Skip to content

Add ability to use Q5_0, Q5_1, and IQ4_NL for quantized K cache #6183

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 7 commits into from
Mar 21, 2024
Merged

Add ability to use Q5_0, Q5_1, and IQ4_NL for quantized K cache #6183

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
5e09ce4
k_cache: be able to use Q5_0
Kawrakow Mar 20, 2024
5d8822b
k_cache: be able to use Q5_1 on CODA
Kawrakow Mar 20, 2024
fef4a23
k_cache: be able to use Q5_0 on Metal
Kawrakow Mar 20, 2024
d68030b
k_cache: be able to use Q5_1 on Metal
Kawrakow Mar 20, 2024
9711e1e
k_cache: be able to use IQ4_NL - just CUDA for now
Kawrakow Mar 20, 2024
d8a498d
k_cache: be able to use IQ4_NL on Metal
Kawrakow Mar 20, 2024
9e1bda9
k_cache: add newly added supported types to llama-bench and CUDA supp…
Kawrakow Mar 20, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions common/common.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1590,6 +1590,9 @@ static ggml_type kv_cache_type_from_str(const std::string & s) {
if (s == "q4_1") {
return GGML_TYPE_Q4_1;
}
if (s == "iq4_nl") {
return GGML_TYPE_IQ4_NL;
}
if (s == "q5_0") {
return GGML_TYPE_Q5_0;
}
Expand Down
3 changes: 3 additions & 0 deletions examples/llama-bench/llama-bench.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -249,6 +249,9 @@ static ggml_type ggml_type_from_name(const std::string & s) {
if (s == "q5_1") {
return GGML_TYPE_Q5_1;
}
if (s == "iq4_nl") {
return GGML_TYPE_IQ4_NL;
}

return GGML_TYPE_COUNT;
}
Expand Down
165 changes: 165 additions & 0 deletions ggml-cuda.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6762,6 +6762,123 @@ static __device__ void cpy_blck_f32_q4_1(const char * cxi, char * cdsti) {
}
}

static __device__ void cpy_blck_f32_q5_0(const char * cxi, char * cdsti) {
const float * xi = (const float *) cxi;
block_q5_0 * dsti = (block_q5_0 *) cdsti;

float amax = 0.0f;
float vmax = 0.0f;

for (int j = 0; j < QK5_0; ++j) {
const float v = xi[j];
if (amax < fabsf(v)) {
amax = fabsf(v);
vmax = v;
}
}

const float d = vmax / -16;
const float id = d ? 1.0f/d : 0.0f;

dsti->d = d;

uint32_t qh = 0;
for (int j = 0; j < QK5_0/2; ++j) {
const float x0 = xi[0 + j]*id;
const float x1 = xi[QK5_0/2 + j]*id;

const uint8_t xi0 = min(31, (int8_t)(x0 + 16.5f));
const uint8_t xi1 = min(31, (int8_t)(x1 + 16.5f));

dsti->qs[j] = (xi0 & 0xf) | ((xi1 & 0xf) << 4);
qh |= ((xi0 & 0x10u) >> 4) << (j + 0);
qh |= ((xi1 & 0x10u) >> 4) << (j + QK5_0/2);
}
memcpy(dsti->qh, &qh, sizeof(qh));
}

static __device__ void cpy_blck_f32_q5_1(const char * cxi, char * cdsti) {
const float * xi = (const float *) cxi;
block_q5_1 * dsti = (block_q5_1 *) cdsti;

float min = xi[0];
float max = xi[0];

for (int j = 1; j < QK5_1; ++j) {
const float v = xi[j];
min = v < min ? v : min;
max = v > max ? v : max;
}

const float d = (max - min) / 31;
const float id = d ? 1.0f/d : 0.0f;

dsti->dm.x = d;
dsti->dm.y = min;

uint32_t qh = 0;
for (int j = 0; j < QK5_1/2; ++j) {
const float x0 = (xi[0 + j] - min)*id;
const float x1 = (xi[QK5_1/2 + j] - min)*id;

