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

Adding Q5_K - scalar, AVX2, CUDA

Browse source 
Performance is ~20% lower compared to Q4_K on the CPU.
This is to be expected, considering that we are memory bound
on the CPU and the 5-bit model is ~22% larger than the 4-bit.
On the GPU, single token prediction is about the same as Q4_0
for both, single token and batch prediction.
This commit is contained in:
Iwan Kawrakow 2023-05-29 18:57:04 +03:00
parent cf221afb55
commit b835d0f49f
8 changed files with 440 additions and 12 deletions
Download patch file Download diff file Expand all files Collapse all files

6
Makefile
View file

@ -40,8 +40,10 @@ endif
# #
# keep standard at C11 and C++11 # keep standard at C11 and C++11
CFLAGS = -I. -O3 -std=c11 -fPIC #OPT = -Ofast
CXXFLAGS = -I. -I./examples -O3 -std=c++11 -fPIC OPT = -O3
CFLAGS = -I. $(OPT) -std=c11 -fPIC
CXXFLAGS = -I. -I./examples $(OPT) -std=c++11 -fPIC
LDFLAGS = LDFLAGS =
ifdef LLAMA_DEBUG ifdef LLAMA_DEBUG

1
examples/quantize/quantize.cpp
View file

@ -14,6 +14,7 @@ static const std::map<std::string, llama_ftype> LLAMA_FTYPE_MAP = {
{"q8_0", LLAMA_FTYPE_MOSTLY_Q8_0}, {"q8_0", LLAMA_FTYPE_MOSTLY_Q8_0},
{"q3_K", LLAMA_FTYPE_MOSTLY_Q3_K}, {"q3_K", LLAMA_FTYPE_MOSTLY_Q3_K},
{"q4_K", LLAMA_FTYPE_MOSTLY_Q4_K}, {"q4_K", LLAMA_FTYPE_MOSTLY_Q4_K},
{"q5_K", LLAMA_FTYPE_MOSTLY_Q5_K},
{"q6_K", LLAMA_FTYPE_MOSTLY_Q6_K}, {"q6_K", LLAMA_FTYPE_MOSTLY_Q6_K},
}; };

