ggml : add CLBlast q5_0, q5_1, q8_0 dequant kernels (#1225)

* Implement q5_0, q5_1 and q8_0

* Work around q5_0 OpenCL issue

* Fix q8_0 dequant kernel

* Move cl kernels into ggml-opencl.c

* Use two memcpy calls for q5_0 buffer transfer
This commit is contained in:
0cc4m 2023-04-30 20:34:52 +02:00 committed by GitHub
parent 6bc4400e67
commit 76a884920a
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 205 additions and 78 deletions

View file

@ -1,63 +0,0 @@
#define MULTILINE_QUOTE(...) #__VA_ARGS__
const char * clblast_dequant = MULTILINE_QUOTE(
struct block_q4_0
{
float d;
uchar qs[16];
};
__kernel void dequantize_row_q4_0(__global struct block_q4_0* blocks, __global float* result) {
const uint i = get_global_id(0) / 32;
const uint l = get_local_id(0);
const float d = blocks[i].d;
const uchar vi = blocks[i].qs[l];
const uint index = i*32 + l*2;
result[index + 0] = ((vi & 0xf) - 8)*d;
result[index + 1] = ((vi >> 4) - 8)*d;
}
struct block_q4_1
{
float d;
float m;
uchar qs[16];
};
__kernel void dequantize_row_q4_1(__global struct block_q4_1* blocks, __global float* result) {
const uint i = get_global_id(0) / 32;
const uint l = get_local_id(0);
const float d = blocks[i].d;
const float m = blocks[i].m;
const uchar vi = blocks[i].qs[l];
const uint index = i*32 + l*2;
result[index + 0] = (vi & 0xf) * d + m;
result[index + 1] = (vi >> 4) * d + m;
}
struct block_q4_2
{
ushort d;
uchar qs[8];
};
__kernel void dequantize_row_q4_2(__global struct block_q4_2* blocks, __global float* result) {
const uint i = get_global_id(0) / 16;
const uint l = get_local_id(0);
const float d = vload_half(0, (__global half*) &blocks[i].d);;
const uchar vi = blocks[i].qs[l];
const uint index = i*16 + l*2;
result[index + 0] = ((vi & 0xf) - 8)*d;
result[index + 1] = ((vi >> 4) - 8)*d;
}
);

