vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

ggml : initial ARM_NEON 2x F16 Q4_0 implementation

Browse source
This commit is contained in:
Georgi Gerganov 2023-04-17 16:17:06 +03:00
parent efd05648c8
commit 7da998da01
No known key found for this signature in database
GPG key ID: 449E073F9DC10735
2 changed files with 234 additions and 102 deletions
Download patch file Download diff file Expand all files Collapse all files

9
Makefile
View file

@ -31,10 +31,15 @@ endif
# #
# keep standard at C11 and C++11 # keep standard at C11 and C++11
CFLAGS = -I. -O3 -DNDEBUG -std=c11 -fPIC CFLAGS = -I. -O3 -std=c11 -fPIC
CXXFLAGS = -I. -I./examples -O3 -DNDEBUG -std=c++11 -fPIC CXXFLAGS = -I. -I./examples -O3 -std=c++11 -fPIC
LDFLAGS = LDFLAGS =
ifndef LLAMA_NO_NDEBUG
CFLAGS += -DNDEBUG
CXXFLAGS += -DNDEBUG
endif
# warnings # warnings
CFLAGS += -Wall -Wextra -Wpedantic -Wcast-qual -Wdouble-promotion -Wshadow -Wstrict-prototypes -Wpointer-arith -Wno-unused-function CFLAGS += -Wall -Wextra -Wpedantic -Wcast-qual -Wdouble-promotion -Wshadow -Wstrict-prototypes -Wpointer-arith -Wno-unused-function
CXXFLAGS += -Wall -Wextra -Wpedantic -Wcast-qual -Wno-unused-function -Wno-multichar CXXFLAGS += -Wall -Wextra -Wpedantic -Wcast-qual -Wno-unused-function -Wno-multichar

327
ggml.c
View file

@ -572,23 +572,27 @@ uint8x8_t vzip2_u8(uint8x8_t a, uint8x8_t b) {
#define QK4_0 32 #define QK4_0 32
typedef struct { typedef struct {
float d; // delta ggml_fp16_t d0; // delta 0
ggml_fp16_t d1; // delta 1
uint8_t qs[QK4_0 / 2]; // nibbles / quants uint8_t qs[QK4_0 / 2]; // nibbles / quants
} block_q4_0; } block_q4_0;
static_assert(sizeof(block_q4_0) == sizeof(float) + QK4_0 / 2, "wrong q4_0 block size/padding"); static_assert(sizeof(block_q4_0) == 2*sizeof(ggml_fp16_t) + QK4_0 / 2, "wrong q4_0 block size/padding");
#define QK4_1 32 #define QK4_1 32
typedef struct { typedef struct {
float d; // delta float d; // delta
float m; // min float m; // min
uint8_t qs[QK4_1 / 2]; // nibbles / quants
uint8_t qs[QK4_1 / 2]; // nibbles / quants
} block_q4_1; } block_q4_1;
static_assert(sizeof(block_q4_1) == sizeof(float) * 2 + QK4_1 / 2, "wrong q4_1 block size/padding"); static_assert(sizeof(block_q4_1) == sizeof(float) * 2 + QK4_1 / 2, "wrong q4_1 block size/padding");
#define QK8_0 32 #define QK8_0 32
typedef struct { typedef struct {
float d; // delta float d; // delta
int8_t qs[QK8_0]; // quants
int8_t qs[QK8_0]; // quants
} block_q8_0; } block_q8_0;