94
ggml-cuda.cu
View file

@ -105,6 +105,15 @@ typedef struct {
} block_q4_K; } block_q4_K;
static_assert(sizeof(block_q4_K) == 2*sizeof(ggml_fp16_t) + 3*QK_K/64 + QK_K/2, "wrong q4_K block size/padding"); static_assert(sizeof(block_q4_K) == 2*sizeof(ggml_fp16_t) + 3*QK_K/64 + QK_K/2, "wrong q4_K block size/padding");
typedef struct {
half d; // super-block scale for quantized scales
half dmin; // super-block scale for quantized mins
uint8_t scales[3*QK_K/64]; // scales, quantized with 6 bits
uint8_t qh[QK_K/8]; // quants, high bit
uint8_t qs[QK_K/2]; // quants, low 4 bits
} block_q5_K;
static_assert(sizeof(block_q5_K) == 2*sizeof(ggml_fp16_t) + 3*QK_K/64 + QK_K/2 + QK_K/8, "wrong q5_K block size/padding");
typedef struct { typedef struct {
uint8_t ql[QK_K/2]; // quants, lower 4 bits uint8_t ql[QK_K/2]; // quants, lower 4 bits
uint8_t qh[QK_K/4]; // quants, upper 2 bits uint8_t qh[QK_K/4]; // quants, upper 2 bits
@ -355,6 +364,76 @@ static __device__ void vec_dot_q4_K(const void * vx, const int ib, const int iqs
} }
static __global__ void dequantize_block_q5_K(const void * vx, float * yy) {
const block_q5_K * x = (const block_q5_K *) vx;
const int i = blockIdx.x;
// assume 64 threads - this is very slightly better than the one below
const int tid = threadIdx.x;
const int il = tid/16; // il is in 0...3
const int ir = tid%16; // ir is in 0...15
const int is = 2*il; // is is in 0...6
float * y = yy + i*QK_K + 64*il + 2*ir;
const float dall = x[i].d;
const float dmin = x[i].dmin;
const uint8_t * ql = x[i].qs + 32*il + 2*ir;
const uint8_t * qh = x[i].qh + 2*ir;
uint8_t sc, m;
get_scale_min_k4(is + 0, x[i].scales, sc, m);
const float d1 = dall * sc; const float m1 = dmin * m;
get_scale_min_k4(is + 1, x[i].scales, sc, m);
const float d2 = dall * sc; const float m2 = dmin * m;
uint8_t hm = 1 << (2*il);
y[ 0] = d1 * ((ql[ 0] & 0xF) + (qh[ 0] & hm ? 16 : 0)) - m1;
y[ 1] = d1 * ((ql[ 1] & 0xF) + (qh[ 1] & hm ? 16 : 0)) - m1;
hm <<= 1;
y[32] = d2 * ((ql[ 0] >> 4) + (qh[ 0] & hm ? 16 : 0)) - m2;
y[33] = d2 * ((ql[ 1] >> 4) + (qh[ 1] & hm ? 16 : 0)) - m2;
}
static __device__ void vec_dot_q5_K(const void * vx, const int ib, const int iqs, const float * yy, float & result) {
const block_q5_K * x = (const block_q5_K *) vx;
// iqs is in 0...248 in steps of 8 =>
const int j = iqs / 64; // j is in 0...3
const int ir = (iqs - 64*j)/2; // ir is in 0...28 in steps of 4
const int is = 2*j; // is is in 0...6 in steps of 2
const float * y = yy + 64*j + ir;
const uint8_t * ql = x[ib].qs + 32*j + ir;
const uint8_t * qh = x[ib].qh + ir;
const float dall = x[ib].d;
const float dmin = x[ib].dmin;
uint8_t sc, m;
get_scale_min_k4(is + 0, x[ib].scales, sc, m);
const float d1 = dall * sc;
const float m1 = dmin * m;
get_scale_min_k4(is + 1, x[ib].scales, sc, m);
const float d2 = dall * sc;
const float m2 = dmin * m;
uint8_t hm = 1 << is;
float sum = 0;
for (int k = 0; k < 4; ++k) {
sum += y[k + 0] * (d1 * ((ql[k] & 0xF) + (qh[k] & hm ? 16 : 0)) - m1);
}
hm <<= 1;
for (int k = 0; k < 4; ++k) {
sum += y[k + 32] * (d2 * ((ql[k] >> 4) + (qh[k] & hm ? 16 : 0)) - m2);
}
result = sum;
}
static __global__ void dequantize_block_q6_K(const void * vx, float * yy) { static __global__ void dequantize_block_q6_K(const void * vx, float * yy) {
const block_q6_K * x = (const block_q6_K *) vx; const block_q6_K * x = (const block_q6_K *) vx;
@ -556,6 +635,11 @@ static void dequantize_row_q4_K_cuda(const void * vx, float * y, const int k, cu
dequantize_block_q4_K<<<nb, 32, 0, stream>>>(vx, y); dequantize_block_q4_K<<<nb, 32, 0, stream>>>(vx, y);
} }
static void dequantize_row_q5_K_cuda(const void * vx, float * y, const int k, cudaStream_t stream) {
const int nb = k / QK_K;
dequantize_block_q5_K<<<nb, 64, 0, stream>>>(vx, y);
}
static void dequantize_row_q6_K_cuda(const void * vx, float * y, const int k, cudaStream_t stream) { static void dequantize_row_q6_K_cuda(const void * vx, float * y, const int k, cudaStream_t stream) {
const int nb = k / QK_K; const int nb = k / QK_K;
dequantize_block_q6_K<<<nb, 64, 0, stream>>>(vx, y); dequantize_block_q6_K<<<nb, 64, 0, stream>>>(vx, y);
@ -613,6 +697,12 @@ static void dequantize_mul_mat_vec_q4_K_cuda(const void * vx, const float * y, f
dequantize_mul_mat_vec_k<32, vec_dot_q4_K><<<nrows/2, block_dims, 0, stream>>>(vx, y, dst, ncols); dequantize_mul_mat_vec_k<32, vec_dot_q4_K><<<nrows/2, block_dims, 0, stream>>>(vx, y, dst, ncols);
} }
static void dequantize_mul_mat_vec_q5_K_cuda(const void * vx, const float * y, float * dst, const int ncols, const int nrows, cudaStream_t stream) {
GGML_ASSERT(ncols % QK_K == 0);
const dim3 block_dims(32, 2, 1);
dequantize_mul_mat_vec_k<32, vec_dot_q5_K><<<nrows/2, block_dims, 0, stream>>>(vx, y, dst, ncols);
}
static void dequantize_mul_mat_vec_q6_K_cuda(const void * vx, const float * y, float * dst, const int ncols, const int nrows, cudaStream_t stream) { static void dequantize_mul_mat_vec_q6_K_cuda(const void * vx, const float * y, float * dst, const int ncols, const int nrows, cudaStream_t stream) {
GGML_ASSERT(ncols % QK_K == 0); GGML_ASSERT(ncols % QK_K == 0);
const dim3 block_dims(32, 2, 1); const dim3 block_dims(32, 2, 1);
@ -648,6 +738,8 @@ static to_fp32_cuda_t ggml_get_to_fp32_cuda(ggml_type type) {
return dequantize_row_q3_K_cuda; return dequantize_row_q3_K_cuda;
case GGML_TYPE_Q4_K: case GGML_TYPE_Q4_K:
return dequantize_row_q4_K_cuda; return dequantize_row_q4_K_cuda;
case GGML_TYPE_Q5_K:
return dequantize_row_q5_K_cuda;
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
return dequantize_row_q6_K_cuda; return dequantize_row_q6_K_cuda;
case GGML_TYPE_F16: case GGML_TYPE_F16:
@ -673,6 +765,8 @@ static dequantize_mul_mat_vec_cuda_t ggml_get_dequantize_mul_mat_vec_cuda(ggml_t
return dequantize_mul_mat_vec_q3_K_cuda; return dequantize_mul_mat_vec_q3_K_cuda;
case GGML_TYPE_Q4_K: case GGML_TYPE_Q4_K:
return dequantize_mul_mat_vec_q4_K_cuda; return dequantize_mul_mat_vec_q4_K_cuda;
case GGML_TYPE_Q5_K:
return dequantize_mul_mat_vec_q5_K_cuda;
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
return dequantize_mul_mat_vec_q6_K_cuda; return dequantize_mul_mat_vec_q6_K_cuda;
case GGML_TYPE_F16: case GGML_TYPE_F16:

