aclnnMatmulAllReduceV2
支持的产品型号
- Atlas A2训练系列产品/Atlas 800I A2推理产品。
说明: 使用该接口时,请确保驱动固件包和CANN包都为配套的8.0.RC2版本或者配套的更高版本,否则将会引发报错,比如BUS ERROR等。
接口原型
每个算子分为两段式接口,必须先调用“aclnnMatmulAllReduceV2GetWorkspaceSize”接口获取计算所需workspace大小以及包含了算子计算流程的执行器,再调用“aclnnMatmulAllReduceV2”接口执行计算。
aclnnStatus aclnnMatmulAllReduceV2GetWorkspaceSize(const aclTensor *x1, const aclTensor *x2, const aclTensor *bias, const aclTensor *x3, const char* group, const char *reduceOp, int64_t commTurn, int64_t streamMode, const aclTensor *output, uint64_t *workspaceSize, aclOpExecutor **executor);
aclnnStatus aclnnMatmulAllReduceV2(void *workspace, uint64_t workspaceSize, aclOpExecutor *executor, const aclrtStream stream);
功能描述
- 算子功能:完成mm + all_reduce_base计算。
- 计算公式:
aclnnMatmulAllReduceV2GetWorkspaceSize
参数说明:
- x1(const aclTensor *, 计算输入):Device侧的2维或3维aclTensor,mm左矩阵。数据类型支持:FLOAT16、BFLOAT16。数据格式支持:ND。不支持非连续输入。
- x2(const aclTensor *, 计算输入):Device侧的2维aclTensor,mm右矩阵。数据类型支持:FLOAT16、BFLOAT16。支持最后两轴转置情况下的非连续tensor传入,数据格式支持:ND。x1的第二维和x2的第一维维度相等,或x2最后两轴转置后,对应维度相等。非连续的Tensor仅支持transpose场景。
- bias(const aclTensor *, 计算输入):bias。数据类型支持:FLOAT16、BFLOAT16。数据格式支持:ND,可选,可为空。非空时shape最后一维和x2最后一维相等
- x3(const aclTensor *, 计算输入):x3,可选,matmul计算后的偏移。数据类型支持:FLOAT16、BFLOAT16。数据格式支持:ND,必选,shape与matmul计算后的shape相同。
- group(const char *, 计算输入):通信域名称。数据类型支持:String。通过Hccl提供的接口获取:extern HcclResult HcclGetCommName(HcclComm comm, char* commName); commName即为group。
- reduceOp(const char *, 计算输入):reduce操作类型。数据类型支持:String。目前仅支持"sum"。
- commTurn(int64_t, 计算输入):通信数据切分数,即总数据量/单次通信量。数据类型支持:int64_t。当前版本仅支持输入0。
- streamMode(int64_t,计算输入):Host侧的整型,acl流模式的枚举,当前只支持枚举值1,类型支持:int64_t。
- output(aclTensor *, 输出):计算+通信的结果。数据类型支持FLOAT16、BFLOAT16, 且数据类型同x1、x2输入。shape第一维与x1第一维相等,最后一维和x2最后一维相等
- workspaceSize(uint64_t *, 出参):返回需要在Device侧申请的workspace大小。
- executor(aclOpExecutor **, 出参):返回op执行器,包含了算子计算流程。
返回值:
返回aclnnStatus状态码,具体参见aclnn返回码。
aclnnMatmulAllReduceV2
参数说明:
- workspace(void*,入参):在Device侧申请的workspace内存地址。
- workspaceSize(uint64_t,入参):在Device侧申请的workspace大小,由第一段接口aclnnMatmulAllReduceV2GetWorkspaceSize。
- executor(aclOpExecutor*,入参):op执行器,包含了算子计算流程。
- stream(const aclrtStream,入参):指定执行任务的AscendCL stream流。
返回值:
返回aclnnStatus状态码,具体参见aclnn返回码。
约束与限制
- 增量场景不使能MC2,全量场景使能MC2
- 输入x1可为2维或者3维,且不为空Tensor,其维度为(b, s, k)或者(m, k)。x2必须是2维,且不为空Tensor。其维度为(k, n),轴满足mm算子入参要求,k轴相等。bias若非空,bias为1维。
- 在Atlas A2训练系列产品/Atlas 800I A2推理产品上,b*s、m、k、n的值均不得超过2147483647(INT32_MAX)。
- x3以及输出output除最后一维皆与输入x1除最后一维相等,x3以及输出output的最后1维与输入x2的最后1维相等。bias若非空,shape大小与output最后一维相等。
- 传入的x1、x2、x3或者output不为空指针。
- x1和x2、x3、output、bias(非空场景)的数据类型和数据格式需要在支持的范围之内。
- 只支持x2矩阵转置/不转置,x1矩阵支持不转置场景。
- Atlas A2训练系列产品/Atlas 800I A2推理产品支持1、2、4、8卡,并且仅支持hccs链路all mesh组网。
- 一个模型中的通算融合MC2算子,仅支持相同通信域。
调用示例
#include <iostream>
#include <vector>
#include <thread>
#include "aclnnop/aclnn_matmul_all_reduce_v2.h"
int ndev = 8;
#define CHECK_RET(cond, return_expr) \
do { \
if (!(cond)) { \
return_expr; \
} \
} while (0)
#define LOG_PRINT(message, ...) \
do { \
printf(message, ##__VA_ARGS__); \
} while (0)
int64_t GetShapeSize(const std::vector<int64_t> &shape) {
int64_t shapeSize = 1;
for (auto i: shape) {
shapeSize *= i;
}
return shapeSize;
}
template<typename T>
int CreateAclTensor(const std::vector<T> &hostData, const std::vector<int64_t> &shape, void **deviceAddr,
aclDataType dataType, aclTensor **tensor) {