const uint8_t xi0 = (uint8_t)(x0 + 0.5f);
const uint8_t xi1 = (uint8_t)(x1 + 0.5f);

dsti->qs[j] = (xi0 & 0xf) | ((xi1 & 0xf) << 4);
qh |= ((xi0 & 0x10u) >> 4) << (j + 0);
qh |= ((xi1 & 0x10u) >> 4) << (j + QK5_1/2);
}
memcpy(dsti->qh, &qh, sizeof(qh));
}

static __device__ __forceinline__ int best_index_int8(int n, const int8_t * val, float x) {
if (x <= val[0]) return 0;
if (x >= val[n-1]) return n-1;
int ml = 0, mu = n-1;
while (mu-ml > 1) {
int mav = (ml+mu)/2;
if (x < val[mav]) mu = mav; else ml = mav;
}
return x - val[mu-1] < val[mu] - x ? mu-1 : mu;
}

static __device__ void cpy_blck_f32_iq4_nl(const char * cxi, char * cdsti) {
const float * xi = (const float *) cxi;
block_iq4_nl * dsti = (block_iq4_nl *) cdsti;

float amax = 0.0f;
float vmax = 0.0f;

for (int j = 0; j < QK4_NL; ++j) {
const float v = xi[j];
if (amax < fabsf(v)) {
amax = fabsf(v);
vmax = v;
}
}

float d = vmax / kvalues_iq4nl[0];
const float id = d ? 1.0f/d : 0.0f;

float sumqx = 0, sumq2 = 0;
for (int j = 0; j < QK4_NL/2; ++j) {
const float x0 = xi[0 + j]*id;
const float x1 = xi[QK4_NL/2 + j]*id;
const uint8_t xi0 = best_index_int8(16, kvalues_iq4nl, x0);
const uint8_t xi1 = best_index_int8(16, kvalues_iq4nl, x1);
dsti->qs[j] = xi0 | (xi1 << 4);
const float v0 = kvalues_iq4nl[xi0];
const float v1 = kvalues_iq4nl[xi1];
const float w0 = xi[0 + j]*xi[0 + j];
const float w1 = xi[QK4_NL/2 + j]*xi[QK4_NL/2 + j];
sumqx += w0*v0*xi[j] + w1*v1*xi[QK4_NL/2 + j];
sumq2 += w0*v0*v0 + w1*v1*v1;
}

dsti->d = sumq2 > 0 ? sumqx/sumq2 : d;
}


template <cpy_kernel_t cpy_blck, int qk>
static __global__ void cpy_f32_q(const char * cx, char * cdst, const int ne,
const int ne00, const int ne01, const int ne02, const int nb00, const int nb01, const int nb02,
Expand Down Expand Up @@ -8495,6 +8612,39 @@ static void ggml_cpy_f32_q4_1_cuda(
(cx, cdst, ne, ne00, ne01, ne02, nb00, nb01, nb02, nb03, ne10, ne11, ne12, nb10, nb11, nb12, nb13);
}

static void ggml_cpy_f32_q5_0_cuda(
const char * cx, char * cdst, const int ne,
const int ne00, const int ne01, const int ne02, const int nb00, const int nb01, const int nb02,
const int nb03, const int ne10, const int ne11, const int ne12, const int nb10, const int nb11, const int nb12, const int nb13, cudaStream_t stream) {

GGML_ASSERT(ne % QK5_0 == 0);
const int num_blocks = ne / QK5_0;
cpy_f32_q<cpy_blck_f32_q5_0, QK5_0><<<num_blocks, 1, 0, stream>>>
(cx, cdst, ne, ne00, ne01, ne02, nb00, nb01, nb02, nb03, ne10, ne11, ne12, nb10, nb11, nb12, nb13);
}

static void ggml_cpy_f32_q5_1_cuda(
const char * cx, char * cdst, const int ne,
const int ne00, const int ne01, const int ne02, const int nb00, const int nb01, const int nb02,
const int nb03, const int ne10, const int ne11, const int ne12, const int nb10, const int nb11, const int nb12, const int nb13, cudaStream_t stream) {

