + qcur += 4*row_size; + src += 4*n_per_row; + } + return nrows*row_size; +} + +void dequantize_row_iq4_ks_r4(const block_iq4_ks_r4 * x, float * y, int64_t k) { + auto n_per_row = k/4; + float * y4[4] = {y, y + n_per_row, y + 2*n_per_row, y + 3*n_per_row}; + int nblock = n_per_row/QK_K; + const float * dptr = (const float *)x; + x = (const block_iq4_ks_r4 *)(dptr + 4); + for (int ibl = 0; ibl < nblock; ++ibl) { + for (int k = 0; k < 4; ++k) { + const float d = dptr[k]; + for (int ib = 0; ib < QK_K/32; ++ib) { + float dl = d * ((x[ibl].scales[4*ib + k] & 254) - 127); + auto values = iq4k_values + ((x[ibl].scales[4*ib + k] & 1) << 4); + for (int i = 0; i < 4; ++i) { + y4[k][QK_K*ibl+32*ib+i+ 0] = dl * values[x[ibl].qs[64*ib+4*k+i+ 0] & 0xf]; + y4[k][QK_K*ibl+32*ib+i+ 8] = dl * values[x[ibl].qs[64*ib+4*k+i+ 0] >> 4]; + y4[k][QK_K*ibl+32*ib+i+16] = dl * values[x[ibl].qs[64*ib+4*k+i+16] & 0xf]; + y4[k][QK_K*ibl+32*ib+i+24] = dl * values[x[ibl].qs[64*ib+4*k+i+16] >> 4]; + y4[k][QK_K*ibl+32*ib+i+ 4] = dl * values[x[ibl].qs[64*ib+4*k+i+32] & 0xf]; + y4[k][QK_K*ibl+32*ib+i+12] = dl * values[x[ibl].qs[64*ib+4*k+i+32] >> 4]; + y4[k][QK_K*ibl+32*ib+i+20] = dl * values[x[ibl].qs[64*ib+4*k+i+48] & 0xf]; + y4[k][QK_K*ibl+32*ib+i+28] = dl * values[x[ibl].qs[64*ib+4*k+i+48] >> 4]; + } + } + } + } +} + +void vec_dot_iq4_ks_r4_q8_k(int n, float * s, size_t bs, const void * vx, size_t bx, const void * vy, size_t by, int nrc) { +#if GGML_USE_IQK_MULMAT + if (iqk_mul_mat(1, 1, n, GGML_TYPE_IQ4_KS_R4, vx, 0, GGML_TYPE_Q8_K, vy, 0, s, 0, 0, 1)) { + return; + } +#endif + GGML_ASSERT(n%QK4_NL == 0); + GGML_ASSERT(nrc == 1); + GGML_UNUSED(bs); + GGML_UNUSED(bx); + GGML_UNUSED(by); +} + +// // ========================================= iq2_bn_r4 // void quantize_row_iq2_bn_r4_ref(const float * x, block_iq2_bn * y, int64_t k) { diff --git a/ggml/src/iqk/iqk_quantize.h b/ggml/src/iqk/iqk_quantize.h index b8604caa..bcec8432 100644 --- a/ggml/src/iqk/iqk_quantize.h +++ b/ggml/src/iqk/iqk_quantize.h @@ -163,6 +163,12 @@ size_t quantize_iq2_k_r4(const float * GGML_RESTRICT src, void * GGML_RESTRICT d void dequantize_row_iq2_k_r4(const block_iq2_k_r4 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); void vec_dot_iq2_k_r4_q8_k(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void quantize_row_iq4_ks_r4_ref(const float * GGML_RESTRICT x, block_iq4_ks_r4 * GGML_RESTRICT y, int64_t k); +void quantize_row_iq4_ks_r4(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +size_t quantize_iq4_ks_r4(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +void dequantize_row_iq4_ks_r4(const block_iq4_ks_r4 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void vec_dot_iq4_ks_r4_q8_k(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); + void quantize_row_q8_k_r8_ref(const float * GGML_RESTRICT x, block_q8_k_r8 * GGML_RESTRICT y, int64_t k); void quantize_row_q8_k_r8(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); size_t