35
ggml.c
View file

@ -1582,6 +1582,14 @@ static const quantize_fns_t quantize_fns[GGML_TYPE_COUNT] = {
.vec_dot_q = ggml_vec_dot_q4_K_q8_K, .vec_dot_q = ggml_vec_dot_q4_K_q8_K,
.vec_dot_type = GGML_TYPE_Q8_K, .vec_dot_type = GGML_TYPE_Q8_K,
}, },
[GGML_TYPE_Q5_K] = {
.dequantize_row_q = (dequantize_row_q_t) dequantize_row_q5_K,
.quantize_row_q = quantize_row_q5_K,
.quantize_row_q_reference = (quantize_row_q_t) quantize_row_q5_K_reference,
.quantize_row_q_dot = quantize_row_q8_K,
.vec_dot_q = ggml_vec_dot_q5_K_q8_K,
.vec_dot_type = GGML_TYPE_Q8_K,
},
[GGML_TYPE_Q6_K] = { [GGML_TYPE_Q6_K] = {
.dequantize_row_q = (dequantize_row_q_t) dequantize_row_q6_K, .dequantize_row_q = (dequantize_row_q_t) dequantize_row_q6_K,
.quantize_row_q = quantize_row_q6_K, .quantize_row_q = quantize_row_q6_K,
@ -3471,13 +3479,14 @@ static const int GGML_BLCK_SIZE[GGML_TYPE_COUNT] = {
[GGML_TYPE_Q8_1] = QK8_1, [GGML_TYPE_Q8_1] = QK8_1,
[GGML_TYPE_Q3_K] = QK_K, [GGML_TYPE_Q3_K] = QK_K,
[GGML_TYPE_Q4_K] = QK_K, [GGML_TYPE_Q4_K] = QK_K,
[GGML_TYPE_Q5_K] = QK_K,
[GGML_TYPE_Q6_K] = QK_K, [GGML_TYPE_Q6_K] = QK_K,
[GGML_TYPE_Q8_K] = QK_K, [GGML_TYPE_Q8_K] = QK_K,
[GGML_TYPE_I8] = 1, [GGML_TYPE_I8] = 1,
[GGML_TYPE_I16] = 1, [GGML_TYPE_I16] = 1,
[GGML_TYPE_I32] = 1, [GGML_TYPE_I32] = 1,
}; };
static_assert(GGML_TYPE_COUNT == 17, "GGML_BLCK_SIZE is outdated"); static_assert(GGML_TYPE_COUNT == 18, "GGML_BLCK_SIZE is outdated");
static const size_t GGML_TYPE_SIZE[GGML_TYPE_COUNT] = { static const size_t GGML_TYPE_SIZE[GGML_TYPE_COUNT] = {
[GGML_TYPE_F32] = sizeof(float), [GGML_TYPE_F32] = sizeof(float),
@ -3489,13 +3498,14 @@ static const size_t GGML_TYPE_SIZE[GGML_TYPE_COUNT] = {
[GGML_TYPE_Q8_0] = sizeof(block_q8_0), [GGML_TYPE_Q8_0] = sizeof(block_q8_0),
[GGML_TYPE_Q8_1] = sizeof(block_q8_1), [GGML_TYPE_Q8_1] = sizeof(block_q8_1),
[GGML_TYPE_Q4_K] = sizeof(block_q4_K), [GGML_TYPE_Q4_K] = sizeof(block_q4_K),
[GGML_TYPE_Q5_K] = sizeof(block_q5_K),
[GGML_TYPE_Q6_K] = sizeof(block_q6_K), [GGML_TYPE_Q6_K] = sizeof(block_q6_K),
[GGML_TYPE_Q8_K] = sizeof(block_q8_K), [GGML_TYPE_Q8_K] = sizeof(block_q8_K),
[GGML_TYPE_I8] = sizeof(int8_t), [GGML_TYPE_I8] = sizeof(int8_t),
[GGML_TYPE_I16] = sizeof(int16_t), [GGML_TYPE_I16] = sizeof(int16_t),
[GGML_TYPE_I32] = sizeof(int32_t), [GGML_TYPE_I32] = sizeof(int32_t),
}; };
static_assert(GGML_TYPE_COUNT == 17, "GGML_TYPE_SIZE is outdated"); static_assert(GGML_TYPE_COUNT == 18, "GGML_TYPE_SIZE is outdated");
static const char * GGML_TYPE_NAME[GGML_TYPE_COUNT] = { static const char * GGML_TYPE_NAME[GGML_TYPE_COUNT] = {
@ -3509,13 +3519,14 @@ static const char * GGML_TYPE_NAME[GGML_TYPE_COUNT] = {
[GGML_TYPE_Q8_1] = "q8_1", [GGML_TYPE_Q8_1] = "q8_1",
[GGML_TYPE_Q3_K] = "q3_K", [GGML_TYPE_Q3_K] = "q3_K",
[GGML_TYPE_Q4_K] = "q4_K", [GGML_TYPE_Q4_K] = "q4_K",
[GGML_TYPE_Q5_K] = "q5_K",
[GGML_TYPE_Q6_K] = "q6_K", [GGML_TYPE_Q6_K] = "q6_K",
[GGML_TYPE_Q8_K] = "q8_K", [GGML_TYPE_Q8_K] = "q8_K",
[GGML_TYPE_I8] = "i8", [GGML_TYPE_I8] = "i8",
[GGML_TYPE_I16] = "i16", [GGML_TYPE_I16] = "i16",
[GGML_TYPE_I32] = "i32", [GGML_TYPE_I32] = "i32",
}; };
static_assert(GGML_TYPE_COUNT == 17, "GGML_TYPE_NAME is outdated"); static_assert(GGML_TYPE_COUNT == 18, "GGML_TYPE_NAME is outdated");
static bool GGML_IS_QUANTIZED[GGML_TYPE_COUNT] = { static bool GGML_IS_QUANTIZED[GGML_TYPE_COUNT] = {
[GGML_TYPE_F32] = false, [GGML_TYPE_F32] = false,
@ -3527,13 +3538,14 @@ static bool GGML_IS_QUANTIZED[GGML_TYPE_COUNT] = {
[GGML_TYPE_Q8_0] = true, [GGML_TYPE_Q8_0] = true,
[GGML_TYPE_Q8_1] = true, [GGML_TYPE_Q8_1] = true,
[GGML_TYPE_Q4_K] = true, [GGML_TYPE_Q4_K] = true,
[GGML_TYPE_Q5_K] = true,
[GGML_TYPE_Q6_K] = true, [GGML_TYPE_Q6_K] = true,
[GGML_TYPE_Q8_K] = true, [GGML_TYPE_Q8_K] = true,
[GGML_TYPE_I8] = false, [GGML_TYPE_I8] = false,
[GGML_TYPE_I16] = false, [GGML_TYPE_I16] = false,
[GGML_TYPE_I32] = false, [GGML_TYPE_I32] = false,
}; };
static_assert(GGML_TYPE_COUNT == 17, "GGML_IS_QUANTIZED is outdated"); static_assert(GGML_TYPE_COUNT == 18, "GGML_IS_QUANTIZED is outdated");
static const char * GGML_OP_NAME[GGML_OP_COUNT] = { static const char * GGML_OP_NAME[GGML_OP_COUNT] = {
"NONE", "NONE",
