mirror of
https://github.com/microsoft/mscclpp.git
synced 2026-05-13 01:36:10 +00:00
127 lines
9.3 KiB
Plaintext
127 lines
9.3 KiB
Plaintext
// Copyright (c) Microsoft Corporation.
|
|
// Licensed under the MIT license.
|
|
|
|
#include <assert.h>
|
|
|
|
#if defined(__HIP_PLATFORM_AMD__)
|
|
#include <hip/hip_fp16.h>
|
|
#else
|
|
#include <cuda_fp16.h>
|
|
#endif
|
|
|
|
// Numerical Recipes ranqd1, Chapter 7.1, §An Even Quicker Generator, Eq. 7.1.6
|
|
// parameters from Knuth and H. W. Lewis
|
|
static __device__ unsigned int ranqd1(unsigned int seed) {
|
|
const unsigned int a = 1664525;
|
|
const unsigned int c = 1013904223;
|
|
return a * seed + c;
|
|
}
|
|
|
|
// fill/test kernel pairs must have the same thread block size to
|
|
// match their random number series.
|
|
|
|
#define FILL_DATA(FuncNameType, DataType) \
|
|
extern "C" __global__ void __launch_bounds__(1024, 1) \
|
|
fill_data_##FuncNameType(DataType* input_buf, size_t num_elems, int rank, int seq) { \
|
|
unsigned int seed = (unsigned int)(blockIdx.x * blockDim.x + threadIdx.x + rank + seq); \
|
|
for (size_t i = blockIdx.x * blockDim.x + threadIdx.x; i < num_elems; i += blockDim.x * gridDim.x) { \
|
|
seed = ranqd1(seed); \
|
|
input_buf[i] = DataType(seed % blockDim.x) / DataType(blockDim.x); \
|
|
} \
|
|
}
|
|
|
|
FILL_DATA(float16, __half)
|
|
FILL_DATA(float32, float)
|
|
FILL_DATA(int32, int)
|
|
|
|
#define TEST_DATA_ALL_GATHER(FuncNameType, DataType) \
|
|
extern "C" __global__ void __launch_bounds__(1024, 1) test_data_all_gather_##FuncNameType( \
|
|
DataType* result_buf, DataType* test_buf, size_t num_elems, int num_ranks, int my_rank, int seq) { \
|
|
for (int rank = 0; rank < num_ranks; rank++) { \
|
|
size_t rank_offset = rank * num_elems; \
|
|
unsigned int seed = (unsigned int)(blockIdx.x * blockDim.x + threadIdx.x + rank + seq); \
|
|
for (size_t i = blockIdx.x * blockDim.x + threadIdx.x; i < num_elems; i += blockDim.x * gridDim.x) { \
|
|
seed = ranqd1(seed); \
|
|
test_buf[rank_offset + i] = DataType(seed % blockDim.x) / DataType(blockDim.x); \
|
|
assert(result_buf[rank_offset + i] == test_buf[rank_offset + i]); \
|
|
} \
|
|
} \
|
|
}
|
|
|
|
TEST_DATA_ALL_GATHER(float16, __half)
|
|
TEST_DATA_ALL_GATHER(float32, float)
|
|
TEST_DATA_ALL_GATHER(int32, int)
|
|
|
|
#define TEST_DATA_ALL_REDUCE(FuncNameType, DataType, Atol, Rtol) \
|
|
extern "C" __global__ void __launch_bounds__(1024, 1) test_data_all_reduce_##FuncNameType( \
|
|
DataType* result_buf, DataType* test_buf, size_t num_elems, int num_ranks, int my_rank, int seq) { \
|
|
for (int rank = 0; rank < num_ranks; rank++) { \
|
|
unsigned int seed = (unsigned int)(blockIdx.x * blockDim.x + threadIdx.x + rank + seq); \
|
|
for (size_t i = blockIdx.x * blockDim.x + threadIdx.x; i < num_elems; i += blockDim.x * gridDim.x) { \
|
|
if (rank == 0) { \
|
|
test_buf[i] = 0; \
|
|
} \
|
|
seed = ranqd1(seed); \
|
|
test_buf[i] += DataType(seed % blockDim.x) / DataType(blockDim.x); \
|