static_assert(sizeof(block_q8_0) == sizeof(float) + QK8_0, "wrong q8_0 block size/padding"); static_assert(sizeof(block_q8_0) == sizeof(float) + QK8_0, "wrong q8_0 block size/padding");
@ -596,26 +600,36 @@ static_assert(sizeof(block_q8_0) == sizeof(float) + QK8_0, "wrong q8_0 block siz
// reference implementation for deterministic creation of model files // 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) { static void quantize_row_q4_0_reference(const float * restrict x, block_q4_0 * restrict y, int k) {
assert(k % QK4_0 == 0); assert(k % QK4_0 == 0);
const int nb = k / QK4_0; const int nb = k / QK4_0;
uint8_t pp[QK4_0/2];
for (int i = 0; i < nb; i++) { for (int i = 0; i < nb; i++) {
float amax = 0.0f; // absolute max float amax0 = 0.0f; // absolute max
float amax1 = 0.0f; // absolute max
for (int l = 0; l < QK4_0; l++) { for (int l = 0; l < QK4_0/2; l++) {
const float v = x[i*QK4_0 + l]; const float v = x[i*QK4_0 + l];
amax = MAX(amax, fabsf(v)); amax0 = MAX(amax0, fabsf(v));
} }
const float d = amax / ((1 << 3) - 1); for (int l = QK4_0/2; l < QK4_0; l++) {
const float id = d ? 1.0f/d : 0.0f; const float v = x[i*QK4_0 + l];
amax1 = MAX(amax1, fabsf(v));
}
y[i].d = d;
for (int l = 0; l < QK4_0; l += 2) { const float d0 = amax0 / ((1 << 3) - 1);
const float v0 = x[i*QK4_0 + l + 0]*id; const float d1 = amax1 / ((1 << 3) - 1);
const float v1 = x[i*QK4_0 + l + 1]*id;
const float id0 = d0 ? 1.0f/d0 : 0.0f;
const float id1 = d1 ? 1.0f/d1 : 0.0f;
y[i].d0 = GGML_FP32_TO_FP16(d0);
y[i].d1 = GGML_FP32_TO_FP16(d1);
for (int l = 0; l < QK4_0/2; l += 2) {
const float v0 = x[i*QK4_0 + l + 0]*id0;
const float v1 = x[i*QK4_0 + l + 1]*id0;
const uint8_t vi0 = (int8_t)roundf(v0) + 8; const uint8_t vi0 = (int8_t)roundf(v0) + 8;
const uint8_t vi1 = (int8_t)roundf(v1) + 8; const uint8_t vi1 = (int8_t)roundf(v1) + 8;
@ -623,15 +637,27 @@ static void quantize_row_q4_0_reference(const float * restrict x, block_q4_0 * r
assert(vi0 < 16); assert(vi0 < 16);
assert(vi1 < 16); assert(vi1 < 16);
pp[l/2] = vi0 | (vi1 << 4); y[i].qs[l/2] = vi0 | (vi1 << 4);
} }
memcpy(y[i].qs, pp, sizeof(pp)); for (int l = QK4_0/2; l < QK4_0; l += 2) {
const float v0 = x[i*QK4_0 + l + 0]*id1;
const float v1 = x[i*QK4_0 + l + 1]*id1;
const uint8_t vi0 = (int8_t)roundf(v0) + 8;
const uint8_t vi1 = (int8_t)roundf(v1) + 8;
assert(vi0 < 16);
assert(vi1 < 16);
y[i].qs[l/2] = vi0 | (vi1 << 4);
}
} }
} }
static void quantize_row_q4_0(const float * restrict x, void * restrict vy, int k) { static void quantize_row_q4_0(const float * restrict x, void * restrict vy, int k) {
