added pure HC gpu stream implmentation

This commit is contained in:
Peter Steinbach 2017-01-03 11:43:12 +01:00
parent edd65dacb1
commit 7621f86701
4 changed files with 292 additions and 0 deletions

View File

@ -10,6 +10,15 @@ if(NOT DEFINED HIP_PATH)
endif()
set(CMAKE_MODULE_PATH "${HIP_PATH}/cmake" ${CMAKE_MODULE_PATH})
if(NOT DEFINED HCC_PATH)
if(NOT DEFINED ENV{HCC_PATH})
set(HCC_PATH "/opt/rocm/" CACHE PATH "Path to which HCC has been installed")
else()
set(HCC_PATH $ENV{HCC_PATH} CACHE PATH "Path to which HCC has been installed")
endif()
endif()
set(CMAKE_CXX_STANDARD 11)
set(CMAKE_CXX_STANDARD_REQUIRED ON)
@ -49,6 +58,26 @@ else()
message("Skipping HIP...")
endif()
#-------------------------------------------------------------------------------
# HCC
#-------------------------------------------------------------------------------
find_program(HCC_BINARY hcc HINTS ${HCC_PATH}/bin PATHS ${HCC_PATH}/bin)
if(EXISTS ${HCC_BINARY})
#can the following be tied to the target only?
set(CMAKE_CXX_COMPILER ${HCC_BINARY})
# list(APPEND CMAKE_CXX_FLAGS -hc -I /usr/include/c++/v1 -std=c++11 -stdlib=libc++ -I/opt/rocm/hcc-lc/include)
# list(APPEND CMAKE_EXE_LINKER_FLAGS -hc -std=c++11 -L/opt/rocm/hcc-lc/lib -Wl,--rpath=/opt/rocm/hcc-lc/lib -lc++ -lc++abi -ldl -lpthread -Wl,--whole-archive -lmcwamp -Wl,--no-whole-archive -mcmodel=small )
add_executable(gpu-stream-hc main.cpp HCStream.cpp)
set_target_properties(gpu-stream-hc PROPERTIES
COMPILE_FLAGS "-hc -I /usr/include/c++/v1 -std=c++11 -stdlib=libc++ -I/opt/rocm/hcc-lc/include"
LINK_FLAGS "-hc -std=c++11 -L/opt/rocm/hcc-lc/lib -Wl,--rpath=/opt/rocm/hcc-lc/lib -lc++ -lc++abi -ldl -lpthread -Wl,--whole-archive -lmcwamp -Wl,--no-whole-archive -mcmodel=small"
)
target_compile_definitions(gpu-stream-hc PUBLIC HC)
else()
message("Skipping HC...")
endif()
#-------------------------------------------------------------------------------
# CUDA
#-------------------------------------------------------------------------------