GGML_ASSERT(ne % QK5_1 == 0);
const int num_blocks = ne / QK5_1;
cpy_f32_q<cpy_blck_f32_q5_1, QK5_1><<<num_blocks, 1, 0, stream>>>
(cx, cdst, ne, ne00, ne01, ne02, nb00, nb01, nb02, nb03, ne10, ne11, ne12, nb10, nb11, nb12, nb13);
}

static void ggml_cpy_f32_iq4_nl_cuda(
const char * cx, char * cdst, const int ne,
const int ne00, const int ne01, const int ne02, const int nb00, const int nb01, const int nb02,
const int nb03, const int ne10, const int ne11, const int ne12, const int nb10, const int nb11, const int nb12, const int nb13, cudaStream_t stream) {

GGML_ASSERT(ne % QK4_NL == 0);
const int num_blocks = ne / QK4_NL;
cpy_f32_q<cpy_blck_f32_iq4_nl, QK4_NL><<<num_blocks, 1, 0, stream>>>
(cx, cdst, ne, ne00, ne01, ne02, nb00, nb01, nb02, nb03, ne10, ne11, ne12, nb10, nb11, nb12, nb13);
}

static void ggml_cpy_f16_f16_cuda(
const char * cx, char * cdst, const int ne,
const int ne00, const int ne01, const int ne02, const int nb00, const int nb01, const int nb02,
Expand Down Expand Up @@ -10893,6 +11043,12 @@ static void ggml_cuda_cpy(ggml_backend_cuda_context & ctx, const ggml_tensor * s
ggml_cpy_f32_q4_0_cuda(src0_ddc, src1_ddc, ne, ne00, ne01, ne02, nb00, nb01, nb02, nb03, ne10, ne11, ne12, nb10, nb11, nb12, nb13, main_stream);
} else if (src0->type == GGML_TYPE_F32 && src1->type == GGML_TYPE_Q4_1) {
ggml_cpy_f32_q4_1_cuda(src0_ddc, src1_ddc, ne, ne00, ne01, ne02, nb00, nb01, nb02, nb03, ne10, ne11, ne12, nb10, nb11, nb12, nb13, main_stream);
} else if (src0->type == GGML_TYPE_F32 && src1->type == GGML_TYPE_Q5_0) {
ggml_cpy_f32_q5_0_cuda(src0_ddc, src1_ddc, ne, ne00, ne01, ne02, nb00, nb01, nb02, nb03, ne10, ne11, ne12, nb10, nb11, nb12, nb13, main_stream);
} else if (src0->type == GGML_TYPE_F32 && src1->type == GGML_TYPE_IQ4_NL) {
ggml_cpy_f32_iq4_nl_cuda(src0_ddc, src1_ddc, ne, ne00, ne01, ne02, nb00, nb01, nb02, nb03, ne10, ne11, ne12, nb10, nb11, nb12, nb13, main_stream);
} else if (src0->type == GGML_TYPE_F32 && src1->type == GGML_TYPE_Q5_1) {
ggml_cpy_f32_q5_1_cuda(src0_ddc, src1_ddc, ne, ne00, ne01, ne02, nb00, nb01, nb02, nb03, ne10, ne11, ne12, nb10, nb11, nb12, nb13, main_stream);
} else if (src0->type == GGML_TYPE_F16 && src1->type == GGML_TYPE_F16) {
ggml_cpy_f16_f16_cuda (src0_ddc, src1_ddc, ne, ne00, ne01, ne02, nb00, nb01, nb02, nb03, ne10, ne11, ne12, nb10, nb11, nb12, nb13, main_stream);
} else if (src0->type == GGML_TYPE_F16 && src1->type == GGML_TYPE_F32) {
Expand Down Expand Up @@ -11309,6 +11465,15 @@ GGML_CALL static bool ggml_backend_cuda_supports_op(ggml_backend_t backend, cons
if (src0_type == GGML_TYPE_F32 && src1_type == GGML_TYPE_Q4_1) {
return true;
}
if (src0_type == GGML_TYPE_F32 && src1_type == GGML_TYPE_Q5_0) {
return true;
}
if (src0_type == GGML_TYPE_F32 && src1_type == GGML_TYPE_Q5_1) {
return true;
}
if (src0_type == GGML_TYPE_F32 && src1_type == GGML_TYPE_IQ4_NL) {
return true;
}
if (src0_type == GGML_TYPE_F16 && src1_type == GGML_TYPE_F16) {
return true;
}
Expand Down
28 changes: 17 additions & 11 deletions ggml-metal.m
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -173,8 +173,9 @@
GGML_METAL_KERNEL_TYPE_CPY_F32_Q8_0,
GGML_METAL_KERNEL_TYPE_CPY_F32_Q4_0,
GGML_METAL_KERNEL_TYPE_CPY_F32_Q4_1,
//GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_0,
//GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_1,
GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_0,
GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_1,
GGML_METAL_KERNEL_TYPE_CPY_F32_IQ4_NL,
GGML_METAL_KERNEL_TYPE_CPY_F16_F16,
GGML_METAL_KERNEL_TYPE_CPY_F16_F32,
GGML_METAL_KERNEL_TYPE_CONCAT,
Expand Down Expand Up @@ -598,8 +599,9 @@ static void ggml_metal_log(enum ggml_log_level level, const char * format, ...){
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_Q8_0, cpy_f32_q8_0, true);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_Q4_0, cpy_f32_q4_0, true);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_Q4_1, cpy_f32_q4_1, true);
//GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_0, cpy_f32_q5_0, true);
//GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_1, cpy_f32_q5_1, true);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_0, cpy_f32_q5_0, true);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_1, cpy_f32_q5_1, true);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_IQ4_NL, cpy_f32_iq4_nl, true);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F16_F16, cpy_f16_f16, true);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F16_F32, cpy_f16_f32, true);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CONCAT, concat, true);
Expand Down Expand Up @@ -739,6 +741,9 @@ static bool ggml_metal_supports_op(const struct ggml_metal_context * ctx, const
case GGML_TYPE_Q8_0:
case GGML_TYPE_Q4_0:
case GGML_TYPE_Q4_1:
case GGML_TYPE_Q5_0:
case GGML_TYPE_Q5_1:
case GGML_TYPE_IQ4_NL:
return true;
default:
return false;
Expand Down Expand Up @@ -2431,13 +2436,14 @@ static enum ggml_status ggml_metal_graph_compute(
GGML_ASSERT(ne0 % ggml_blck_size(dst->type) == 0);

switch (dstt) {
case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_F16].pipeline; break;
case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_F32].pipeline; break;
case GGML_TYPE_Q8_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_Q8_0].pipeline; break;
case GGML_TYPE_Q4_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_Q4_0].pipeline; break;
case GGML_TYPE_Q4_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_Q4_1].pipeline; break;
//case GGML_TYPE_Q5_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_0].pipeline; break;
//case GGML_TYPE_Q5_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_1].pipeline; break;
case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_F16].pipeline; break;
case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_F32].pipeline; break;
case GGML_TYPE_Q8_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_Q8_0].pipeline; break;
case GGML_TYPE_Q4_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_Q4_0].pipeline; break;
case GGML_TYPE_Q4_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_Q4_1].pipeline; break;
case GGML_TYPE_Q5_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_0].pipeline; break;
case GGML_TYPE_Q5_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_1].pipeline; break;
case GGML_TYPE_IQ4_NL: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_IQ4_NL].pipeline; break;
default: GGML_ASSERT(false && "not implemented");
};
} break;
Expand Down
Loading