@ -3842,6 +3854,7 @@ enum ggml_type ggml_ftype_to_ggml_type(enum ggml_ftype ftype) {
case GGML_FTYPE_MOSTLY_Q8_0: wtype = GGML_TYPE_Q8_0; break; case GGML_FTYPE_MOSTLY_Q8_0: wtype = GGML_TYPE_Q8_0; break;
case GGML_FTYPE_MOSTLY_Q3_K: wtype = GGML_TYPE_Q3_K; break; case GGML_FTYPE_MOSTLY_Q3_K: wtype = GGML_TYPE_Q3_K; break;
case GGML_FTYPE_MOSTLY_Q4_K: wtype = GGML_TYPE_Q4_K; break; case GGML_FTYPE_MOSTLY_Q4_K: wtype = GGML_TYPE_Q4_K; break;
case GGML_FTYPE_MOSTLY_Q5_K: wtype = GGML_TYPE_Q5_K; break;
case GGML_FTYPE_MOSTLY_Q6_K: wtype = GGML_TYPE_Q6_K; break; case GGML_FTYPE_MOSTLY_Q6_K: wtype = GGML_TYPE_Q6_K; break;
case GGML_FTYPE_UNKNOWN: wtype = GGML_TYPE_COUNT; break; case GGML_FTYPE_UNKNOWN: wtype = GGML_TYPE_COUNT; break;
case GGML_FTYPE_MOSTLY_Q4_1_SOME_F16: wtype = GGML_TYPE_COUNT; break; case GGML_FTYPE_MOSTLY_Q4_1_SOME_F16: wtype = GGML_TYPE_COUNT; break;
@ -7628,6 +7641,7 @@ static void ggml_compute_forward_add(
case GGML_TYPE_Q8_0: case GGML_TYPE_Q8_0:
case GGML_TYPE_Q3_K: case GGML_TYPE_Q3_K:
case GGML_TYPE_Q4_K: case GGML_TYPE_Q4_K:
case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
{ {
ggml_compute_forward_add_q_f32(params, src0, src1, dst); ggml_compute_forward_add_q_f32(params, src0, src1, dst);
@ -7934,6 +7948,7 @@ static void ggml_compute_forward_add1(
case GGML_TYPE_Q8_1: case GGML_TYPE_Q8_1:
case GGML_TYPE_Q3_K: case GGML_TYPE_Q3_K:
case GGML_TYPE_Q4_K: case GGML_TYPE_Q4_K:
case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
{ {
ggml_compute_forward_add1_q_f32(params, src0, src1, dst); ggml_compute_forward_add1_q_f32(params, src0, src1, dst);
@ -8059,6 +8074,7 @@ static void ggml_compute_forward_acc(
case GGML_TYPE_Q8_1: case GGML_TYPE_Q8_1:
case GGML_TYPE_Q3_K: case GGML_TYPE_Q3_K:
case GGML_TYPE_Q4_K: case GGML_TYPE_Q4_K:
case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
default: default:
{ {
@ -10155,6 +10171,7 @@ static void ggml_compute_forward_mul_mat(
case GGML_TYPE_Q8_1: case GGML_TYPE_Q8_1:
case GGML_TYPE_Q3_K: case GGML_TYPE_Q3_K:
case GGML_TYPE_Q4_K: case GGML_TYPE_Q4_K:
case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
{ {
ggml_compute_forward_mul_mat_q_f32(params, src0, src1, dst); ggml_compute_forward_mul_mat_q_f32(params, src0, src1, dst);
@ -10341,6 +10358,7 @@ static void ggml_compute_forward_set(
case GGML_TYPE_Q8_1: case GGML_TYPE_Q8_1:
case GGML_TYPE_Q3_K: case GGML_TYPE_Q3_K:
case GGML_TYPE_Q4_K: case GGML_TYPE_Q4_K:
case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
default: default:
{ {
@ -10509,6 +10527,7 @@ static void ggml_compute_forward_get_rows(
case GGML_TYPE_Q8_1: case GGML_TYPE_Q8_1:
case GGML_TYPE_Q3_K: case GGML_TYPE_Q3_K:
case GGML_TYPE_Q4_K: case GGML_TYPE_Q4_K:
case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
{ {
ggml_compute_forward_get_rows_q(params, src0, src1, dst); ggml_compute_forward_get_rows_q(params, src0, src1, dst);
@ -11058,6 +11077,7 @@ static void ggml_compute_forward_alibi(
case GGML_TYPE_Q8_1: case GGML_TYPE_Q8_1:
case GGML_TYPE_Q3_K: case GGML_TYPE_Q3_K:
case GGML_TYPE_Q4_K: case GGML_TYPE_Q4_K:
case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
case GGML_TYPE_Q8_K: case GGML_TYPE_Q8_K:
case GGML_TYPE_I8: case GGML_TYPE_I8:
@ -11133,6 +11153,7 @@ static void ggml_compute_forward_clamp(
case GGML_TYPE_Q8_1: case GGML_TYPE_Q8_1:
case GGML_TYPE_Q3_K: case GGML_TYPE_Q3_K:
case GGML_TYPE_Q4_K: case GGML_TYPE_Q4_K:
case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
case GGML_TYPE_Q8_K: case GGML_TYPE_Q8_K:
case GGML_TYPE_I8: case GGML_TYPE_I8:
@ -16150,6 +16171,12 @@ size_t ggml_quantize_chunk(enum ggml_type type, const float * src, void * dst, i
block_q4_K * block = (block_q4_K*)dst + start / QK_K; block_q4_K * block = (block_q4_K*)dst + start / QK_K;
result = ggml_quantize_q4_K(src + start, block, n, n, hist); result = ggml_quantize_q4_K(src + start, block, n, n, hist);
} break; } break;
case GGML_TYPE_Q5_K:
{
GGML_ASSERT(start % QK_K == 0);
block_q5_K * block = (block_q5_K*)dst + start / QK_K;
result = ggml_quantize_q5_K(src + start, block, n, n, hist);
} break;
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
{ {
GGML_ASSERT(start % QK_K == 0); GGML_ASSERT(start % QK_K == 0);