213
HCStream.cpp Normal file
View File

@ -0,0 +1,213 @@
// Copyright (c) 2015-16 Peter Steinbach, MPI CBG Scientific Computing Facility
//
// For full license terms please see the LICENSE file distributed with this
// source code
#include <codecvt>
#include <vector>
#include <locale>
#include "HCStream.h"
//#include "hc.hpp"
#define TBSIZE 1024
std::string getDeviceName(const hc::accelerator& _acc)
{
std::wstring_convert<std::codecvt_utf8<wchar_t>, wchar_t> converter;
std::string value = converter.to_bytes(_acc.get_description());
return value;
}
void listDevices(void)
{
// Get number of devices
std::vector<hc::accelerator> accs = hc::accelerator::get_all();
// Print device names
if (accs.empty())
{
std::cerr << "No devices found." << std::endl;
}
else
{
std::cout << std::endl;
std::cout << "Devices:" << std::endl;
for (int i = 0; i < accs.size(); i++)
{
std::cout << i << ": " << getDeviceName(accs[i]) << std::endl;
}
std::cout << std::endl;
}
}
// void check_error(void)
// {
// hipError_t err = hipGetLastError();
// if (err != hipSuccess)
// {
// std::cerr << "Error: " << hipGetErrorString(err) << std::endl;
// exit(err);
// }
// }
template <class T>
HCStream<T>::HCStream(const unsigned int ARRAY_SIZE, const int device_index):
array_size(ARRAY_SIZE),
d_a(ARRAY_SIZE),
d_b(ARRAY_SIZE),
d_c(ARRAY_SIZE)
{
// The array size must be divisible by TBSIZE for kernel launches
if (ARRAY_SIZE % TBSIZE != 0)
{
std::stringstream ss;
ss << "Array size must be a multiple of " << TBSIZE;
throw std::runtime_error(ss.str());
}
// // Set device
std::vector<hc::accelerator> accs = hc::accelerator::get_all();
auto current = accs[device_index];
std::cout << "Using HC device " << getDeviceName(current) << std::endl;
// // The array size must be divisible by TBSIZE for kernel launches
// if (ARRAY_SIZE % TBSIZE != 0)
// {
// std::stringstream ss;
// ss << "Array size must be a multiple of " << TBSIZE;
// throw std::runtime_error(ss.str());
// }
// // Set device
// int count;
// hipGetDeviceCount(&count);
// check_error();
// if (device_index >= count)
// throw std::runtime_error("Invalid device index");
// hipSetDevice(device_index);
// check_error();
// // Print out device information
// std::cout << "Using HIP device " << getDeviceName(device_index) << std::endl;
// std::cout << "Driver: " << getDeviceDriver(device_index) << std::endl;
// array_size = ARRAY_SIZE;
// // Check buffers fit on the device
// hipDeviceProp_t props;
// hipGetDeviceProperties(&props, 0);
// if (props.totalGlobalMem < 3*ARRAY_SIZE*sizeof(T))
// throw std::runtime_error("Device does not have enough memory for all 3 buffers");
// // Create device buffers
// hipMalloc(&d_a, ARRAY_SIZE*sizeof(T));
// check_error();
// hipMalloc(&d_b, ARRAY_SIZE*sizeof(T));
// check_error();
// hipMalloc(&d_c, ARRAY_SIZE*sizeof(T));
// check_error();
}
template <class T>
HCStream<T>::~HCStream()
{
}
template <class T>
void HCStream<T>::write_arrays(const std::vector<T>& a, const std::vector<T>& b, const std::vector<T>& c)
{
hc::copy(a.cbegin(),a.cend(),d_a);
hc::copy(b.cbegin(),b.cend(),d_b);
hc::copy(c.cbegin(),c.cend(),d_c);
}
template <class T>
void HCStream<T>::read_arrays(std::vector<T>& a, std::vector<T>& b, std::vector<T>& c)
{
// Copy device memory to host
hc::copy(d_a,a.begin());
hc::copy(d_b,b.begin());
hc::copy(d_c,c.begin());
}
template <class T>
void HCStream<T>::copy()
{
try{
// launch a GPU kernel to compute the saxpy in parallel
hc::completion_future future_kernel = hc::parallel_for_each(hc::extent<1>(array_size)
, [&](hc::index<1> i) [[hc]] {
d_c[i] = d_a[i];
});
future_kernel.wait();
}
catch(std::exception& e){
std::cout << e.what() << std::endl;
throw;
}
}
template <class T>
void HCStream<T>::mul()
{
const T scalar = 0.3;
try{
// launch a GPU kernel to compute the saxpy in parallel
hc::completion_future future_kernel = hc::parallel_for_each(hc::extent<1>(array_size)
, [&](hc::index<1> i) [[hc]] {
d_b[i] = scalar*d_c[i];
});
future_kernel.wait();
}
catch(std::exception& e){
std::cout << e.what() << std::endl;
throw;
}
}
template <class T>
void HCStream<T>::add()
{
try{
// launch a GPU kernel to compute the saxpy in parallel
hc::completion_future future_kernel = hc::parallel_for_each(hc::extent<1>(array_size)
, [&](hc::index<1> i) [[hc]] {
d_c[i] = d_a[i]+d_b[i];
});
future_kernel.wait();
}
catch(std::exception& e){
std::cout << e.what() << std::endl;
throw;
}
}
template <class T>
void HCStream<T>::triad()
{
const T scalar = 0.3;
try{
// launch a GPU kernel to compute the saxpy in parallel
hc::completion_future future_kernel = hc::parallel_for_each(hc::extent<1>(array_size)
, [&](hc::index<1> i) [[hc]] {
d_a[i] = d_b[i] + scalar*d_c[i];
});
future_kernel.wait();
}
catch(std::exception& e){
std::cout << e.what() << std::endl;
throw;
}
}
template class HCStream<float>;
template class HCStream<double>;

44
HCStream.h Normal file
View File

@ -0,0 +1,44 @@
// Copyright (c) 2015-16 Tom Deakin, Simon McIntosh-Smith,
// University of Bristol HPC
//
// For full license terms please see the LICENSE file distributed with this
// source code
#pragma once
#include <iostream>
#include <stdexcept>
#include <sstream>
#include "Stream.h"
#include "hc.hpp"
#define IMPLEMENTATION_STRING "HC"
template <class T>
class HCStream : public Stream<T>
{
protected:
// Size of arrays
unsigned int array_size;
// Device side pointers to arrays
hc::array<T,1> d_a;
hc::array<T,1> d_b;
hc::array<T,1> d_c;
public:
HCStream(const unsigned int, const int);
~HCStream();
virtual void copy() override;
virtual void add() override;
virtual void mul() override;
virtual void triad() override;
virtual void write_arrays(const std::vector<T>& a, const std::vector<T>& b, const std::vector<T>& c) override;
virtual void read_arrays(std::vector<T>& a, std::vector<T>& b, std::vector<T>& c) override;
};

View File

@ -22,6 +22,8 @@
#include "CUDAStream.h"
#elif defined(HIP)
#include "HIPStream.h"
#elif defined(HC)
#include "HCStream.h"
#elif defined(OCL)
#include "OCLStream.h"
#elif defined(USE_RAJA)
@ -105,6 +107,10 @@ void run()
// Use the HIP implementation
stream = new HIPStream<T>(ARRAY_SIZE, deviceIndex);
#elif defined(HC)
// Use the HC implementation
stream = new HCStream<T>(ARRAY_SIZE, deviceIndex);
#elif defined(OCL)
// Use the OpenCL implementation
stream = new OCLStream<T>(ARRAY_SIZE, deviceIndex);