|
} \
|
|
} \
|
|
for (size_t i = blockIdx.x * blockDim.x + threadIdx.x; i < num_elems; i += blockDim.x * gridDim.x) { \
|
|
float diff = abs(float(result_buf[i]) - float(test_buf[i])); \
|
|
float expected = abs(float(test_buf[i])); \
|
|
assert(diff <= num_ranks * ((Atol) + (Rtol) * expected)); \
|
|
} \
|
|
}
|
|
|
|
TEST_DATA_ALL_REDUCE(float16, __half, 1.0 / 128, 1.0 / 128)
|
|
TEST_DATA_ALL_REDUCE(float32, float, 1e-7, 1e-7)
|
|
TEST_DATA_ALL_REDUCE(int32, int, 0, 0)
|
|
|
|
#define TEST_DATA_REDUCE_SCATTER(FuncNameType, DataType, Atol, Rtol) \
|
|
extern "C" __global__ void __launch_bounds__(1024, 1) test_data_reduce_scatter_##FuncNameType( \
|
|
DataType* result_buf, DataType* test_buf, size_t num_elems, int num_ranks, int my_rank, int seq) { \
|
|
int nem_elems_per_rank = num_elems / num_ranks; \
|
|
int offset = nem_elems_per_rank * my_rank; \
|
|
for (int rank = 0; rank < num_ranks; rank++) { \
|
|
unsigned int seed = (unsigned int)(blockIdx.x * blockDim.x + threadIdx.x + rank + seq); \
|
|
for (size_t i = blockIdx.x * blockDim.x + threadIdx.x; i < num_elems; i += blockDim.x * gridDim.x) { \
|
|
if (rank == 0) { \
|
|
test_buf[i] = 0; \
|
|
} \
|
|
seed = ranqd1(seed); \
|
|
test_buf[i] += DataType(seed % blockDim.x) / DataType(blockDim.x); \
|
|
} \
|
|
} \
|
|
for (size_t i = blockIdx.x * blockDim.x + threadIdx.x; i < num_elems; i += blockDim.x * gridDim.x) { \
|
|
if (i >= offset && i < offset + nem_elems_per_rank) { \
|
|
float diff = abs(float(result_buf[i - offset]) - float(test_buf[i])); \
|
|
float expected = abs(float(test_buf[i])); \
|
|
assert(diff <= num_ranks * ((Atol) + (Rtol) * expected)); \
|
|
} \
|
|
} \
|
|
}
|
|
|
|
TEST_DATA_REDUCE_SCATTER(float16, __half, 1.0 / 128, 1.0 / 128)
|
|
TEST_DATA_REDUCE_SCATTER(float32, float, 1e-7, 1e-7)
|
|
TEST_DATA_REDUCE_SCATTER(int32, int, 0, 0)
|
|
|
|
#define TEST_DATA_ALL_TO_ALL(FuncNameType, DataType) \
|
|
extern "C" __global__ void __launch_bounds__(1024, 1) test_data_all_to_all_##FuncNameType( \
|
|
DataType* result_buf, DataType* test_buf, size_t num_elems, int num_ranks, int my_rank, int seq) { \
|
|
int nem_elems_per_rank = num_elems / num_ranks; \
|
|
int offset = nem_elems_per_rank * my_rank; \
|
|
for (int rank = 0; rank < num_ranks; rank++) { \
|
|
size_t rank_offset = rank * nem_elems_per_rank; \
|
|
unsigned int seed = (unsigned int)(blockIdx.x * blockDim.x + threadIdx.x + rank + seq); \
|
|
for (size_t i = blockIdx.x * blockDim.x + threadIdx.x; i < num_elems; i += blockDim.x * gridDim.x) { \
|
|
seed = ranqd1(seed); \
|
|
if (i >= my_rank * nem_elems_per_rank && i < (my_rank + 1) * nem_elems_per_rank) { \
|
|
test_buf[rank_offset + i - offset] = DataType(seed % blockDim.x) / DataType(blockDim.x); \
|
|
assert(result_buf[rank_offset + i - offset] == test_buf[rank_offset + i - offset]); \
|
|
} \
|
|
} \
|
|
} \
|
|
}
|
|
|
|
TEST_DATA_ALL_TO_ALL(float16, __half)
|
|
TEST_DATA_ALL_TO_ALL(float32, float)
|
|
TEST_DATA_ALL_TO_ALL(int32, int) |