doubango/plugins/pluginCUDA/plugin_cuda_utils.cxx

/* Copyright (C) 2013 Mamadou DIOP
* Copyright (C) 2013 Doubango Telecom <http://www.doubango.org>
*
* This file is part of Open Source Doubango Framework.
*
* DOUBANGO is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* DOUBANGO is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with DOUBANGO.
*/
#include "plugin_cuda_utils.h"

#include "tsk_debug.h"

#include <NVEncoderAPI.h>
#include <cuda.h>
#include <cuda_runtime_api.h>

bool CudaUtils::g_bStarted = false;
bool CudaUtils::g_bH264Checked = false;
bool CudaUtils::g_bH264Supported = false;
int CudaUtils::g_nCores = 0;

HRESULT CudaUtils::Startup()
{
    if(!g_bStarted) {
        CUresult cuResult = CUDA_SUCCESS;
        HRESULT hr = CoInitializeEx(NULL, COINIT_MULTITHREADED);
        if(SUCCEEDED(hr) || hr == 0x80010106) { // 0x80010106 when called from managed code (e.g. Boghe) - More info: http://support.microsoft.com/kb/824480
            if((cuResult = cuInit(0)) != CUDA_SUCCESS) {
                TSK_DEBUG_ERROR("cuInit() failed with error code = %08x", cuResult);
                hr = E_FAIL;
            }
            else {
                hr = S_OK;
            }
        }
        g_bStarted = true;
        return hr;
    }
    return S_OK;
}

HRESULT CudaUtils::Shutdown()
{
    // cuDeinit();
    return S_OK;
}

bool CudaUtils::IsH264Supported()
{
    if(g_bH264Checked) {
        return g_bH264Supported;
    }

    HRESULT hr = S_OK;

    CHECK_HR(hr = Startup());

    g_bH264Checked = true;

    NVEncoder pEncoder = NULL;

    CHECK_HR(hr = NVGetHWEncodeCaps());
    CHECK_HR(hr = NVCreateEncoder(&pEncoder));
    // Both Base and Main profiles *must* be supported
    CHECK_HR(hr = NVIsSupportedCodecProfile(pEncoder, NV_CODEC_TYPE_H264, NVVE_H264_PROFILE_BASELINE));
    CHECK_HR(hr = NVIsSupportedCodecProfile(pEncoder, NV_CODEC_TYPE_H264, NVVE_H264_PROFILE_MAIN));

    g_bH264Supported = true;

bail:
    if(pEncoder) {
        NVDestroyEncoder(pEncoder);
        pEncoder = NULL;
    }

    return g_bH264Supported;
}

int CudaUtils::ConvertSMVer2Cores(int nMajor, int nMinor)
{
    if(g_nCores != 0) {
        return g_nCores;
    }

    // Defines for GPU Architecture types (using the SM version to determine the # of cores per SM
    typedef struct {
        int SM; // 0xMm (hexidecimal notation), M = SM Major version, and m = SM minor version
        int Cores;
    } sSMtoCores;

    sSMtoCores nGpuArchCoresPerSM[] = {
        { 0x10,  8 }, // Tesla Generation (SM 1.0) G80 class
        { 0x11,  8 }, // Tesla Generation (SM 1.1) G8x class
        { 0x12,  8 }, // Tesla Generation (SM 1.2) G9x class
        { 0x13,  8 }, // Tesla Generation (SM 1.3) GT200 class
        { 0x20, 32 }, // Fermi Generation (SM 2.0) GF100 class
        { 0x21, 48 }, // Fermi Generation (SM 2.1) GF10x class
        { 0x30, 192}, // Kepler Generation (SM 3.0) GK10x class
        { 0x35, 192}, // Kepler Generation (SM 3.5) GK11x class
    };

    int index = 0;

    while (nGpuArchCoresPerSM[index].SM != -1) {
        if (nGpuArchCoresPerSM[index].SM == ((nMajor << 4) + nMinor)) {
            g_nCores = nGpuArchCoresPerSM[index].Cores;
            break;
        }

        index++;
    }

    // If we don't find the values, we default use the previous one to run properly
    TSK_DEBUG_INFO("MapSMtoCores for SM %d.%d is undefined.  Default to use %d Cores/SM", nMajor, nMinor, nGpuArchCoresPerSM[7].Cores);
    g_nCores = nGpuArchCoresPerSM[7].Cores;

    return g_nCores;
}

int CudaUtils::GetMaxGflopsDeviceId()
{
    int device_count = 0;
    cudaGetDeviceCount( &device_count );

    cudaDeviceProp device_properties;
    int max_gflops_device = 0;
    int max_gflops = 0;

    int current_device = 0;
    cudaGetDeviceProperties( &device_properties, current_device );
    max_gflops = device_properties.multiProcessorCount * device_properties.clockRate;
    ++current_device;

    while( current_device < device_count ) {
        cudaGetDeviceProperties( &device_properties, current_device );
        int gflops = device_properties.multiProcessorCount * device_properties.clockRate;
        if( gflops > max_gflops ) {
            max_gflops        = gflops;
            max_gflops_device = current_device;
        }
        ++current_device;
    }

    return max_gflops_device;
}