From edb6c8bb66dd2777dfe704d5f84cad4960ebe3e4 Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Thu, 4 May 2023 20:53:14 +0300 Subject: [PATCH] ggml : remove Q4_1 bit shuffling (ARM NEON + reference) --- ggml.c | 426 ++++++++++++++++----------------------------------------- 1 file changed, 116 insertions(+), 310 deletions(-) diff --git a/ggml.c b/ggml.c index 40cec55a6..49e9083f0 100644 --- a/ggml.c +++ b/ggml.c @@ -837,17 +837,19 @@ static_assert(sizeof(block_q8_1) == 3*sizeof(float) + QK8_1, "wrong q8_1 block s // reference implementation for deterministic creation of model files static void quantize_row_q4_0_reference(const float * restrict x, block_q4_0 * restrict y, int k) { - assert(QK4_0 / 16 == 0); - assert(k % QK4_0 == 0); + static const int qk = QK4_0; - const int nb = k / QK4_0; + assert(qk / 16 == 0); + assert(k % qk == 0); + + const int nb = k / qk; for (int i = 0; i < nb; i++) { float amax = 0.0f; // absolute max float max = 0.0f; - for (int l = 0; l < QK4_0; l++) { - const float v = x[i*QK4_0 + l]; + for (int l = 0; l < qk; l++) { + const float v = x[i*qk + l]; if (amax < fabsf(v)) { amax = fabsf(v); max = v; @@ -862,9 +864,9 @@ static void quantize_row_q4_0_reference(const float * restrict x, block_q4_0 * r uint64_t qs[QK4_0 / 16] = {0}; // pack first half of weights into low nibbles and second half into high nibbles - for (int l = 0; l < QK4_0/2; ++l) { - const float v0 = x[i*QK4_0 + 0 + l]*id; - const float v1 = x[i*QK4_0 + QK4_0/2 + l]*id; + for (int l = 0; l < qk/2; ++l) { + const float v0 = x[i*qk + 0 + l]*id; + const float v1 = x[i*qk + qk/2 + l]*id; const uint64_t vi0 = MIN(15, (int8_t)(v0 + 8.5f)); const uint64_t vi1 = MIN(15, (int8_t)(v1 + 8.5f)); @@ -877,176 +879,55 @@ static void quantize_row_q4_0_reference(const float * restrict x, block_q4_0 * r } } -static void quantize_row_q4_0(const float * restrict x, void * restrict vy, int k) { - assert(k % QK4_0 == 0); - - block_q4_0 * restrict y = vy; - +static void quantize_row_q4_0(const float * restrict x, void * restrict y, int k) { quantize_row_q4_0_reference(x, y, k); } -static void quantize_row_q4_1_reference(const float * restrict x, void * restrict vy, int k) { - assert(k % QK4_1 == 0); - const int nb = k / QK4_1; +static void quantize_row_q4_1_reference(const float * restrict x, block_q4_1 * restrict y, int k) { + const int qk = QK4_1; - block_q4_1 * restrict y = vy; + assert(qk / 16 == 0); + assert(k % qk == 0); - uint8_t pp[QK4_1/2]; + const int nb = k / qk; for (int i = 0; i < nb; i++) { float min = FLT_MAX; float max = -FLT_MAX; - for (int l = 0; l < QK4_1; l++) { - const float v = x[i*QK4_1 + l]; + for (int l = 0; l < qk; l++) { + const float v = x[i*qk + l]; + if (v < min) min = v; if (v > max) max = v; } - const float d = (max - min) / ((1 << 4) - 1); + const float d = (max - min) / ((1 << 4) - 1); const float id = d ? 