auto size = GetShapeSize(shape) * sizeof(T);
// 调用aclrtMalloc申请device侧内存
auto ret = aclrtMalloc(deviceAddr, size, ACL_MEM_MALLOC_HUGE_FIRST);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtMalloc failed. ERROR: %d\n", ret); return ret);
// 调用aclrtMemcpy将host侧数据拷贝到device侧内存上
ret = aclrtMemcpy(*deviceAddr, size, hostData.data(), size, ACL_MEMCPY_HOST_TO_DEVICE);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtMemcpy failed. ERROR: %d\n", ret); return ret);
// 计算连续tensor的strides
std::vector<int64_t> strides(shape.size(), 1);
for (int64_t i = shape.size() - 2; i >= 0; i--) {
strides[i] = shape[i + 1] * strides[i + 1];
}
// 调用aclCreateTensor接口创建aclTensor
*tensor = aclCreateTensor(shape.data(), shape.size(), dataType, strides.data(), 0, aclFormat::ACL_FORMAT_ND,
shape.data(), shape.size(), *deviceAddr);
return 0;
}
struct Args {
uint32_t rankId;
HcclComm hcclComm;
aclrtStream stream;
aclrtContext context;
};
int launchOneThreadMatmulAllReduce(Args &args) {
int ret;
ret = aclrtSetCurrentContext(args.context);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSetCurrentContext failed. ERROR: %d\n", ret); return ret);
char hcom_name[128];
ret = HcclGetCommName(args.hcclComm, hcom_name);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("[ERROR] HcclGetCommName failed. ret = %d \n", ret); return -1);
LOG_PRINT("[INFO] rank %d hcom: %s stream: %p, context : %p\n", args.rankId, hcom_name, args.stream,
args.context);
std::vector<int64_t> x1Shape = {32, 64};
std::vector<int64_t> x2Shape = {64, 128};
std::vector<int64_t> biasShape = {128};
std::vector<int64_t> x3Shape = {32, 128};
std::vector<int64_t> outShape = {32, 128};
void *x1DeviceAddr = nullptr;
void *x2DeviceAddr = nullptr;
void *biasDeviceAddr = nullptr;
void *x3DeviceAddr = nullptr;
void *outDeviceAddr = nullptr;
aclTensor *x1 = nullptr;
aclTensor *x2 = nullptr;
aclTensor *bias = nullptr;
aclTensor *x3 = nullptr;
aclTensor *out = nullptr;
int64_t commTurn = 0;
int64_t streamMode = 1;
uint64_t workspaceSize = 0;
aclOpExecutor *executor;
void *workspaceAddr = nullptr;
long long x1ShapeSize = GetShapeSize(x1Shape);
long long x2ShapeSize = GetShapeSize(x2Shape);
long long biasShapeSize = GetShapeSize(biasShape);
long long x3ShapeSize = GetShapeSize(x3Shape);
long long outShapeSize = GetShapeSize(outShape);
std::vector<int16_t> x1HostData(x1ShapeSize, 1);
std::vector<int16_t> x2HostData(x2ShapeSize, 1);
std::vector<int16_t> biasHostData(biasShapeSize, 1);
std::vector<int16_t> x3HostData(x3ShapeSize, 1);
std::vector<int16_t> outHostData(outShapeSize, 0);
// 创建 tensor
ret = CreateAclTensor(x1HostData, x1Shape, &x1DeviceAddr, aclDataType::ACL_FLOAT16, &x1);
CHECK_RET(ret == ACL_SUCCESS, return ret);
ret = CreateAclTensor(x2HostData, x2Shape, &x2DeviceAddr, aclDataType::ACL_FLOAT16, &x2);
CHECK_RET(ret == ACL_SUCCESS, return ret);
ret = CreateAclTensor(biasHostData, biasShape, &biasDeviceAddr, aclDataType::ACL_FLOAT16, &bias);
CHECK_RET(ret == ACL_SUCCESS, return ret);
ret = CreateAclTensor(x3HostData, x3Shape, &x3DeviceAddr, aclDataType::ACL_FLOAT16, &x3);
CHECK_RET(ret == ACL_SUCCESS, return ret);
ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT16, &out);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 调用第一段接口, x3位置传入out