assert(k % QK4_0 == 0); assert(k % QK4_0 == 0);
const int nb = k / QK4_0; const int nb = k / QK4_0;
block_q4_0 * restrict y = vy; block_q4_0 * restrict y = vy;
@ -678,24 +704,42 @@ static void quantize_row_q4_0(const float * restrict x, void * restrict vy, int
for (int i = 0; i < nb; i++) { for (int i = 0; i < nb; i++) {
float32x4_t srcv [8]; float32x4_t srcv [8];
float32x4_t asrcv[8]; float32x4_t asrcv[8];
float32x4_t amaxv[8];
float32x4_t amaxv0[4];
float32x4_t amaxv1[4];
for (int l = 0; l < 8; l++) srcv[l] = vld1q_f32(x + i*32 + 4*l); for (int l = 0; l < 8; l++) srcv[l] = vld1q_f32(x + i*32 + 4*l);
for (int l = 0; l < 8; l++) asrcv[l] = vabsq_f32(srcv[l]); for (int l = 0; l < 8; l++) asrcv[l] = vabsq_f32(srcv[l]);
for (int l = 0; l < 4; l++) amaxv[2*l] = vmaxq_f32(asrcv[2*l], asrcv[2*l+1]); for (int l = 0; l < 2; l++) amaxv0[2*l] = vmaxq_f32(asrcv [2*l], asrcv [2*l+1]);
for (int l = 0; l < 2; l++) amaxv[4*l] = vmaxq_f32(amaxv[4*l], amaxv[4*l+2]); for (int l = 0; l < 1; l++) amaxv0[4*l] = vmaxq_f32(amaxv0[4*l], amaxv0[4*l+2]);
for (int l = 0; l < 1; l++) amaxv[8*l] = vmaxq_f32(amaxv[8*l], amaxv[8*l+4]);
const float amax = vmaxvq_f32(amaxv[0]); for (int l = 0; l < 2; l++) amaxv1[2*l] = vmaxq_f32(asrcv [4+2*l], asrcv [4+2*l+1]);
for (int l = 0; l < 1; l++) amaxv1[4*l] = vmaxq_f32(amaxv1[4*l], amaxv1[4*l+2]);
const float d = amax / ((1 << 3) - 1); const float amax0 = vmaxvq_f32(amaxv0[0]);
const float id = d ? 1.0f/d : 0.0f; const float amax1 = vmaxvq_f32(amaxv1[0]);
y[i].d = d; const float d0 = amax0 / ((1 << 3) - 1);
const float d1 = amax1 / ((1 << 3) - 1);
for (int l = 0; l < 8; l++) { const float id0 = d0 ? 1.0f/d0 : 0.0f;
const float32x4_t v = vmulq_n_f32(srcv[l], id); const float id1 = d1 ? 1.0f/d1 : 0.0f;
y[i].d0 = d0;
y[i].d1 = d1;
for (int l = 0; l < 4; l++) {
const float32x4_t v = vmulq_n_f32(srcv[l], id0);
const float32x4_t vf = vaddq_f32(v, vdupq_n_f32(8.5f));
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);
}
for (int l = 4; l < 8; l++) {
const float32x4_t v = vmulq_n_f32(srcv[l], id1);
const float32x4_t vf = vaddq_f32(v, vdupq_n_f32(8.5f)); const float32x4_t vf = vaddq_f32(v, vdupq_n_f32(8.5f));
const int32x4_t vi = vcvtq_s32_f32(vf); const int32x4_t vi = vcvtq_s32_f32(vf);
@ -1237,70 +1281,94 @@ static void dequantize_row_q4_0(const void * restrict vx, float * restrict y, in
} }
#elif defined(__ARM_NEON) #elif defined(__ARM_NEON)
for (int i = 0; i < nb; i++) { for (int i = 0; i < nb; i++) {
const float32x4_t vd = vdupq_n_f32(x[i].d); const float32x4_t vd_0 = vdupq_n_f32(GGML_FP16_TO_FP32(x[i].d0));