1.0f/d : 0.0f; y[i].d = d; y[i].m = min; - for (int l = 0; l < QK4_1; l += 2) { - const float v0 = (x[i*QK4_1 + l + 0] - min)*id; - const float v1 = (x[i*QK4_1 + l + 1] - min)*id; + uint64_t qs[QK4_1 / 16] = {0}; - const uint8_t vi0 = roundf(v0); - const uint8_t vi1 = roundf(v1); + // pack first half of weights into low nibbles and second half into high nibbles + for (int l = 0; l < qk/2; ++l) { + const float v0 = (x[i*qk + 0 + l] - min)*id; + const float v1 = (x[i*qk + qk/2 + l] - min)*id; - assert(vi0 < 16); - assert(vi1 < 16); + const uint64_t vi0 = MIN(15, (int8_t)(v0 + 0.5f)); + const uint64_t vi1 = MIN(15, (int8_t)(v1 + 0.5f)); - pp[l/2] = vi0 | (vi1 << 4); + qs[l/8] |= vi0 << (8*(l & 7)); + qs[l/8] |= vi1 << (8*(l & 7) + 4); } - memcpy(y[i].qs, pp, sizeof(pp)); + memcpy(y[i].qs, qs, sizeof(qs)); } } -static void quantize_row_q4_1(const float * restrict x, void * restrict vy, int k) { - assert(k % QK4_1 == 0); - - const int nb = k / QK4_1; - - block_q4_1 * restrict y = vy; - -#if defined(__AVX2__) - for (int i = 0; i < nb; i++) { - // Load elements into 4 AVX vectors - __m256 v0 = _mm256_loadu_ps( x ); - __m256 v1 = _mm256_loadu_ps( x + 8 ); - __m256 v2 = _mm256_loadu_ps( x + 16 ); - __m256 v3 = _mm256_loadu_ps( x + 24 ); - x += 32; - - // Compute max for the block - __m256 vmax; - vmax = _mm256_max_ps( v0, v1 ); - vmax = _mm256_max_ps( vmax, v2 ); - vmax = _mm256_max_ps( vmax, v3 ); - - __m128 max4 = _mm_max_ps( _mm256_extractf128_ps( vmax, 1 ), _mm256_castps256_ps128( vmax ) ); - max4 = _mm_max_ps( max4, _mm_movehl_ps( max4, max4 ) ); - max4 = _mm_max_ss( max4, _mm_movehdup_ps( max4 ) ); - const float maxScalar = _mm_cvtss_f32( max4 ); - - // Compute min for the block - __m256 vmin; - vmin = _mm256_min_ps( v0, v1 ); - vmin = _mm256_min_ps( vmin, v2 ); - vmin = _mm256_min_ps( vmin, v3 ); - - __m128 min4 = _mm_min_ps( _mm256_extractf128_ps( vmin, 1 ), _mm256_castps256_ps128( vmin ) ); - min4 = _mm_min_ps( min4, _mm_movehl_ps( min4, min4 ) ); - min4 = _mm_min_ss( min4, _mm_movehdup_ps( min4 ) ); - const float minScalar = _mm_cvtss_f32( min4 ); - - // Quantize these floats - const float d = (maxScalar - minScalar) / ((1 << 4) - 1); - const float id = d ? 1.0f/d : 0.0f; - - y[i].m = minScalar; - y[i].d = d; - - // x = (x-min)*id - const __m256 mul = _mm256_set1_ps( id ); - const __m256 off = _mm256_set1_ps( minScalar ); - v0 = _mm256_mul_ps( _mm256_sub_ps( v0, off ), mul ); - v1 = _mm256_mul_ps( _mm256_sub_ps( v1, off ), mul ); - v2 = _mm256_mul_ps( _mm256_sub_ps( v2, off ), mul ); - v3 = _mm256_mul_ps( _mm256_sub_ps( v3, off ), mul ); - - // Round to nearest integer - v0 = _mm256_round_ps( v0, _MM_ROUND_NEAREST ); - v1 = _mm256_round_ps( v1, _MM_ROUND_NEAREST ); - v2 = _mm256_round_ps( v2, _MM_ROUND_NEAREST ); - v3 = _mm256_round_ps( v3, _MM_ROUND_NEAREST ); - - // Convert floats to integers - __m256i i0 = _mm256_cvtps_epi32( v0 ); - __m256i i1 = _mm256_cvtps_epi32( v1 ); - __m256i i2 = _mm256_cvtps_epi32( v2 ); - __m256i i3 = _mm256_cvtps_epi32( v3 ); - - // Convert int32 to int16 - i0 = _mm256_packs_epi32( i0, i1 ); // 0, 1, 2, 3, 8, 9, 10, 