《CUDA By Example》【Chapter 03】CUDA C简介

3.2 第一个CUDA程序

simple_kernel.cu

#include "../common/book.h"

__global__ void kernel( void ) {
}

int main( void ) {
    kernel<<<1,1>>>(); //不是传给设备代码的参数，而是告诉运行时如何启动代码
    printf( "Hello, World!\n" );
    return 0;
}

simple_kernel_params.cu

#include "../common/book.h"

__global__ void add( int a, int b, int *c ) {
    *c = a + b;
}

int main( void ) {
    int c;
    int *dev_c;
    HANDLE_ERROR( cudaMalloc( (void**)&dev_c, sizeof(int) ) );
    //第一个参数，待分配内存地址的变量；
    //HANDLE_ERROR为本书自定义的宏，判断函数调用是否返回了错误值；

    add<<<1,1>>>( 2, 7, dev_c );

    HANDLE_ERROR( cudaMemcpy( &c, dev_c, sizeof(int),
                              cudaMemcpyDeviceToHost ) );
    //因为主机端不能读写cudaMalloc()返回的指针内存，所以主机和设备间通信的数据通过专用的命令cudaMemcpy()来实现；
    printf( "2 + 7 = %d\n", c );
    HANDLE_ERROR( cudaFree( dev_c ) );

    return 0;
}

CUDA的简单和强大，很大来源于淡化了主机代码和设备代码的差异。然而，一定不能在主机代码中对cudaMalloc（）返回的指针进行解引用。主机代码可以将这个指针作为参数传递，但是不可以用这个指针读取和写入内存。

设备指针的限制：
- 可以将cudaMalloc()分配的指针传递给设备上执行的函数 可以在设备代码
- 使用cudaMalloc()分配的指针进行内存读/写操作
- 可以将cudaMalloc()分配的指针传递给在主机上执行的函数

3.3 查询设备

enum_gpu.cu

#include "../common/book.h"

int main( void ) {
    cudaDeviceProp  prop;

    int count;
    HANDLE_ERROR( cudaGetDeviceCount( &count ) );
    for (int i=0; i< count; i++) {
        HANDLE_ERROR( cudaGetDeviceProperties( &prop, i ) );
        printf( "   --- General Information for device %d ---\n", i );
        printf( "Name:  %s\n", prop.name );
        printf( "Compute capability:  %d.%d\n", prop.major, prop.minor );
        printf( "Clock rate:  %d\n", prop.clockRate );
        printf( "Device copy overlap:  " );
        if (prop.deviceOverlap)
            printf( "Enabled\n" );
        else
            printf( "Disabled\n");
        printf( "Kernel execution timeout :  " );
        if (prop.kernelExecTimeoutEnabled)
            printf( "Enabled\n" );
        else
            printf( "Disabled\n" );

        printf( "   --- Memory Information for device %d ---\n", i );
        printf( "Total global mem:  %ld\n", prop.totalGlobalMem );
        printf( "Total constant Mem:  %ld\n", prop.totalConstMem );
        printf( "Max mem pitch:  %ld\n", prop.memPitch );
        printf( "Texture Alignment:  %ld\n", prop.textureAlignment );

        printf( "   --- MP Information for device %d ---\n", i );
        printf( "Multiprocessor count:  %d\n",
                    prop.multiProcessorCount );
        printf( "Shared mem per mp:  %ld\n", prop.sharedMemPerBlock );
        printf( "Registers per mp:  %d\n", prop.regsPerBlock );
        printf( "Threads in warp:  %d\n", prop.warpSize );
        printf( "Max threads per block:  %d\n",
                    prop.maxThreadsPerBlock );
        printf( "Max thread dimensions:  (%d, %d, %d)\n",
                    prop.maxThreadsDim[0], prop.maxThreadsDim[1],
                    prop.maxThreadsDim[2] );
        printf( "Max grid dimensions:  (%d, %d, %d)\n",
                    prop.maxGridSize[0], prop.maxGridSize[1],
                    prop.maxGridSize[2] );
        printf( "\n" );
    }
}

   --- General Information for device 0 ---
Name:  Tesla P100-PCIE-16GB
Compute capability:  6.0
Clock rate:  1328500
Device copy overlap:  Enabled
Kernel execution timeout :  Disabled
   --- Memory Information for device 0 ---
Total global mem:  17071734784
Total constant Mem:  65536
Max mem pitch:  2147483647
Texture Alignment:  512
   --- MP Information for device 0 ---
Multiprocessor count:  56
Shared mem per mp:  49152
Registers per mp:  65536
Threads in warp:  32
Max threads per block:  1024
Max thread dimensions:  (1024, 1024, 64)
Max grid dimensions:  (2147483647, 65535, 65535)

   --- General Information for device 1 ---
Name:  Tesla P100-PCIE-16GB
Compute capability:  6.0
Clock rate:  1328500
Device copy overlap:  Enabled
Kernel execution timeout :  Disabled
   --- Memory Information for device 1 ---
Total global mem:  17071734784
Total constant Mem:  65536
Max mem pitch:  2147483647
Texture Alignment:  512
   --- MP Information for device 1 ---
Multiprocessor count:  56
Shared mem per mp:  49152
Registers per mp:  65536
Threads in warp:  32
Max threads per block:  1024
Max thread dimensions:  (1024, 1024, 64)
Max grid dimensions:  (2147483647, 65535, 65535)

   --- General Information for device 2 ---
Name:  Tesla P100-PCIE-16GB
Compute capability:  6.0
Clock rate:  1328500
Device copy overlap:  Enabled
Kernel execution timeout :  Disabled
   --- Memory Information for device 2 ---
Total global mem:  17071734784
Total constant Mem:  65536
Max mem pitch:  2147483647
Texture Alignment:  512
   --- MP Information for device 2 ---
Multiprocessor count:  56
Shared mem per mp:  49152
Registers per mp:  65536
Threads in warp:  32
Max threads per block:  1024
Max thread dimensions:  (1024, 1024, 64)
Max grid dimensions:  (2147483647, 65535, 65535)

   --- General Information for device 3 ---
Name:  Tesla P100-PCIE-16GB
Compute capability:  6.0
Clock rate:  1328500
Device copy overlap:  Enabled
Kernel execution timeout :  Disabled
   --- Memory Information for device 3 ---
Total global mem:  17071734784
Total constant Mem:  65536
Max mem pitch:  2147483647
Texture Alignment:  512
   --- MP Information for device 3 ---
Multiprocessor count:  56
Shared mem per mp:  49152
Registers per mp:  65536
Threads in warp:  32
Max threads per block:  1024
Max thread dimensions:  (1024, 1024, 64)
Max grid dimensions:  (2147483647, 65535, 65535)

3.4 设备属性的使用