const float32x4_t vd_1 = vdupq_n_f32(GGML_FP16_TO_FP32(x[i].d1));
const uint8_t * restrict pp = x[i].qs; const uint8_t * restrict pp = x[i].qs;
for (int l = 0; l < QK4_0; l += 16) { // Load 16x4-bit integers into 8x8-bit integers
// Load 16x4-bit integers into 8x8-bit integers const uint8x8_t v8_0 = vld1_u8(pp + 0);
const uint8x8_t v8 = vld1_u8(pp + l/2); const uint8x8_t v8_1 = vld1_u8(pp + 8);
// Expand 4-bit qs to 8-bit bytes // Expand 4-bit qs to 8-bit bytes
const uint8x8_t v0 = vand_u8(v8, vdup_n_u8(0x0f)); const uint8x8_t v0_0 = vand_u8 (v8_0, vdup_n_u8(0x0f));
const uint8x8_t v1 = vshr_n_u8(v8, 4); const uint8x8_t v1_0 = vshr_n_u8(v8_0, 4);
const uint8x8_t v0_1 = vand_u8 (v8_1, vdup_n_u8(0x0f));
const uint8x8_t v1_1 = vshr_n_u8(v8_1, 4);
// Convert to signed 8-bit integers // Convert to signed 8-bit integers
const int8x8_t vs_0 = vreinterpret_s8_u8(v0); const int8x8_t vs_0_0 = vreinterpret_s8_u8(v0_0);
const int8x8_t vs_1 = vreinterpret_s8_u8(v1); const int8x8_t vs_1_0 = vreinterpret_s8_u8(v1_0);
const int8x8_t vs_0_1 = vreinterpret_s8_u8(v0_1);
const int8x8_t vs_1_1 = vreinterpret_s8_u8(v1_1);
// Subtract 8 from each byte // Subtract 8 from each byte
const int8x8_t vb_0 = vsub_s8(vs_0, vdup_n_s8(8)); const int8x8_t vb_0_0 = vsub_s8(vs_0_0, vdup_n_s8(8));
const int8x8_t vb_1 = vsub_s8(vs_1, vdup_n_s8(8)); const int8x8_t vb_1_0 = vsub_s8(vs_1_0, vdup_n_s8(8));
const int8x8_t vb_0_1 = vsub_s8(vs_0_1, vdup_n_s8(8));
const int8x8_t vb_1_1 = vsub_s8(vs_1_1, vdup_n_s8(8));
// Interleave and combine // Interleave and combine
const int8x8_t vx_0 = vzip1_s8(vb_0, vb_1); const int8x8_t vx_0_0 = vzip1_s8(vb_0_0, vb_1_0);
const int8x8_t vx_1 = vzip2_s8(vb_0, vb_1); const int8x8_t vx_1_0 = vzip2_s8(vb_0_0, vb_1_0);
const int8x8_t vx_0_1 = vzip1_s8(vb_0_1, vb_1_1);
const int8x8_t vx_1_1 = vzip2_s8(vb_0_1, vb_1_1);
const int8x16_t vq = vcombine_s8(vx_0, vx_1); const int8x16_t vq_0 = vcombine_s8(vx_0_0, vx_1_0);
const int8x16_t vq_1 = vcombine_s8(vx_0_1, vx_1_1);
// convert to 2x int16x8_t // convert to 2x int16x8_t
const int16x8_t vi_0 = vmovl_s8(vget_low_s8 (vq)); const int16x8_t vi_0_0 = vmovl_s8(vget_low_s8 (vq_0));
const int16x8_t vi_1 = vmovl_s8(vget_high_s8(vq)); const int16x8_t vi_1_0 = vmovl_s8(vget_high_s8(vq_0));
const int16x8_t vi_0_1 = vmovl_s8(vget_low_s8 (vq_1));
const int16x8_t vi_1_1 = vmovl_s8(vget_high_s8(vq_1));
// convert to 4x float32x4_t // convert to 4x float32x4_t
const float32x4_t vf_0 = vcvtq_f32_s32(vmovl_s16(vget_low_s16 (vi_0))); const float32x4_t vf_0_0 = vcvtq_f32_s32(vmovl_s16(vget_low_s16 (vi_0_0)));