11, 4, 5, 6, 7, 12, 13, 14, 15 - i2 = _mm256_packs_epi32( i2, i3 ); // 16, 17, 18, 19, 24, 25, 26, 27, 20, 21, 22, 23, 28, 29, 30, 31 - // Convert int16 to int8 - i0 = _mm256_packs_epi16( i0, i2 ); // 0, 1, 2, 3, 8, 9, 10, 11, 16, 17, 18, 19, 24, 25, 26, 27, 4, 5, 6, 7, 12, 13, 14, 15, 20, 21, 22, 23, 28, 29, 30, 31 - - // We got our precious signed bytes, but the order is now wrong - // These AVX2 pack instructions process 16-byte pieces independently - // The following instruction is fixing the order - const __m256i perm = _mm256_setr_epi32( 0, 4, 1, 5, 2, 6, 3, 7 ); - i0 = _mm256_permutevar8x32_epi32( i0, perm ); - - // Compress the vector into 4 bit/value, and store - __m128i res = packNibbles( i0 ); - _mm_storeu_si128( ( __m128i* )y[i].qs, res ); - } -#elif __ARM_NEON - for (int i = 0; i < nb; i++) { - float32x4_t srcv[8]; - float32x4_t minv[8]; - float32x4_t maxv[8]; - - for (int l = 0; l < 8; l++) srcv[l] = vld1q_f32(x + i*QK4_1 + 4*l); - - for (int l = 0; l < 4; l++) minv[2*l] = vminq_f32(srcv[2*l], srcv[2*l + 1]); - for (int l = 0; l < 2; l++) minv[4*l] = vminq_f32(minv[4*l], minv[4*l + 2]); - for (int l = 0; l < 1; l++) minv[8*l] = vminq_f32(minv[8*l], minv[8*l + 4]); - - for (int l = 0; l < 4; l++) maxv[2*l] = vmaxq_f32(srcv[2*l], srcv[2*l + 1]); - for (int l = 0; l < 2; l++) maxv[4*l] = vmaxq_f32(maxv[4*l], maxv[4*l + 2]); - for (int l = 0; l < 1; l++) maxv[8*l] = vmaxq_f32(maxv[8*l], maxv[8*l + 4]); - - const float min = vminvq_f32(minv[0]); - const float max = vmaxvq_f32(maxv[0]); - - const float d = (max - min) / ((1 << 4) - 1); - const float id = d ? 1.0f/d : 0.0f; - - y[i].d = d; - y[i].m = min; - - const float32x4_t minv0 = vdupq_n_f32(min); - - for (int l = 0; l < 8; l++) { - const float32x4_t v = vmulq_n_f32(vsubq_f32(srcv[l], minv0), id); - const float32x4_t vf = vaddq_f32(v, vdupq_n_f32(0.5f)); // needed to round to nearest - const int32x4_t vi = vcvtq_s32_f32(vf); - - y[i].qs[2*l + 0] = vgetq_lane_s32(vi, 0) | (vgetq_lane_s32(vi, 1) << 4); - y[i].qs[2*l + 1] = vgetq_lane_s32(vi, 2) | (vgetq_lane_s32(vi, 3) << 4); - } - } -#else - // scalar - quantize_row_q4_1_reference(x, vy, k); -#endif +static void quantize_row_q4_1(const float * restrict x, void * restrict y, int k) { + quantize_row_q4_1_reference(x, y, k); } // reference implementation for deterministic creation of model files @@ -1546,13 +1427,13 @@ static void quantize_row_q8_1(const float * restrict x, void * restrict vy, int #endif } -static void dequantize_row_q4_0(const void * restrict vx, float * restrict y, int k) { - assert(QK4_0 / 16 == 0); - assert(k % QK4_0 == 0); +static void dequantize_row_q4_0(const block_q4_0 * restrict x, float * restrict y, int k) { + static const int qk = QK4_0; - const int nb = k / QK4_0; + assert(qk / 16 == 0); + assert(k % qk == 0); - const block_q4_0 * restrict x = vx; + const int nb = k / qk; for (int i = 0; i < nb; i++) { const float d = x[i].d; @@ -1560,126 +1441,49 @@ static void dequantize_row_q4_0(const