欢迎您访问程序员文章站本站旨在为大家提供分享程序员计算机编程知识!
您现在的位置是: 首页

NVIDIA CUDA clock.cu 代码  

程序员文章站 2022-06-04 16:05:24
...

//VC++ 2010 + CUDA 4.1


#ifndef _CLOCK_KERNEL_H_
#define _CLOCK_KERNEL_H_

// This kernel computes a standard parallel reduction and evaluates the
// time it takes to do that for each block. The timing results are stored
// in device memory.
__global__ static void timedReduction(const float * input, float * output, clock_t * timer)
{
// __shared__ float shared[2 * blockDim.x];
extern __shared__ float shared[];

const int tid = threadIdx.x;
const int bid = blockIdx.x;

if (tid == 0) timer[bid] = clock();

// Copy input.
shared[tid] = input[tid];
shared[tid + blockDim.x] = input[tid + blockDim.x];

// Perform reduction to find minimum.
for(int d = blockDim.x; d > 0; d /= 2)
{
__syncthreads();

if (tid < d)
{
float f0 = shared[tid];
float f1 = shared[tid + d];

if (f1 < f0) {
shared[tid] = f1;
}
}
}

// Write result.
if (tid == 0) output[bid] = shared[0];

__syncthreads();

if (tid == 0) timer[bid+gridDim.x] = clock();
}

#endif // _CLOCK_KERNEL_H_





#include <stdio.h>
#include <stdlib.h>

#include <shrQATest.h>
#include <cutil_inline.h>

#include "clock_kernel.cu"

// This example shows how to use the clock function to measure the performance of
// a kernel accurately.
//
// Blocks are executed in parallel and out of order. Since there's no synchronization
// mechanism between blocks, we measure the clock once for each block. The clock
// samples are written to device memory.

#define NUM_BLOCKS 64
#define NUM_THREADS 256

// It's interesting to change the number of blocks and the number of threads to
// understand how to keep the hardware busy.
//
// Here are some numbers I get on my G80:
// blocks - clocks
// 1 - 3096
// 8 - 3232
// 16 - 3364
// 32 - 4615
// 64 - 9981
//
// With less than 16 blocks some of the multiprocessors of the device are idle. With
// more than 16 you are using all the multiprocessors, but there's only one block per
// multiprocessor and that doesn't allow you to hide the latency of the memory. With
// more than 32 the speed scales linearly.

int main(int argc, char** argv)
{
shrQAStart(argc, argv);

// use command-line specified CUDA device, otherwise use device with highest Gflops/s
if ( cutCheckCmdLineFlag(argc, (const char **)argv, "device")) {
int devID = cutilDeviceInit(argc, argv);
if (devID < 0) {
printf("No CUDA Capable devices found, exiting...\n");
shrQAFinishExit(argc, (const char **)argv, QA_WAIVED);
}
} else {
cudaSetDevice( cutGetMaxGflopsDeviceId() );
}

float * dinput = NULL;
float * doutput = NULL;
clock_t * dtimer = NULL;

clock_t timer[NUM_BLOCKS * 2];
float input[NUM_THREADS * 2];

for (int i = 0; i < NUM_THREADS * 2; i++)
{
input[i] = (float)i;
}

cutilSafeCall(cudaMalloc((void**)&dinput, sizeof(float) * NUM_THREADS * 2));
cutilSafeCall(cudaMalloc((void**)&doutput, sizeof(float) * NUM_BLOCKS));
cutilSafeCall(cudaMalloc((void**)&dtimer, sizeof(clock_t) * NUM_BLOCKS * 2));

cutilSafeCall(cudaMemcpy(dinput, input, sizeof(float) * NUM_THREADS * 2, cudaMemcpyHostToDevice));

timedReduction<<<NUM_BLOCKS, NUM_THREADS, sizeof(float) * 2 * NUM_THREADS>>>(dinput, doutput, dtimer);

//cutilSafeCall(cudaMemcpy(output, doutput, sizeof(float) * NUM_BLOCKS, cudaMemcpyDeviceToHost));
cutilSafeCall(cudaMemcpy(timer, dtimer, sizeof(clock_t) * NUM_BLOCKS * 2, cudaMemcpyDeviceToHost));

cutilSafeCall(cudaFree(dinput));
cutilSafeCall(cudaFree(doutput));
cutilSafeCall(cudaFree(dtimer));


// Compute the difference between the last block end and the first block start.
clock_t minStart = timer[0];
clock_t maxEnd = timer[NUM_BLOCKS];

for (int i = 1; i < NUM_BLOCKS; i++)
{
minStart = timer[i] < minStart ? timer[i] : minStart;
maxEnd = timer[NUM_BLOCKS+i] > maxEnd ? timer[NUM_BLOCKS+i] : maxEnd;
}

printf("time = %d\n", maxEnd - minStart);

cutilDeviceReset();

// This test always passes.
shrQAFinishExit(argc, (const char **)argv, QA_PASSED);
}