const float32x4_t vf_1 = vcvtq_f32_s32(vmovl_s16(vget_high_s16(vi_0))); const float32x4_t vf_1_0 = vcvtq_f32_s32(vmovl_s16(vget_high_s16(vi_0_0)));
const float32x4_t vf_2 = vcvtq_f32_s32(vmovl_s16(vget_low_s16 (vi_1))); const float32x4_t vf_2_0 = vcvtq_f32_s32(vmovl_s16(vget_low_s16 (vi_1_0)));
const float32x4_t vf_3 = vcvtq_f32_s32(vmovl_s16(vget_high_s16(vi_1))); const float32x4_t vf_3_0 = vcvtq_f32_s32(vmovl_s16(vget_high_s16(vi_1_0)));
const float32x4_t vf_0_1 = vcvtq_f32_s32(vmovl_s16(vget_low_s16 (vi_0_1)));
const float32x4_t vf_1_1 = vcvtq_f32_s32(vmovl_s16(vget_high_s16(vi_0_1)));
const float32x4_t vf_2_1 = vcvtq_f32_s32(vmovl_s16(vget_low_s16 (vi_1_1)));
const float32x4_t vf_3_1 = vcvtq_f32_s32(vmovl_s16(vget_high_s16(vi_1_1)));
// Multiply by d // Multiply by d
const float32x4_t r0 = vmulq_f32(vf_0, vd); const float32x4_t r0_0 = vmulq_f32(vf_0_0, vd_0);
const float32x4_t r1 = vmulq_f32(vf_1, vd); const float32x4_t r1_0 = vmulq_f32(vf_1_0, vd_0);
const float32x4_t r2 = vmulq_f32(vf_2, vd); const float32x4_t r2_0 = vmulq_f32(vf_2_0, vd_0);
const float32x4_t r3 = vmulq_f32(vf_3, vd); const float32x4_t r3_0 = vmulq_f32(vf_3_0, vd_0);
const float32x4_t r0_1 = vmulq_f32(vf_0_1, vd_1);
const float32x4_t r1_1 = vmulq_f32(vf_1_1, vd_1);
const float32x4_t r2_1 = vmulq_f32(vf_2_1, vd_1);
const float32x4_t r3_1 = vmulq_f32(vf_3_1, vd_1);
// Store // Store
vst1q_f32(y + i*QK4_0 + l + 0, r0); vst1q_f32(y + i*QK4_0 + 0, r0_0);
vst1q_f32(y + i*QK4_0 + l + 4, r1); vst1q_f32(y + i*QK4_0 + 4, r1_0);
vst1q_f32(y + i*QK4_0 + l + 8, r2); vst1q_f32(y + i*QK4_0 + 8, r2_0);
vst1q_f32(y + i*QK4_0 + l + 12, r3); vst1q_f32(y + i*QK4_0 + 12, r3_0);
} vst1q_f32(y + i*QK4_0 + 16, r0_1);
vst1q_f32(y + i*QK4_0 + 20, r1_1);
vst1q_f32(y + i*QK4_0 + 24, r2_1);
vst1q_f32(y + i*QK4_0 + 28, r3_1);
} }
#else #else
// scalar // scalar
for (int i = 0; i < nb; i++) { for (int i = 0; i < nb; i++) {
const float d = x[i].d; const float d0 = GGML_FP16_TO_FP32(x[i].d0);
const float d1 = GGML_FP16_TO_FP32(x[i].d1);
const uint8_t * restrict pp = x[i].qs; const uint8_t * restrict pp = x[i].qs;
for (int l = 0; l < QK4_0; l += 2) { for (int l = 0; l < QK4_0/2; l += 2) {
const uint8_t vi = pp[l/2]; const uint8_t vi = pp[l/2];
const int8_t vi0 = vi & 0xf; const int8_t vi0 = vi & 0xf;
const int8_t vi1 = vi >> 4; const int8_t vi1 = vi >> 4;
const float v0 = (vi0 - 8)*d; const float v0 = (vi0 - 8)*d0;
const float v1 = (vi1 - 8)*d; const float v1 = (vi1 - 8)*d0;