9
ggml.h
View file

@ -244,9 +244,9 @@ extern "C" {
// k-quantizations // k-quantizations
GGML_TYPE_Q3_K = 10, GGML_TYPE_Q3_K = 10,
GGML_TYPE_Q4_K = 11, GGML_TYPE_Q4_K = 11,
//GGML_TYPE_Q5_K = 12, GGML_TYPE_Q5_K = 12,
GGML_TYPE_Q6_K = 12, GGML_TYPE_Q6_K = 13,
GGML_TYPE_Q8_K = 13, GGML_TYPE_Q8_K = 14,
GGML_TYPE_I8, GGML_TYPE_I8,
GGML_TYPE_I16, GGML_TYPE_I16,
GGML_TYPE_I32, GGML_TYPE_I32,
@ -272,7 +272,8 @@ extern "C" {
GGML_FTYPE_MOSTLY_Q5_1 = 9, // except 1d tensors GGML_FTYPE_MOSTLY_Q5_1 = 9, // except 1d tensors
GGML_FTYPE_MOSTLY_Q3_K = 10, // except 1d tensors GGML_FTYPE_MOSTLY_Q3_K = 10, // except 1d tensors
GGML_FTYPE_MOSTLY_Q4_K = 11, // except 1d tensors GGML_FTYPE_MOSTLY_Q4_K = 11, // except 1d tensors
GGML_FTYPE_MOSTLY_Q6_K = 12, // except 1d tensors GGML_FTYPE_MOSTLY_Q5_K = 12, // except 1d tensors
GGML_FTYPE_MOSTLY_Q6_K = 13, // except 1d tensors
}; };
// available tensor operations: // available tensor operations:

299
k_quants.c
View file

@ -499,6 +499,135 @@ size_t ggml_quantize_q4_K(const float * restrict src, void * restrict dst, int n
return (n/QK_K*sizeof(block_q4_K)); return (n/QK_K*sizeof(block_q4_K));
} }
// ====================== 5-bit (de)-quantization
void quantize_row_q5_K_reference(const float * restrict x, block_q5_K * restrict y, int k) {
assert(k % QK_K == 0);
const int nb = k / QK_K;
uint8_t L[QK_K];
float mins[QK_K/32];
float scales[QK_K/32];
for (int i = 0; i < nb; i++) {
float max_scale = 0; // as we are deducting the min, scales are always positive
float max_min = 0;
for (int j = 0; j < QK_K/32; ++j) {
scales[j] = make_qkx1_quants(32, 31, x + 32*j, L + 32*j, &mins[j], 5);
float scale = scales[j];
if (scale > max_scale) {
max_scale = scale;
}
float min = mins[j];
if (min > max_min) {
max_min = min;
}
}
float inv_scale = max_scale > 0 ? 63.f/max_scale : 0.f;
float inv_min = max_min > 0 ? 63.f/max_min : 0.f;
for (int j = 0; j < QK_K/32; ++j) {
uint8_t ls = nearest_int(inv_scale*scales[j]);
uint8_t lm = nearest_int(inv_min*mins[j]);
ls = MIN(63, ls);
lm = MIN(63, lm);
if (j < 4) {
y[i].scales[j] = ls;
y[i].scales[j+4] = lm;
} else {
y[i].scales[j+4] = (ls & 0xF) | ((lm & 0xF) << 4);
y[i].scales[j-4] |= ((ls >> 4) << 6);
y[i].scales[j-0] |= ((lm >> 4) << 6);
}
}
y[i].d = ggml_fp32_to_fp16(max_scale/63.f);
y[i].dmin = ggml_fp32_to_fp16(max_min/63.f);
uint8_t sc, m;
for (int j = 0; j < QK_K/32; ++j) {
get_scale_min_k4(j, y[i].scales, &sc, &m);
const float d = ggml_fp16_to_fp32(y[i].d) * sc;
if (!d) continue;
const float dm = ggml_fp16_to_fp32(y[i].dmin) * m;
for (int ii = 0; ii < 32; ++ii) {
int l = nearest_int((x[32*j + ii] + dm)/d);
l = MAX(0, MIN(31, l));
L[32*j + ii] = l;
}
}
uint8_t * restrict qh = y[i].qh;
uint8_t * restrict ql = y[i].qs;
memset(qh, 0, QK_K/8);
uint8_t m1 = 1, m2 = 2;
for (int n = 0; n < QK_K; n += 64) {
for (int j = 0; j < 32; ++j) {
int l1 = L[n + j];
if (l1 > 15) {
l1 -= 16; qh[j] |= m1;
}
int l2 = L[n + j + 32];
if (l2 > 15) {
l2 -= 16; qh[j] |= m2;
}
ql[j] = l1 | (l2 << 4);
}
m1 <<= 2; m2 <<= 2;
ql += 32;
}
x += QK_K;
}
}
void dequantize_row_q5_K(const block_q5_K * restrict x, float * restrict y, int k) {
assert(k % QK_K == 0);
const int nb = k / QK_K;
for (int i = 0; i < nb; i++) {
const float d = ggml_fp16_to_fp32(x[i].d);
const float min = ggml_fp16_to_fp32(x[i].dmin);
const uint8_t * ql = x[i].qs;
const uint8_t * qh = x[i].qh;
int is = 0;
uint8_t sc, m;
uint8_t u1 = 1, u2 = 2;
for (int j = 0; j < QK_K; j += 64) {
get_scale_min_k4(is + 0, x[i].scales, &sc, &m);
const float d1 = d * sc; const float m1 = min * m;
get_scale_min_k4(is + 1, x[i].scales, &sc, &m);
const float d2 = d * sc; const float m2 = min * m;
for (int l = 0; l < 32; ++l) *y++ = d1 * ((ql[l] & 0xF) + (qh[l] & u1 ? 16 : 0)) - m1;
for (int l = 0; l < 32; ++l) *y++ = d2 * ((ql[l] >> 4) + (qh[l] & u2 ? 16 : 0)) - m2;
ql += 32; is += 2;
u1 <<= 2; u2 <<= 2;
}
}
}
void quantize_row_q5_K(const float * restrict x, void * restrict vy, int k) {
assert(k % QK_K == 0);
block_q5_K * restrict y = vy;
quantize_row_q5_K_reference(x, y, k);
}
size_t ggml_quantize_q5_K(const float * restrict src, void * restrict dst, int n, int k, int64_t * restrict hist) {
assert(k % QK_K == 0);
const int nb = k / QK_K;
(void)hist;
for (int j = 0; j < nb; j += k) {
block_q5_K * restrict y = (block_q5_K *)dst + j/QK_K;
quantize_row_q5_K_reference(src + j, y, k);
}
return (n/QK_K*sizeof(block_q5_K));
}
// ====================== 6-bit (de)-quantization // ====================== 6-bit (de)-quantization
void quantize_row_q6_K_reference(const float * restrict x, block_q6_K * restrict y, int k) { void quantize_row_q6_K_reference(const float * restrict x, block_q6_K * restrict y, int k) {
@ -1058,6 +1187,176 @@ void ggml_vec_dot_q4_K_q8_K(const int n, float * restrict s, const void * restri
#endif #endif
} }
void ggml_vec_dot_q5_K_q8_K(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) {
assert(n % QK_K == 0);
const block_q5_K * restrict x = vx;
const block_q8_K * restrict y = vy;
const int nb = n / QK_K;
#ifdef z__ARM_NEON
GGML_ASSERT(false);
#elif defined __AVX2__
static const uint32_t kmask1 = 0x3f3f3f3f;
static const uint32_t kmask2 = 0x0f0f0f0f;
static const uint32_t kmask3 = 0x03030303;
const __m256i m4 = _mm256_set1_epi8(0xF);
const __m128i mzero = _mm_setzero_si128();
const __m256i mone = _mm256_set1_epi8(1);
__m256 acc = _mm256_setzero_ps();
uint32_t utmp[4];
float summs = 0.f;
for (int i = 0; i < nb; ++i) {
const float d = y[i].d * ggml_fp16_to_fp32(x[i].d);
const float dmin = -y[i].d * ggml_fp16_to_fp32(x[i].dmin);
const uint8_t * restrict q5 = x[i].qs;
const int8_t * restrict q8 = y[i].qs;