quantize_q8_k_r8(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); diff --git a/include/llama.h b/include/llama.h index 7267fbd4..5480f4e9 100644 --- a/include/llama.h +++ b/include/llama.h @@ -197,6 +197,7 @@ extern "C" { LLAMA_FTYPE_MOSTLY_IQ3_K_R4 = 339, // except 1d tensors LLAMA_FTYPE_MOSTLY_IQ4_K_R4 = 340, // except 1d tensors LLAMA_FTYPE_MOSTLY_IQ5_K_R4 = 341, // except 1d tensors + LLAMA_FTYPE_MOSTLY_IQ4_KS_R4 = 345, // except 1d tensors LLAMA_FTYPE_MOSTLY_Q8_K_R8 = 399, // except 1d tensors LLAMA_FTYPE_GUESSED = 1024, // not specified in the model file diff --git a/src/llama.cpp b/src/llama.cpp index f04d7f6f..3f0ceb0a 100644 --- a/src/llama.cpp +++ b/src/llama.cpp @@ -3868,6 +3868,7 @@ struct llama_model_loader { case GGML_TYPE_Q8_0_R4: ftype = LLAMA_FTYPE_MOSTLY_Q8_0_R4; break; case GGML_TYPE_IQ4_XS: ftype = LLAMA_FTYPE_MOSTLY_IQ4_XS; break; case GGML_TYPE_IQ4_KS: ftype = LLAMA_FTYPE_MOSTLY_IQ4_KS; break; + case GGML_TYPE_IQ4_KS_R4:ftype = LLAMA_FTYPE_MOSTLY_IQ4_KS_R4; break; case GGML_TYPE_IQ4_KSS: ftype = LLAMA_FTYPE_MOSTLY_IQ4_KSS; break; case GGML_TYPE_IQ2_K: ftype = LLAMA_FTYPE_MOSTLY_IQ2_K; break; case GGML_TYPE_IQ2_K_R4:ftype = LLAMA_FTYPE_MOSTLY_IQ2_K_R4;break; @@ -4593,6 +4594,7 @@ static std::string llama_model_ftype_name(llama_ftype ftype) { case LLAMA_FTYPE_MOSTLY_Q8_0_R4: return "Q8_0_R4 - 8.5 bpw"; case LLAMA_FTYPE_MOSTLY_IQ4_XS: return "IQ4_XS - 4.25 bpw"; case LLAMA_FTYPE_MOSTLY_IQ4_KS: return "IQ4_KS - 4.25 bpw"; + case LLAMA_FTYPE_MOSTLY_IQ4_KS_R4:return "IQ4_KS_R4 - 4.25 bpw"; case LLAMA_FTYPE_MOSTLY_IQ4_KSS: return "IQ4_KSS - 4.0 bpw"; case LLAMA_FTYPE_MOSTLY_IQ2_K: return "IQ2_K - 2.375 bpw"; case LLAMA_FTYPE_MOSTLY_IQ2_K_R4: return "IQ2_K_R4 - 2.375 bpw"; @@ -15794,7 +15796,7 @@ static ggml_type llama_tensor_get_type(quantize_state_internal & qs, ggml_type n new_type = !qs.has_output ? GGML_TYPE_IQ4_K : GGML_TYPE_Q5_K; } else if ((ftype == LLAMA_FTYPE_MOSTLY_IQ3_S || ftype == LLAMA_FTYPE_MOSTLY_IQ3_M || ftype == LLAMA_FTYPE_MOSTLY_IQ4_XS || - ftype == LLAMA_FTYPE_MOSTLY_IQ4_KS || ftype == LLAMA_FTYPE_MOSTLY_IQ4_KSS) && !qs.has_output) { + ftype == LLAMA_FTYPE_MOSTLY_IQ4_KS || ftype == LLAMA_FTYPE_MOSTLY_IQ4_KSS || ftype == LLAMA_FTYPE_MOSTLY_IQ4_KS_R4) && !qs.has_output) { new_type = GGML_TYPE_IQ5_K; } else if (new_type != GGML_TYPE_Q8_0 && new_type != GGML_TYPE_Q8_0_R4 && new_type != GGML_TYPE_IQ6_K && new_type != GGML_TYPE_Q6_K_R4 && @@ -15859,6 +15861,9 @@ static ggml_type llama_tensor_get_type(quantize_state_internal & qs, ggml_type n else if (new_type == GGML_TYPE_IQ5_K_R4) { new_type = GGML_TYPE_IQ5_K; } + else if (new_type == GGML_TYPE_IQ4_KS_R4) { + new_type = GGML_TYPE_IQ4_KS; + } else if (new_type == GGML_TYPE_Q4_0_R4) { new_type = GGML_TYPE_Q4_0; } @@ -15949,6 +15954,9 @@ static ggml_type llama_tensor_get_type(quantize_state_internal & qs, ggml_type n ftype == LLAMA_FTYPE_MOSTLY_IQ4_KS || ftype == LLAMA_FTYPE_MOSTLY_IQ4_KSS) && qs.model.hparams.n_gqa() >= 2) { new_type = GGML_TYPE_IQ5_K; } + else if (ftype == LLAMA_FTYPE_MOSTLY_IQ4_KS_R4 && qs.model.hparams.n_gqa() >= 2) { + new_type = GGML_TYPE_IQ5_K_R4; + } else if (ftype == LLAMA_FTYPE_MOSTLY_IQ4_K && qs.model.hparams.n_gqa() >= 2) { new_type = GGML_TYPE_IQ5_K; } @@ -16053,6 +16061,9 @@ static ggml_type llama_tensor_get_type(quantize_state_internal & qs, ggml_type n ftype == LLAMA_FTYPE_MOSTLY_IQ4_NL_R4 || ftype == LLAMA_FTYPE_MOSTLY_IQ4_XS_R4)) { new_type = GGML_TYPE_Q5_K; } + else if (ftype == LLAMA_FTYPE_MOSTLY_IQ4_KS_R4 && i_layer < n_layer/8 && !qs.has_imatrix) { + new_type = GGML_TYPE_Q5_K_R4; + } else if (ftype == LLAMA_FTYPE_MOSTLY_Q5_K_M && use_more_bits(i_layer, n_layer)) new_type = GGML_TYPE_Q6_K; else if (ftype == LLAMA_FTYPE_MOSTLY_Q4_K_S && arch != LLM_ARCH_FALCON && i_layer < n_layer/8) { new_type = GGML_TYPE_Q5_K; @@ -16155,7 +16166,7 @@ static ggml_type llama_tensor_get_type(quantize_state_internal & qs, ggml_type n new_type == GGML_TYPE_IQ2_KS || new_type == GGML_TYPE_IQ4_KSS || new_type == GGML_TYPE_Q6_K_R4 || new_type == GGML_TYPE_Q5_K_R4 || new_type == GGML_TYPE_Q3_K_R4 || new_type == GGML_TYPE_Q2_K_R4 || new_type == GGML_TYPE_IQ4_K_R4|| new_type == GGML_TYPE_Q8_K_R8 || new_type == GGML_TYPE_IQ3_K_R4|| - new_type == GGML_TYPE_IQ2_K_R4|| new_type == GGML_TYPE_IQ5_K_R4) { + new_type == GGML_TYPE_IQ2_K_R4|| new_type == GGML_TYPE_IQ5_K_R4|| new_type == GGML_TYPE_IQ4_KS_R4) { int nx = tensor->ne[0]; int ny = tensor->ne[1]; if (nx % QK_K != 0) { @@ -16191,6 +16202,7 @@ static ggml_type llama_tensor_get_type(quantize_state_internal & qs, ggml_type n case GGML_TYPE_IQ3_K_R4: case GGML_TYPE_IQ4_KSS: case GGML_TYPE_IQ4_KS: + case GGML_TYPE_IQ4_KS_R4: case GGML_TYPE_IQ4_XS_R4: case GGML_TYPE_IQ4_XS: new_type = GGML_TYPE_IQ4_NL; break; case GGML_TYPE_IQ4_K: @@ -16322,6 +16334,7 @@ static void llama_model_quantize_internal(const std::string & fname_inp, const s case LLAMA_FTYPE_MOSTLY_Q8_0_R4: default_type = GGML_TYPE_Q8_0_R4; break; case LLAMA_FTYPE_MOSTLY_IQ4_XS: default_type = GGML_TYPE_IQ4_XS; break; case LLAMA_FTYPE_MOSTLY_IQ4_KS: default_type = GGML_TYPE_IQ4_KS; break; + case LLAMA_FTYPE_MOSTLY_IQ4_KS_R4:default_type = GGML_TYPE_IQ4_KS_R4;break; case LLAMA_FTYPE_MOSTLY_IQ4_KSS: default_type = GGML_TYPE_IQ4_KSS; break; case LLAMA_FTYPE_MOSTLY_IQ2_K: default_type = GGML_TYPE_IQ2_K; break; case LLAMA_FTYPE_MOSTLY_IQ2_K_R4:default_type = GGML_TYPE_IQ2_K_R4;break; @@ -16752,6 +16765,10 @@ static void llama_model_quantize_internal(const std::string & fname_inp, const s if (tensor->ne[1] % 4 != 0) new_type = GGML_TYPE_IQ5_K; else chunk_size_multiplier = 4; } + else if (new_type == GGML_TYPE_IQ4_KS_R4) { + if (tensor->ne[1] % 4 != 0) new_type = GGML_TYPE_IQ4_KS; + else chunk_size_multiplier = 4; + } else if (new_type == GGML_TYPE_BF16_R16) { if (tensor->ne[1] % 16 != 0) new_type = GGML_TYPE_BF16; else chunk_size_multiplier = 16; |