void * restrict vx, float * restrict y, in // unpack nibbles into bytes uint64_t qs[QK4_0 / 8] = {0}; - memcpy(qs + 0, x[i].qs, sizeof(x[i].qs)); - memcpy(qs + QK4_0 / 16, x[i].qs, sizeof(x[i].qs)); + memcpy(qs + 0, x[i].qs, sizeof(x[i].qs)); - for (int l = 0; l < QK4_0 / 16; ++l) { - qs[l ] = (qs[l ] & 0x0F0F0F0F0F0F0F0FULL) >> 0; - qs[l + QK4_0/16] = (qs[l + QK4_0/16] & 0xF0F0F0F0F0F0F0F0ULL) >> 4; + for (int l = 0; l < qk / 16; ++l) { + qs[l + qk/16] = (qs[l] & 0xF0F0F0F0F0F0F0F0ULL) >> 4; + qs[l + 0 ] = (qs[l] & 0x0F0F0F0F0F0F0F0FULL) >> 0; } const uint8_t * restrict qsp = (const uint8_t * restrict) qs; - for (int l = 0; l < QK4_0; ++l) { - y[i*QK4_0 + l] = (qsp[l] - 8)*d; + for (int l = 0; l < qk; ++l) { + y[i*qk + l] = (qsp[l] - 8)*d; } } } -static void dequantize_row_q4_1(const void * restrict vx, float * restrict y, int k) { - assert(k % QK4_1 == 0); - const int nb = k / QK4_1; +static void dequantize_row_q4_1(const block_q4_1 * restrict x, float * restrict y, int k) { + static const int qk = QK4_1; - const block_q4_1 * restrict x = vx; + assert(qk / 16 == 0); + assert(k % qk == 0); -#if defined(__AVX2__) - for (int i = 0; i < nb; i++) { - const __m256 d_v = _mm256_broadcast_ss(&x[i].d); - const __m256 d_m = _mm256_broadcast_ss(&x[i].m); + const int nb = k / qk; - const uint8_t * restrict pp = x[i].qs; - - for (int l = 0; l < QK4_1; l += 32) { - // Load 32x4-bit integers into 32x8-bit integers - __m256i vx8 = bytes_from_nibbles_32(pp+l/2); - - // Convert to 16-bit int - const __m256i vx16_lo = _mm256_cvtepi8_epi16(_mm256_extracti128_si256(vx8, 0)); - const __m256i vx16_hi = _mm256_cvtepi8_epi16(_mm256_extracti128_si256(vx8, 1)); - - // Convert to 32-bit int -> float 32 - const __m256 vf[4] = { - _mm256_cvtepi32_ps(_mm256_cvtepi16_epi32(_mm256_extracti128_si256(vx16_lo, 0))), - _mm256_cvtepi32_ps(_mm256_cvtepi16_epi32(_mm256_extracti128_si256(vx16_lo, 1))), - _mm256_cvtepi32_ps(_mm256_cvtepi16_epi32(_mm256_extracti128_si256(vx16_hi, 0))), - _mm256_cvtepi32_ps(_mm256_cvtepi16_epi32(_mm256_extracti128_si256(vx16_hi, 1))) - }; - - // Scale, add m and store - for (int j = 0; j < 4; j++) { - const __m256 result = _mm256_add_ps(_mm256_mul_ps(vf[j], d_v), d_m); - _mm256_storeu_ps(y + i * QK4_1 + l + j*8, result); - } - } - } -#elif defined(__ARM_NEON) - for (int i = 0; i < nb; i++) { - const float32x4_t vd = vdupq_n_f32(x[i].d); - const float32x4_t vm = vdupq_n_f32(x[i].m); - - const uint8_t * restrict pp = x[i].qs; - - for (int l = 0; l < QK4_1; l += 16) { - // Load 16x4-bit integers into 8x8-bit integers - const uint8x8_t v8 = vld1_u8(pp + l/2); - - // Expand 4-bit qs to 8-bit bytes - const uint8x8_t v0 = vand_u8(v8, vdup_n_u8(0x0F)); - const uint8x8_t v1 = vshr_n_u8(v8, 4); - - // Interleave and combine - const uint8x8_t vx_0 = vzip1_u8(v0, v1); - const uint8x8_t vx_1 = vzip2_u8(v0, v1); - - const uint8x16_t vq = vcombine_u8(vx_0, vx_1); - - // convert to 2x uint16x8_t - const uint16x8_t vi_0 = vmovl_u8(vget_low_u8 (vq)); - const