ret = aclnnMatmulAllReduceV2GetWorkspaceSize(x1, x2, bias, x3, hcom_name, "sum", commTurn, streamMode, out, &workspaceSize, &executor);
CHECK_RET(ret == ACL_SUCCESS,
LOG_PRINT("aclnnMatmulAllReduceV2GetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
// 根据第一段接口计算出的workspaceSize申请device内存
if (workspaceSize > 0) {
ret = aclrtMalloc(&workspaceAddr, workspaceSize, ACL_MEM_MALLOC_HUGE_FIRST);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("allocate workspace failed. ERROR: %d\n", ret); return ret);
}
// 调用第二段接口
ret = aclnnMatmulAllReduceV2(workspaceAddr, workspaceSize, executor, args.stream);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnMatmulAllReduceV2 failed. ERROR: %d\n", ret); return ret);
//(固定写法)同步等待任务执行结束
ret = aclrtSynchronizeStreamWithTimeout(args.stream, 10000);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);
LOG_PRINT("device%d aclnnMatmulAllReduceV2 execute success \n", args.rankId);
// 释放device资源,需要根据具体API的接口定义修改
if (x1 != nullptr) {
aclDestroyTensor(x1);
}
if (x2 != nullptr) {
aclDestroyTensor(x2);
}
if (bias != nullptr) {
aclDestroyTensor(bias);
}
if (x3 != nullptr) {
aclDestroyTensor(x3);
}
if (out != nullptr) {
aclDestroyTensor(out);
}
if (x1DeviceAddr != nullptr) {
aclrtFree(x1DeviceAddr);
}
if (x2DeviceAddr != nullptr) {
aclrtFree(x2DeviceAddr);
}
if (biasDeviceAddr != nullptr) {
aclrtFree(biasDeviceAddr);
}
if (x3DeviceAddr != nullptr) {
aclrtFree(x3DeviceAddr);
}
if (outDeviceAddr != nullptr) {
aclrtFree(outDeviceAddr);
}
if (workspaceSize > 0) {
aclrtFree(workspaceAddr);
}
aclrtDestroyStream(args.stream);
HcclCommDestroy(args.hcclComm);
aclrtDestroyContext(args.context);
aclrtResetDevice(args.rankId);
return 0;
}
int main(int argc, char *argv[]) {
int ret;
int32_t devices[ndev];
for (int i = 0; i < ndev; i++) {
devices[i] = i;
}
HcclComm comms[128];
ret = aclInit(nullptr);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclInit failed. ERROR: %d\n", ret); return ret);
// 初始化集合通信域
for (int i = 0; i < ndev; i++) {
ret = aclrtSetDevice(devices[i]);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSetDevice failed. ERROR: %d\n", ret); return ret);
}
ret = HcclCommInitAll(ndev, devices, comms);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("HcclCommInitAll failed. ERROR: %d\n", ret); return ret);
Args args[ndev];
aclrtStream stream[ndev];
aclrtContext context[ndev];
for (uint32_t rankId = 0; rankId < ndev; rankId++) {
ret = aclrtSetDevice(rankId);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSetDevice failed. ERROR: %d\n", ret); return ret);
ret = aclrtCreateContext(&context[rankId], rankId);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtCreateContext failed. ERROR: %d\n", ret); return ret);
ret = aclrtCreateStream(&stream[rankId]);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtCreateStream failed. ERROR: %d\n", ret); return ret);
}
// 启动多线程
std::vector<std::unique_ptr<std::thread>> threads(ndev);
for (uint32_t rankId = 0; rankId < ndev; rankId++) {
args[rankId].rankId = rankId;
args[rankId].hcclComm = comms[rankId];
args[rankId].stream = stream[rankId];
args[rankId].context = context[rankId];
threads[rankId].reset(new(std::nothrow) std::thread(&launchOneThreadMatmulAllReduce, std::ref(args[rankId])));
}
for (uint32_t rankId = 0; rankId < ndev; rankId++) {
threads[rankId]->join();
}
aclFinalize();
return 0;
}