memcpy(utmp, x[i].scales, 12);
utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4);
const uint32_t uaux = utmp[1] & kmask1;
utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4);
utmp[2] = uaux;
utmp[0] &= kmask1;
const __m256i mins_and_scales = _mm256_cvtepu8_epi16(_mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0]));
const __m256i q8sums = _mm256_loadu_si256((const __m256i*)y[i].bsums);
const __m128i q8s = _mm_hadd_epi16(_mm256_extracti128_si256(q8sums, 0), _mm256_extracti128_si256(q8sums, 1));
const __m128i prod = _mm_madd_epi16(_mm256_extracti128_si256(mins_and_scales, 1), q8s);
const __m128i hsum = _mm_hadd_epi32(_mm_hadd_epi32(prod, mzero), mzero);
summs += dmin * _mm_extract_epi32(hsum, 0);
const __m128i sc128 = _mm256_extracti128_si256(mins_and_scales, 0);
const __m256i scales = _mm256_set_m128i(sc128, sc128);
const __m256i hbits = _mm256_loadu_si256((const __m256i*)x[i].qh);
__m256i hmask = mone;
__m256i sumi = _mm256_setzero_si256();
int bit = 0;
for (int j = 0; j < QK_K/64; ++j) {
const __m256i scale_0 = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+0));
const __m256i scale_1 = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+1));
const __m256i q5bits = _mm256_loadu_si256((const __m256i*)q5); q5 += 32;
const __m256i q5l_0 = _mm256_and_si256(q5bits, m4);
const __m256i q5h_0 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_and_si256(hbits, hmask), bit++), 4);
const __m256i q5_0 = _mm256_add_epi8(q5l_0, q5h_0);
hmask = _mm256_slli_epi16(hmask, 1);
const __m256i q5l_1 = _mm256_and_si256(_mm256_srli_epi16(q5bits, 4), m4);
const __m256i q5h_1 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_and_si256(hbits, hmask), bit++), 4);
const __m256i q5_1 = _mm256_add_epi8(q5l_1, q5h_1);
hmask = _mm256_slli_epi16(hmask, 1);
const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32;
const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32;
__m256i p16_0 = _mm256_maddubs_epi16(q5_0, q8_0);
__m256i p16_1 = _mm256_maddubs_epi16(q5_1, q8_1);
p16_0 = _mm256_madd_epi16(scale_0, p16_0);
p16_1 = _mm256_madd_epi16(scale_1, p16_1);
sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_0, p16_1));
}
__m256 vd = _mm256_set1_ps(d);
acc = _mm256_fmadd_ps(vd, _mm256_cvtepi32_ps(sumi), acc);
}
*s = hsum_float_8(acc) + summs;
#else
static const uint32_t kmask1 = 0x3f3f3f3f;
static const uint32_t kmask2 = 0x0f0f0f0f;
static const uint32_t kmask3 = 0x03030303;
uint32_t utmp[4];
const uint8_t * scales = (const uint8_t*)&utmp[0];
const uint8_t * mins = (const uint8_t*)&utmp[2];
int8_t aux8[QK_K];
int16_t aux16[8];
float sums [8];
int32_t aux32[8];
memset(sums, 0, 8*sizeof(float));
float sumf = 0;
for (int i = 0; i < nb; ++i) {
const uint8_t * restrict q4 = x[i].qs;
const uint8_t * restrict hm = x[i].qh;
const int8_t * restrict q8 = y[i].qs;
memset(aux32, 0, 8*sizeof(int32_t));
int8_t * restrict a = aux8;
uint8_t m = 1;
for (int j = 0; j < QK_K/64; ++j) {
for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] & 0xF);
for (int l = 0; l < 32; ++l) a[l] += (hm[l] & m ? 16 : 0);
a += 32; m <<= 1;
for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] >> 4);
for (int l = 0; l < 32; ++l) a[l] += (hm[l] & m ? 16 : 0);
a += 32; m <<= 1;
q4 += 32;
}
memcpy(utmp, x[i].scales, 12);
utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4);
const uint32_t uaux = utmp[1] & kmask1;
utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4);
utmp[2] = uaux;
utmp[0] &= kmask1;
int sumi = 0;
for (int j = 0; j < QK_K/16; ++j) sumi += y[i].bsums[j] * mins[j/2];
a = aux8;
int is = 0;
for (int j = 0; j < QK_K/32; ++j) {
int32_t scale = scales[is++];
for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l];
for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l];
q8 += 8; a += 8;
for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l];
for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l];
q8 += 8; a += 8;
for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l];
for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l];
q8 += 8; a += 8;
for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l];
for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l];
q8 += 8; a += 8;
}
const float d = ggml_fp16_to_fp32(x[i].d) * y[i].d;
for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l];
const float dmin = ggml_fp16_to_fp32(x[i].dmin) * y[i].d;
sumf -= dmin * sumi;
}
for (int l = 0; l < 8; ++l) sumf += sums[l];
*s = sumf;
#endif
}
void ggml_vec_dot_q6_K_q8_K(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { void ggml_vec_dot_q6_K_q8_K(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) {
assert(n % QK_K == 0); assert(n % QK_K == 0);