uint16x8_t vi_1 = vmovl_u8(vget_high_u8(vq)); - - // convert to 4x float32x4_t - const float32x4_t vf_0 = vcvtq_f32_u32(vmovl_u16(vget_low_u16 (vi_0))); - const float32x4_t vf_1 = vcvtq_f32_u32(vmovl_u16(vget_high_u16(vi_0))); - const float32x4_t vf_2 = vcvtq_f32_u32(vmovl_u16(vget_low_u16 (vi_1))); - const float32x4_t vf_3 = vcvtq_f32_u32(vmovl_u16(vget_high_u16(vi_1))); - - // multiply by d and add m - const float32x4_t r0 = vmlaq_f32(vm, vf_0, vd); - const float32x4_t r1 = vmlaq_f32(vm, vf_1, vd); - const float32x4_t r2 = vmlaq_f32(vm, vf_2, vd); - const float32x4_t r3 = vmlaq_f32(vm, vf_3, vd); - - // Store - vst1q_f32(y + i*QK4_1 + l + 0, r0); - vst1q_f32(y + i*QK4_1 + l + 4, r1); - vst1q_f32(y + i*QK4_1 + l + 8, r2); - vst1q_f32(y + i*QK4_1 + l + 12, r3); - } - } -#else for (int i = 0; i < nb; i++) { const float d = x[i].d; const float m = x[i].m; - const uint8_t * restrict pp = x[i].qs; + // unpack nibbles into bytes + uint64_t qs[QK4_0 / 8] = {0}; - for (int l = 0; l < QK4_1; l += 2) { - const uint8_t vi = pp[l/2]; + memcpy(qs + 0, x[i].qs, sizeof(x[i].qs)); - const int8_t vi0 = vi & 0x0F; - const int8_t vi1 = vi >> 4; + for (int l = 0; l < qk / 16; ++l) { + qs[l + qk/16] = (qs[l] & 0xF0F0F0F0F0F0F0F0ULL) >> 4; + qs[l + 0 ] = (qs[l] & 0x0F0F0F0F0F0F0F0FULL) >> 0; + } - const float v0 = vi0*d + m; - const float v1 = vi1*d + m; + const uint8_t * restrict qsp = (const uint8_t * restrict) qs; - y[i*QK4_1 + l + 0] = v0; - y[i*QK4_1 + l + 1] = v1; - - assert(!isnan(y[i*QK4_1 + l + 0])); - assert(!isnan(y[i*QK4_1 + l + 1])); + for (int l = 0; l < qk; ++l) { + y[i*qk + l] = qsp[l]*d + m; } } -#endif } static void dequantize_row_q4_2(const void * restrict vx, float * restrict y, int k) { @@ -1810,7 +1614,7 @@ static void ggml_vec_dot_q8_0_q8_0(const int n, float * restrict s, const void * static const quantize_fns_t quantize_fns[GGML_TYPE_COUNT] = { [GGML_TYPE_Q4_0] = { - .dequantize_row_q = dequantize_row_q4_0, + .dequantize_row_q = (dequantize_row_q_t) dequantize_row_q4_0, .quantize_row_q = quantize_row_q4_0, .quantize_row_q_reference = (quantize_row_q_t) quantize_row_q4_0_reference, .quantize_row_q_dot = quantize_row_q8_0, @@ -1818,7 +1622,7 @@ static const quantize_fns_t quantize_fns[GGML_TYPE_COUNT] = { .vec_dot_type = GGML_TYPE_Q8_0, }, [GGML_TYPE_Q4_1] = { - .dequantize_row_q = dequantize_row_q4_1, + .dequantize_row_q = (dequantize_row_q_t)dequantize_row_q4_1, .quantize_row_q = quantize_row_q4_1, .quantize_row_q_reference = (quantize_row_q_t) quantize_row_q4_1_reference, .quantize_row_q_dot = quantize_row_q8_1, @@ -2467,9 +2271,10 @@ inline static void ggml_vec_dot_f16(const int n, float * restrict s, ggml_fp16_t } static void ggml_vec_dot_q4_0_q8_0(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { - const int nb = n / QK8_0; + const int qk = QK8_0; + const int nb = n / qk; - assert(n % QK8_0 == 0); + assert(n % qk == 0); assert(nb % 2 == 0); const block_q4_0 * restrict x = vx; @@ -2604,41 +2409,45 @@ static