//printf("d = %f, vi = %d, vi0 = %d, vi1 = %d, v0 = %f, v1 = %f\n", d, vi, vi0, vi1, v0, v1); //printf("d = %f, vi = %d, vi0 = %d, vi1 = %d, v0 = %f, v1 = %f\n", d, vi, vi0, vi1, v0, v1);
@ -1310,6 +1378,22 @@ static void dequantize_row_q4_0(const void * restrict vx, float * restrict y, in
assert(!isnan(y[i*QK4_0 + l + 0])); assert(!isnan(y[i*QK4_0 + l + 0]));
assert(!isnan(y[i*QK4_0 + l + 1])); assert(!isnan(y[i*QK4_0 + l + 1]));
} }
for (int l = QK4_0/2; l < QK4_0; l += 2) {
const uint8_t vi = pp[l/2];
const int8_t vi0 = vi & 0xf;
const int8_t vi1 = vi >> 4;
const float v0 = (vi0 - 8)*d1;
const float v1 = (vi1 - 8)*d1;
y[i*QK4_0 + l + 0] = v0;
y[i*QK4_0 + l + 1] = v1;
assert(!isnan(y[i*QK4_0 + l + 0]));
assert(!isnan(y[i*QK4_0 + l + 1]));
}
} }
#endif #endif
} }
@ -2250,14 +2334,19 @@ static void ggml_vec_dot_q4_0(const int n, float * restrict s, const void * rest
#if defined(__ARM_FEATURE_DOTPROD) #if defined(__ARM_FEATURE_DOTPROD)
// dot product into int32x4_t // dot product into int32x4_t
int32x4_t p_0 = vdotq_s32(vdupq_n_s32(0), v0_0ls, v1_0ls); //int32x4_t p_0 = vdotq_s32(vdupq_n_s32(0), v0_0ls, v1_0ls);
int32x4_t p_1 = vdotq_s32(vdupq_n_s32(0), v0_1ls, v1_1ls); //int32x4_t p_1 = vdotq_s32(vdupq_n_s32(0), v0_1ls, v1_1ls);
p_0 = vdotq_s32(p_0, v0_0hs, v1_0hs); //p_0 = vdotq_s32(p_0, v0_0hs, v1_0hs);
p_1 = vdotq_s32(p_1, v0_1hs, v1_1hs); //p_1 = vdotq_s32(p_1, v0_1hs, v1_1hs);
sum0 += x0->d*y0->d*vaddvq_s32(p_0); //sum0 += x0->d*y0->d*vaddvq_s32(p_0);
sum1 += x1->d*y1->d*vaddvq_s32(p_1); //sum1 += x1->d*y1->d*vaddvq_s32(p_1);
sum0 += (GGML_FP16_TO_FP32(x0->d0)*GGML_FP16_TO_FP32(y0->d0))*vaddvq_s32(vdotq_s32(vdupq_n_s32(0), v0_0ls, v1_0ls));
sum0 += (GGML_FP16_TO_FP32(x0->d1)*GGML_FP16_TO_FP32(y0->d1))*vaddvq_s32(vdotq_s32(vdupq_n_s32(0), v0_0hs, v1_0hs));
sum1 += (GGML_FP16_TO_FP32(x1->d0)*GGML_FP16_TO_FP32(y1->d0))*vaddvq_s32(vdotq_s32(vdupq_n_s32(0), v0_1ls, v1_1ls));
sum1 += (GGML_FP16_TO_FP32(x1->d1)*GGML_FP16_TO_FP32(y1->d1))*vaddvq_s32(vdotq_s32(vdupq_n_s32(0), v0_1hs, v1_1hs));
#else #else
const int16x8_t pl0l = vmull_s8(vget_low_s8 (v0_0ls), vget_low_s8 (v1_0ls)); const int16x8_t pl0l = vmull_s8(vget_low_s8 (v0_0ls), vget_low_s8 (v1_0ls));
const int16x8_t pl0h = vmull_s8(vget_high_s8(v0_0ls), vget_high_s8(v1_0ls)); const int16x8_t pl0h = vmull_s8(vget_high_s8(v0_0ls), vget_high_s8(v1_0ls));
@ -2517,14 +2606,13 @@ static void ggml_vec_dot_q4_0(const int n, float * restrict s, const void * rest
#else #else
// scalar // scalar