5
llama.cpp
View file

@ -509,6 +509,7 @@ struct llama_file_loader {
case GGML_TYPE_Q8_0: case GGML_TYPE_Q8_0:
case GGML_TYPE_Q3_K: case GGML_TYPE_Q3_K:
case GGML_TYPE_Q4_K: case GGML_TYPE_Q4_K:
case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
break; break;
default: { default: {
@ -586,7 +587,7 @@ struct llama_file_saver {
case GGML_TYPE_Q5_1: case GGML_TYPE_Q5_1:
case GGML_TYPE_Q8_0: case GGML_TYPE_Q8_0:
case GGML_TYPE_Q3_K: case GGML_TYPE_Q3_K:
case GGML_TYPE_Q4_K: case GGML_TYPE_Q5_K:
case GGML_TYPE_Q6_K: case GGML_TYPE_Q6_K:
break; break;
default: LLAMA_ASSERT(false); default: LLAMA_ASSERT(false);
@ -906,6 +907,7 @@ static const char *llama_ftype_name(enum llama_ftype ftype) {
case LLAMA_FTYPE_MOSTLY_Q8_0: return "mostly Q8_0"; case LLAMA_FTYPE_MOSTLY_Q8_0: return "mostly Q8_0";
case LLAMA_FTYPE_MOSTLY_Q3_K: return "mostly Q3_K"; case LLAMA_FTYPE_MOSTLY_Q3_K: return "mostly Q3_K";
case LLAMA_FTYPE_MOSTLY_Q4_K: return "mostly Q4_K"; case LLAMA_FTYPE_MOSTLY_Q4_K: return "mostly Q4_K";
case LLAMA_FTYPE_MOSTLY_Q5_K: return "mostly Q5_K";
case LLAMA_FTYPE_MOSTLY_Q6_K: return "mostly Q6_K"; case LLAMA_FTYPE_MOSTLY_Q6_K: return "mostly Q6_K";
default: return "unknown, may not work"; default: return "unknown, may not work";
} }
@ -2074,6 +2076,7 @@ static void llama_model_quantize_internal(const std::string & fname_inp, const s
case LLAMA_FTYPE_MOSTLY_Q8_0: quantized_type = GGML_TYPE_Q8_0; break; case LLAMA_FTYPE_MOSTLY_Q8_0: quantized_type = GGML_TYPE_Q8_0; break;
case LLAMA_FTYPE_MOSTLY_Q3_K: quantized_type = GGML_TYPE_Q3_K; break; case LLAMA_FTYPE_MOSTLY_Q3_K: quantized_type = GGML_TYPE_Q3_K; break;
case LLAMA_FTYPE_MOSTLY_Q4_K: quantized_type = GGML_TYPE_Q4_K; break; case LLAMA_FTYPE_MOSTLY_Q4_K: quantized_type = GGML_TYPE_Q4_K; break;
case LLAMA_FTYPE_MOSTLY_Q5_K: quantized_type = GGML_TYPE_Q5_K; break;
case LLAMA_FTYPE_MOSTLY_Q6_K: quantized_type = GGML_TYPE_Q6_K; break; case LLAMA_FTYPE_MOSTLY_Q6_K: quantized_type = GGML_TYPE_Q6_K; break;
default: throw format("invalid output file type %d\n", ftype); default: throw format("invalid output file type %d\n", ftype);
}; };

3
llama.h
View file

@ -96,7 +96,8 @@ extern "C" {
LLAMA_FTYPE_MOSTLY_Q5_1 = 9, // except 1d tensors LLAMA_FTYPE_MOSTLY_Q5_1 = 9, // except 1d tensors
LLAMA_FTYPE_MOSTLY_Q3_K = 10,// except 1d tensors LLAMA_FTYPE_MOSTLY_Q3_K = 10,// except 1d tensors
LLAMA_FTYPE_MOSTLY_Q4_K = 11,// except 1d tensors LLAMA_FTYPE_MOSTLY_Q4_K = 11,// except 1d tensors
LLAMA_FTYPE_MOSTLY_Q6_K = 12,// except 1d tensors LLAMA_FTYPE_MOSTLY_Q5_K = 12,// except 1d tensors
LLAMA_FTYPE_MOSTLY_Q6_K = 13,// except 1d tensors
}; };
LLAMA_API struct llama_context_params llama_context_default_params(); LLAMA_API struct llama_context_params llama_context_default_params();