View file

@ -3,12 +3,141 @@
#define CL_TARGET_OPENCL_VERSION 110 #define CL_TARGET_OPENCL_VERSION 110
#include <clblast_c.h> #include <clblast_c.h>
#include <stdlib.h>
#include <stdio.h> #include <stdio.h>
#include <string.h> #include <string.h>
#include "ggml.h" #include "ggml.h"
#include "ggml-opencl-dequant.cl" #define MULTILINE_QUOTE(...) #__VA_ARGS__
const char * clblast_dequant = MULTILINE_QUOTE(
struct block_q4_0
{
float d;
uchar qs[16];
};
__kernel void dequantize_row_q4_0(__global struct block_q4_0* blocks, __global float* result) {
const uint i = get_global_id(0) / 32;
const uint l = get_local_id(0);
const float d = blocks[i].d;
const uchar vi = blocks[i].qs[l];
const uint index = i*32 + l*2;
result[index + 0] = ((vi & 0xf) - 8)*d;
result[index + 1] = ((vi >> 4) - 8)*d;
}
struct block_q4_1
{
float d;
float m;
uchar qs[16];
};
__kernel void dequantize_row_q4_1(__global struct block_q4_1* blocks, __global float* result) {
const uint i = get_global_id(0) / 32;
const uint l = get_local_id(0);
const float d = blocks[i].d;
const float m = blocks[i].m;
const uchar vi = blocks[i].qs[l];
const uint index = i*32 + l*2;
result[index + 0] = (vi & 0xf) * d + m;
result[index + 1] = (vi >> 4) * d + m;
}
struct block_q4_2
{
ushort d;
uchar qs[8];
};
__kernel void dequantize_row_q4_2(__global struct block_q4_2* blocks, __global float* result) {
const uint i = get_global_id(0) / 16;
const uint l = get_local_id(0);
const float d = vload_half(0, (__global half*) &blocks[i].d);
const uchar vi = blocks[i].qs[l];
const uint index = i*16 + l*2;
result[index + 0] = ((vi & 0xf) - 8)*d;
result[index + 1] = ((vi >> 4) - 8)*d;
}
struct block_q5_0
{
float d;
uint qh;
uchar qs[16];
};
__kernel void dequantize_row_q5_0(__global struct block_q5_0* blocks, __global float* result) {
const uint i = get_global_id(0) / 32;
const uint l = get_local_id(0);
const float d = blocks[i].d;
const uchar vi = blocks[i].qs[l];
const uint l2 = l * 2;
const uchar vh0 = ((blocks[i].qh & (1 << (l2 + 0))) >> (l2 + 0)) << 4;
const uchar vh1 = ((blocks[i].qh & (1 << (l2 + 1))) >> (l2 + 1)) << 4;
const uint index = i*32 + l2;
result[index + 0] = (((vi & 0xf) | vh0) - 16)*d;
result[index + 1] = (((vi >> 4) | vh1) - 16)*d;
}
struct block_q5_1
{
ushort d;
ushort m;
uint qh;
uchar qs[16];
};
__kernel void dequantize_row_q5_1(__global struct block_q5_1* blocks, __global float* result) {
const uint i = get_global_id(0) / 32;
const uint l = get_local_id(0);
const float d = vload_half(0, (__global half*) &blocks[i].d);
const float m = vload_half(0, (__global half*) &blocks[i].m);
const uchar vi = blocks[i].qs[l];
const uint l2 = l * 2;
const uchar vh0 = ((blocks[i].qh & (1 << (l2 + 0))) >> (l2 + 0)) << 4;
const uchar vh1 = ((blocks[i].qh & (1 << (l2 + 1))) >> (l2 + 1)) << 4;
const uint index = i*32 + l2;
result[index + 0] = ((vi & 0xf) | vh0)*d + m;
result[index + 1] = ((vi >> 4) | vh1)*d + m;
}
struct block_q8_0
{
float d;
char qs[32];
};
__kernel void dequantize_row_q8_0(__global struct block_q8_0* blocks, __global float* result) {
const uint i = get_global_id(0) / 32;
