Shulou(Shulou.com)06/01 Report--

This article is about how to implement CUDA programming under Ubuntu 9.04. The editor thinks it is very practical, so share it with you as a reference and follow the editor to have a look.

The instances of CUDA SDK are in the src directory, and each instance has its own directory, such as deviceuery, and in its directory there is a compile-time Makefile file that compiles a single project. Now we compile all the instances, and after running sudo make under the CUDA_SDK root, you can see the compiled executable under / bin/linux/release, and you can see the results.

This is the result of the operation of deviceQuery:

So I believe readers should have thought that we can use these examples to create our own projects. In another example, there is a template, delete the .cu, .cpp files in the src under this directory, and empty the contents in the obj directory, which becomes an empty CUDA project. You can write the program under src, and then modify the compiled file name in Makefie to compile. The generated execution file is under CUDA_SDK_HOME/bin/linux/release. Here is a test code that performs matrix addition:

1 # include

2 # include

3 # include

4 # include

5 # include

six

7 # define VEC_SIZE 16

eight

9 / / kernel function

10 _ _ global__ void vecAdd (float* dudes A float * dcards Bline float * dudes C)

11 {

12 int index=threadIdx.x

13 DC [index] = DA [index] + dB [index]

14}

fifteen

16 int main ()

17 {

18 / / get the size of the allocated space

19 size_t size=VEC_SIZE*sizeof (float)

twenty

21 / / allocate memory locally

22 float* hype A = (float*) malloc (size)

23 float* hype B = (float*) malloc (size)

24 float* hype C = (float*) malloc (size)

twenty-five

26 / / initialization

27 for (int idol.

28 {

29hA [I] = 1.0

30 h _ B [I] = 2.0

31}

thirty-two

33 / / copy the data from local memory to the device

34 float* d_A

35 cudaMalloc ((void**) & dumbAMagneSize)

36 cudaMemcpy (dudaMemcpyHostToDevice)

thirty-seven

38 float* d_B

39 cudaMalloc ((void**) & dumbbard size)

40 cudaMemcpy (daddy BMagne hobby BJM SizeMagol cudaMemcpyHostToDevice)

forty-one

42 / / allocate the space to store the results

43 float* d_C

44 cudaMalloc ((void**) & dazzling CJournal size)

forty-five

46 / / define 16 threads

47 dim3 dimblock (16)

48 vecAdd (dumbA, dandy, dhob, daddy, etc.)

forty-nine

50 / / copy the calculation results back to main memory

51 cudaMemcpy (hobbies, memcpyDeviceToHost)

fifty-two

53 / / output calculation results

54 for (int juni0bot j

55 {

56 printf ("% f\ t", hC [j])

57}

fifty-eight

59 / / release host and device memory

60 cudaFree (dumbA)

61 cudaFree (dumbb)

62 cudaFree (dumbC)

sixty-three

64 free (hype A)

65 free (hobby B)

66 free (hype C)

sixty-seven

68 return 0

69}

Thank you for reading! This is the end of this article on "how to achieve CUDA programming under Ubuntu 9.04". I hope the above content can be of some help to you, so that you can learn more knowledge. if you think the article is good, you can share it for more people to see!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.