void ggml_vec_dot_q4_0_q8_0(const int n, float * restrict s, const void * #else // scalar float sumf = 0.0; - for (int i = 0; i < nb; i++) { - const float d0 = x[i].d; - const float d1 = y[i].d; - const uint8_t * restrict p0 = x[i].qs; - const int8_t * restrict p1 = y[i].qs; + for (int i = 0; i < nb; i++) { + // unpack nibbles into bytes + uint64_t qs[QK8_0 / 8] = {0}; + + memcpy(qs + 0, x[i].qs, sizeof(x[i].qs)); + + for (int l = 0; l < qk / 16; ++l) { + qs[l + qk/16] = (qs[l] & 0xF0F0F0F0F0F0F0F0ULL) >> 4; + qs[l + 0 ] = (qs[l] & 0x0F0F0F0F0F0F0F0FULL) >> 0; + } + + const uint8_t * restrict px = (const uint8_t * restrict) qs; + const int8_t * restrict py = y[i].qs; int sumi = 0; - for (int j = 0; j < QK8_0/2; j++) { - const uint8_t v0 = p0[j]; - const int i0 = (int8_t) (v0 & 0x0F) - 8; - const int i1 = (int8_t) (v0 >> 4) - 8; - - const int i2 = p1[2*j + 0]; - const int i3 = p1[2*j + 1]; - - sumi += i0*i2 + i1*i3; + for (int j = 0; j < qk; ++j) { + sumi += (px[j] - 8) * py[j]; } - sumf += d0*d1*sumi; + + sumf += (x[i].d*y[i].d)*sumi; } + *s = sumf; #endif } static void ggml_vec_dot_q4_1_q8_1(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { - const int nb = n / QK8_1; + const int qk = QK8_1; + const int nb = n / qk; - assert(n % QK8_1 == 0); + assert(n % qk == 0); assert(nb % 2 == 0); const block_q4_1 * restrict x = vx; const block_q8_1 * restrict y = vy; - // TODO: add AVX / WASM SIMD / etc + // TODO: add WASM SIMD #if defined(__ARM_NEON) float32x4_t sumv0 = vdupq_n_f32(0.0f); float32x4_t sumv1 = vdupq_n_f32(0.0f); @@ -2664,12 +2473,6 @@ static void ggml_vec_dot_q4_1_q8_1(const int n, float * restrict s, const void * const int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); const int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); - // interleave - const int8x16_t v0_0lz = vzip1q_s8(v0_0l, v0_0h); - const int8x16_t v0_0hz = vzip2q_s8(v0_0l, v0_0h); - const int8x16_t v0_1lz = vzip1q_s8(v0_1l, v0_1h); - const int8x16_t v0_1hz = vzip2q_s8(v0_1l, v0_1h); - // load y const int8x16_t v1_0l = vld1q_s8(y0->qs); const int8x16_t v1_0h = vld1q_s8(y0->qs + 16); @@ -2678,21 +2481,21 @@ static void ggml_vec_dot_q4_1_q8_1(const int n, float * restrict s, const void * #if defined(__ARM_FEATURE_DOTPROD) // dot product into int32x4_t - const int32x4_t p_0 = vdotq_s32(vdotq_s32(vdupq_n_s32(0), v0_0lz, v1_0l), v0_0hz, v1_0h); - const int32x4_t p_1 = vdotq_s32(vdotq_s32(vdupq_n_s32(0), v0_1lz, v1_1l), v0_1hz, v1_1h); + const int32x4_t p_0 = vdotq_s32(vdotq_s32(vdupq_n_s32(0), v0_0l, v1_0l), v0_0h, v1_0h); + const int32x4_t p_1 = vdotq_s32(vdotq_s32(vdupq_n_s32(0), v0_1l, v1_1l), v0_1h, v1_1h); sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(p_0), x0->d*y0->d); sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(p_1), x1->d*y1->d); #else - const int16x8_t pl0l = vmull_s8(vget_low_s8 (v0_0lz), vget_low_s8 (v1_0l)); - const int16x8_t pl0h = vmull_s8(vget_high_s8(v0_0lz), vget_high_s8(v1_0l)); - const int16x8_t ph0l = vmull_s8(vget_low_s8 (v0_0hz), vget_low_s8 (v1_0h)); - const int16x8_t ph0h = vmull_s8(vget_high_s8(v0_0hz), vget_high_s8(v1_0h)); + const int16x8_t pl0l = vmull_s8(vget_low_s8 (v0_0l), vget_low_s8 (v1_0l)); + const int16x8_t pl0h = vmull_s8(vget_high_s8(v0_0l), vget_high_s8(v1_0l)); + const int16x8_t ph0l = vmull_s8(vget_low_s8 (v0_0h), vget_low_s8 (v1_0h)); + const int16x8_t ph0h = vmull_s8(vget_high_s8(v0_0h), vget_high_s8(v1_0h)); - const int16x8_t pl1l = vmull_s8(vget_low_s8 (v0_1lz), vget_low_s8 (v1_1l)); - const int16x8_t pl1h = vmull_s8(vget_high_s8(v0_1lz), vget_high_s8(v1_1l)); - const int16x8_t ph1l = vmull_s8(vget_low_s8 (v0_1hz), vget_low_s8 (v1_1h)); - const int16x8_t ph1h = vmull_s8(vget_high_s8(v0_1hz), vget_high_s8(v1_1h)); + const int16x8_t pl1l = vmull_s8(vget_low_s8 (v0_1l), vget_low_s8 (v1_1l)); + const int16x8_t pl1h = vmull_s8(vget_high_s8(v0_1l), vget_high_s8(v1_1l)); + const int16x8_t ph1l = vmull_s8(vget_low_s8 (v0_1h), vget_low_s8 (v1_1h)); + const int16x8_t ph1h = vmull_s8(vget_high_s8(v0_1h), vget_high_s8(v1_1h)); const int32x4_t pl0 = vaddq_s32(vpaddlq_s16(pl0l), vpaddlq_s16(pl0h)); const int32x4_t ph0 = vaddq_s32(vpaddlq_s16(ph0l), vpaddlq_s16(ph0h)); @@ -2738,27 +2541,30 @@ static void ggml_vec_dot_q4_1_q8_1(const int n, float * restrict s, const void * #else // scalar float sumf = 0.0; + for (int i = 0; i < nb; i++) { - const float d0 = x[i].d; - const float m0 = x[i].m; - const float d1 = y[i].d; + // unpack nibbles into bytes + uint64_t qs[QK8_1 / 8] = {0}; - const uint8_t * restrict p0 = x[i].qs; - const int8_t * restrict p1 = y[i].qs; + memcpy(qs + 0, x[i].qs, sizeof(x[i].qs)); - // TODO: this is very slow .. - for (int j = 0; j < QK8_1/2; j++) { - const uint8_t v0 = p0[j]; - - const float f0 = d0*(v0 & 0x0F) + m0; - const float f1 = d0*(v0 >> 4) + m0; - - const float f2 = d1*p1[2*j + 0]; - const float f3 = d1*p1[2*j + 1]; - - sumf += f0*f2 + f1*f3; + for (int l = 0; l < qk / 16; ++l) { + qs[l + qk/16] = (qs[l] & 0xF0F0F0F0F0F0F0F0ULL) >> 4; + qs[l + 0 ] = (qs[l] & 0x0F0F0F0F0F0F0F0FULL) >> 0; } + + const uint8_t * restrict px = (const uint8_t * restrict) qs; + const int8_t * restrict py = y[i].qs; + + int sumi = 0; + + for (int j = 0; j < qk; ++j) { + sumi += px[j]*py[j]; + } + + sumf += (x[i].d*y[i].d)*sumi + x[i].m*(y[i].s0 + y[i].s1); } + *s = sumf; #endif } @@ -12703,7 +12509,7 @@ size_t ggml_quantize_q4_0(const float * src, void * dst, int n, int k, int64_t * const int nb = k / QK4_0; for (int j = 0; j < n; j += k) { - block_q4_0 * restrict y = (block_q4_0 *)dst + j/QK4_0; + block_q4_0 * restrict y = (block_q4_0 *) dst + j/QK4_0; quantize_row_q4_0_reference(src + j, y, k); @@ -12726,7 +12532,7 @@ size_t ggml_quantize_q4_1(const float * src, void * dst, int n, int k, int64_t * const int nb = k / QK4_1; for (int j = 0; j < n; j += k) { - block_q4_1 * restrict y = (block_q4_1 *)dst + j/QK4_1; + block_q4_1 * restrict y = (block_q4_1 *) dst + j/QK4_1; quantize_row_q4_1_reference(src + j, y, k);