const uint l = get_local_id(0);
result[i*32 + l] = blocks[i].qs[l] * blocks[i].d;
}
);
#define CL_CHECK(err, name) \ #define CL_CHECK(err, name) \
do { \ do { \
@ -19,12 +148,26 @@
} \ } \
} while (0) } while (0)
#define QK5_0 32
typedef struct {
ggml_fp16_t d; // delta
uint8_t qh[4]; // 5-th bit of quants
uint8_t qs[QK5_0 / 2]; // nibbles / quants
} block_q5_0;
typedef struct {
float d; // delta
uint32_t qh; // 5-th bit of quants
uint8_t qs[QK5_0 / 2]; // nibbles / quants
} cl_block_q5_0;
static cl_platform_id platform; static cl_platform_id platform;
static cl_device_id device; static cl_device_id device;
static cl_context context; static cl_context context;
static cl_command_queue queue; static cl_command_queue queue;
static cl_program program; static cl_program program;
static cl_kernel kernel_q4_0, kernel_q4_1, kernel_q4_2; static cl_kernel kernel_q4_0, kernel_q4_1, kernel_q4_2, kernel_q5_0, kernel_q5_1, kernel_q8_0;
static cl_mem cl_buffer_a, cl_buffer_qb, cl_buffer_b, cl_buffer_c; static cl_mem cl_buffer_a, cl_buffer_qb, cl_buffer_b, cl_buffer_c;
static size_t cl_size_a = 0, cl_size_qb = 0, cl_size_b = 0, cl_size_c = 0; static size_t cl_size_a = 0, cl_size_qb = 0, cl_size_b = 0, cl_size_c = 0;
@ -97,6 +240,12 @@ void ggml_cl_init(void) {
CL_CHECK(err, "clCreateKernel"); CL_CHECK(err, "clCreateKernel");
kernel_q4_2 = clCreateKernel(program, "dequantize_row_q4_2", &err); kernel_q4_2 = clCreateKernel(program, "dequantize_row_q4_2", &err);
CL_CHECK(err, "clCreateKernel"); CL_CHECK(err, "clCreateKernel");
kernel_q5_0 = clCreateKernel(program, "dequantize_row_q5_0", &err);
CL_CHECK(err, "clCreateKernel");
kernel_q5_1 = clCreateKernel(program, "dequantize_row_q5_1", &err);
CL_CHECK(err, "clCreateKernel");
kernel_q8_0 = clCreateKernel(program, "dequantize_row_q8_0", &err);
CL_CHECK(err, "clCreateKernel");
} }
static void ggml_cl_malloc(size_t req_size, size_t* cur_size, cl_mem_flags flags, cl_mem* buf) { static void ggml_cl_malloc(size_t req_size, size_t* cur_size, cl_mem_flags flags, cl_mem* buf) {
@ -125,6 +274,7 @@ void ggml_cl_sgemm_wrapper(
cl_kernel kernel; cl_kernel kernel;
size_t global = n * k, local, size_qb; size_t global = n * k, local, size_qb;
bool dequant; bool dequant;
cl_block_q5_0* cl_host_b;
switch (btype) { switch (btype) {
case GGML_TYPE_F32: case GGML_TYPE_F32:
@ -146,7 +296,36 @@ void ggml_cl_sgemm_wrapper(
dequant = true; dequant = true;
kernel = kernel_q4_2; kernel = kernel_q4_2;
local = 8; local = 8;
size_qb = global * (sizeof(short) + local) / 16; size_qb = global * (sizeof(ggml_fp16_t) + local) / 16;
break;
case GGML_TYPE_Q5_0:
dequant = true;
kernel = kernel_q5_0;
local = 16;
// For some reason OpenCL seems to be incapable of working with structs of size 22.
// 20 and 24 bytes are fine. Workaround to do the fp16 to fp32 step on CPU...
// TODO Find the reason, fix and remove workaround.
const block_q5_0* b = (const block_q5_0*) host_b;
cl_host_b = (cl_block_q5_0*) malloc(sizeof(cl_block_q5_0) * global / 32);
for (size_t i = 0; i < global / 32; i++) {
cl_host_b[i].d = ggml_fp16_to_fp32(b[i].d);