for (int i = 0; i < nb; i++) { 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 uint8_t * restrict p0 = x[i].qs;
const uint8_t * restrict p1 = y[i].qs; const uint8_t * restrict p1 = y[i].qs;
int sumi = 0; int sumi_0 = 0;
for (int j = 0; j < QK4_0/2; j++) { int sumi_1 = 0;
for (int j = 0; j < QK4_0/4; j++) {
const uint8_t v0 = p0[j]; const uint8_t v0 = p0[j];
const uint8_t v1 = p1[j]; const uint8_t v1 = p1[j];
@ -2534,9 +2622,24 @@ static void ggml_vec_dot_q4_0(const int n, float * restrict s, const void * rest
const int i2 = (v1 & 0xf) - 8; const int i2 = (v1 & 0xf) - 8;
const int i3 = (v1 >> 4) - 8; const int i3 = (v1 >> 4) - 8;
sumi += i0*i2 + i1*i3; sumi_0 += i0*i2 + i1*i3;
} }
sumf += d0 * d1 * sumi;
for (int j = QK4_0/4; j < QK4_0/2; j++) {
const uint8_t v0 = p0[j];
const uint8_t v1 = p1[j];
const int i0 = (v0 & 0xf) - 8;
const int i1 = (v0 >> 4) - 8;
const int i2 = (v1 & 0xf) - 8;
const int i3 = (v1 >> 4) - 8;
sumi_1 += i0*i2 + i1*i3;
}
sumf += (GGML_FP16_TO_FP32(x[i].d0) * GGML_FP16_TO_FP32(y[i].d0)) * sumi_0;
sumf += (GGML_FP16_TO_FP32(x[i].d1) * GGML_FP16_TO_FP32(y[i].d1)) * sumi_1;
} }
#endif #endif
@ -2765,6 +2868,12 @@ static void ggml_vec_dot_q4_0_q8_0(const int n, float * restrict s, const void *
const int8x16_t v0_1ls = vsubq_s8(v0_1l, s8b); const int8x16_t v0_1ls = vsubq_s8(v0_1l, s8b);
const int8x16_t v0_1hs = vsubq_s8(v0_1h, s8b); const int8x16_t v0_1hs = vsubq_s8(v0_1h, s8b);
// interleave
const int8x16_t v0_0lz = vzip1q_s8(v0_0ls, v0_0hs);
const int8x16_t v0_0hz = vzip2q_s8(v0_0ls, v0_0hs);
const int8x16_t v0_1lz = vzip1q_s8(v0_1ls, v0_1hs);
const int8x16_t v0_1hz = vzip2q_s8(v0_1ls, v0_1hs);
// load y // load y
const int8x16_t v1_0l = vld1q_s8(y0->qs); const int8x16_t v1_0l = vld1q_s8(y0->qs);
const int8x16_t v1_0h = vld1q_s8(y0->qs + 16); const int8x16_t v1_0h = vld1q_s8(y0->qs + 16);
@ -2772,21 +2881,26 @@ static void ggml_vec_dot_q4_0_q8_0(const int n, float * restrict s, const void *
const int8x16_t v1_1h = vld1q_s8(y1->qs + 16); const int8x16_t v1_1h = vld1q_s8(y1->qs + 16);
// interleave // interleave
const int8x16_t v1_0ls = vuzp1q_s8(v1_0l, v1_0h); //const int8x16_t v1_0ls = vuzp1q_s8(v1_0l, v1_0h);
const int8x16_t v1_0hs = vuzp2q_s8(v1_0l, v1_0h); //const int8x16_t v1_0hs = vuzp2q_s8(v1_0l, v1_0h);
const int8x16_t v1_1ls = vuzp1q_s8(v1_1l, v1_1h); //const int8x16_t v1_1ls = vuzp1q_s8(v1_1l, v1_1h);
const int8x16_t v1_1hs = vuzp2q_s8(v1_1l, v1_1h); //const int8x16_t v1_1hs = vuzp2q_s8(v1_1l, v1_1h);
#if defined(__ARM_FEATURE_DOTPROD) #if defined(__ARM_FEATURE_DOTPROD)
// dot product into int32x4_t // dot product into int32x4_t
int32x4_t p_0 = vdotq_s32(vdupq_n_s32(0), v0_0ls, v1_0ls); //int32x4_t p_0 = vdotq_s32(vdupq_n_s32(0), v0_0ls, v1_0ls);
int32x4_t p_1 = vdotq_s32(vdupq_n_s32(0), v0_1ls, v1_1ls); //int32x4_t p_1 = vdotq_s32(vdupq_n_s32(0), v0_1ls, v1_1ls);
p_0 = vdotq_s32(p_0, v0_0hs, v1_0hs); //p_0 = vdotq_s32(p_0, v0_0hs, v1_0hs);
p_1 = vdotq_s32(p_1, v0_1hs, v1_1hs); //p_1 = vdotq_s32(p_1, v0_1hs, v1_1hs);
sum0 += x0->d*y0->d*vaddvq_s32(p_0); //sum0 += x0->d*y0->d*vaddvq_s32(p_0);
sum1 += x1->d*y1->d*vaddvq_s32(p_1); //sum1 += x1->d*y1->d*vaddvq_s32(p_1);
sum0 += (GGML_FP16_TO_FP32(x0->d0)*y0->d)*vaddvq_s32(vdotq_s32(vdupq_n_s32(0), v0_0lz, v1_0l));
sum0 += (GGML_FP16_TO_FP32(x0->d1)*y0->d)*vaddvq_s32(vdotq_s32(vdupq_n_s32(0), v0_0hz, v1_0h));
sum1 += (GGML_FP16_TO_FP32(x1->d0)*y1->d)*vaddvq_s32(vdotq_s32(vdupq_n_s32(0), v0_1lz, v1_1l));
sum1 += (GGML_FP16_TO_FP32(x1->d1)*y1->d)*vaddvq_s32(vdotq_s32(vdupq_n_s32(0), v0_1hz, v1_1h));
#else #else
const int16x8_t pl0l = vmull_s8(vget_low_s8 (v0_0ls), vget_low_s8 (v1_0ls)); const int16x8_t pl0l = vmull_s8(vget_low_s8 (v0_0ls), vget_low_s8 (v1_0ls));
const int16x8_t pl0h = vmull_s8(vget_high_s8(v0_0ls), vget_high_s8(v1_0ls)); const int16x8_t pl0h = vmull_s8(vget_high_s8(v0_0ls), vget_high_s8(v1_0ls));
@ -2904,14 +3018,13 @@ static void ggml_vec_dot_q4_0_q8_0(const int n, float * restrict s, const void *
#else #else
// scalar // scalar
for (int i = 0; i < nb; i++) { 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 uint8_t * restrict p0 = x[i].qs;
const int8_t * restrict p1 = y[i].qs; const int8_t * restrict p1 = y[i].qs;
int sumi = 0; int sumi_0 = 0;
for (int j = 0; j < QK8_0/2; j++) { int sumi_1 = 0;
for (int j = 0; j < QK8_0/4; j++) {
const uint8_t v0 = p0[j]; const uint8_t v0 = p0[j];
const int i0 = (int8_t) (v0 & 0xf) - 8; const int i0 = (int8_t) (v0 & 0xf) - 8;
@ -2920,9 +3033,23 @@ static void ggml_vec_dot_q4_0_q8_0(const int n, float * restrict s, const void *
const int i2 = p1[2*j + 0]; const int i2 = p1[2*j + 0];
const int i3 = p1[2*j + 1]; const int i3 = p1[2*j + 1];
sumi += i0*i2 + i1*i3; sumi_0 += i0*i2 + i1*i3;
} }
sumf += d0*d1*sumi;
for (int j = QK8_0/4; j < QK8_0/2; j++) {
const uint8_t v0 = p0[j];
const int i0 = (int8_t) (v0 & 0xf) - 8;
const int i1 = (int8_t) (v0 >> 4) - 8;
const int i2 = p1[2*j + 0];
const int i3 = p1[2*j + 1];
sumi_1 += i0*i2 + i1*i3;
}
sumf += (GGML_FP16_TO_FP32(x[i].d0) * y[i].d) * sumi_0;
sumf += (GGML_FP16_TO_FP32(x[i].d1) * y[i].d) * sumi_1;
} }
#endif #endif