memcpy(&cl_host_b[i].qh, b[i].qh, sizeof(uint32_t));
memcpy(&cl_host_b[i].qs, b[i].qs, QK5_0 / 2);
}
host_b = (const float*) cl_host_b;
size_qb = global * (sizeof(float) + sizeof(uint32_t) + local) / 32;
break;
case GGML_TYPE_Q5_1:
dequant = true;
kernel = kernel_q5_1;
local = 16;
size_qb = global * (sizeof(ggml_fp16_t) * 2 + sizeof(uint32_t) + local) / 32;
break;
case GGML_TYPE_Q8_0:
dequant = true;
kernel = kernel_q8_0;
local = 32;
size_qb = global * (sizeof(float) + local) / 32;
break; break;
default: default:
fprintf(stderr, "Error: Unsupported OpenCL btype %d\n", btype); fprintf(stderr, "Error: Unsupported OpenCL btype %d\n", btype);
@ -171,12 +350,15 @@ void ggml_cl_sgemm_wrapper(
err = clSetKernelArg(kernel, 0, sizeof(cl_mem), &cl_buffer_qb); err = clSetKernelArg(kernel, 0, sizeof(cl_mem), &cl_buffer_qb);
err |= clSetKernelArg(kernel, 1, sizeof(cl_mem), &cl_buffer_b); err |= clSetKernelArg(kernel, 1, sizeof(cl_mem), &cl_buffer_b);
CL_CHECK(err, "clSetKernelArg"); CL_CHECK(err, "clSetKernelArg");
clEnqueueWriteBuffer(queue, cl_buffer_qb, CL_FALSE, 0, size_qb, host_b, 0, NULL, &ev_qb); err = clEnqueueWriteBuffer(queue, cl_buffer_qb, CL_FALSE, 0, size_qb, host_b, 0, NULL, &ev_qb);
CL_CHECK(err, "clEnqueueWriteBuffer qb");
} else { } else {
clEnqueueWriteBuffer(queue, cl_buffer_b, CL_FALSE, 0, size_b, host_b, 0, NULL, &ev_b); err = clEnqueueWriteBuffer(queue, cl_buffer_b, CL_FALSE, 0, size_b, host_b, 0, NULL, &ev_b);
CL_CHECK(err, "clEnqueueWriteBuffer b");
} }
clEnqueueWriteBuffer(queue, cl_buffer_a, CL_FALSE, 0, size_a, host_a, 0, NULL, &ev_a); err = clEnqueueWriteBuffer(queue, cl_buffer_a, CL_FALSE, 0, size_a, host_a, 0, NULL, &ev_a);
CL_CHECK(err, "clEnqueueWriteBuffer a");
if (dequant) { if (dequant) {
err = clEnqueueNDRangeKernel(queue, kernel, 1, NULL, &global, &local, 1, &ev_qb, &ev_b); err = clEnqueueNDRangeKernel(queue, kernel, 1, NULL, &global, &local, 1, &ev_qb, &ev_b);
CL_CHECK(err, "clEnqueueNDRangeKernel"); CL_CHECK(err, "clEnqueueNDRangeKernel");
@ -188,15 +370,20 @@ void ggml_cl_sgemm_wrapper(
clReleaseEvent(ev_b); clReleaseEvent(ev_b);
cl_event ev_sgemm; cl_event ev_sgemm;
CLBlastSgemm((CLBlastLayout)order, CLBlastStatusCode status = CLBlastSgemm((CLBlastLayout)order,
(CLBlastTranspose)trans_a, (CLBlastTranspose)trans_b, (CLBlastTranspose)trans_a, (CLBlastTranspose)trans_b,
m, n, k, m, n, k,
alpha, alpha,
cl_buffer_a, 0, lda, cl_buffer_a, 0, lda,
cl_buffer_b, 0, ldb, cl_buffer_b, 0, ldb,
beta, beta,
cl_buffer_c, 0, ldc, cl_buffer_c, 0, ldc,
&queue, &ev_sgemm); &queue, &ev_sgemm);
if (status != CLBlastSuccess) {
fprintf(stderr, "Error: CLBlast SGEMM %d\n", status);
abort();
}
cl_event ev_c; cl_event ev_c;
clEnqueueReadBuffer(queue, cl_buffer_c, CL_TRUE, 0, size_c, host_c, 1, &ev_sgemm, &ev_c); clEnqueueReadBuffer(queue, cl_buffer_c, CL_TRUE, 0, size_c, host_c, 1, &ev_sgemm, &ev_c);
@ -205,4 +392,7 @@ void ggml_cl_sgemm_wrapper(
clWaitForEvents(1, &ev_c); clWaitForEvents(1, &ev_c);
clReleaseEvent(ev_sgemm); clReleaseEvent(ev_sgemm);
clReleaseEvent(ev_c); clReleaseEvent(ev_c);
if (btype == GGML_TYPE_Q5_0) {
free((void*) cl_host_b);
}
} }