英文:
Jcuda CUDA_ERROR_LAUNCH_OUT_OF_RESOURCES (Too Many Parameters?)
问题
Pointer kernelParams = Pointer.to(
Pointer.to(new int[] {n}),
Pointer.to(new int[] {xRes}),
Pointer.to(new int[] {yRes}),
Pointer.to(new double[] {camX}),
Pointer.to(new double[] {camY}),
Pointer.to(new double[] {camZ}),
Pointer.to(new double[] {camforX}),
Pointer.to(new double[] {camforY}),
Pointer.to(new double[] {camforZ}),
Pointer.to(new double[] {camupX}),
Pointer.to(new double[] {camupY}),
Pointer.to(new double[] {camupZ}),
Pointer.to(new double[] {fov}),
Pointer.to(new double[] {aspectRatio}),
Pointer.to(in14),
Pointer.to(in15),
Pointer.to(in16),
Pointer.to(in17),
Pointer.to(in18),
Pointer.to(in19),
Pointer.to(in20),
Pointer.to(in21),
Pointer.to(in22),
Pointer.to(new double[] {meshposX}),
Pointer.to(new double[] {meshposY}),
Pointer.to(new double[] {meshposZ}),
Pointer.to(out0),
Pointer.to(out1),
Pointer.to(out2)
);
cuLaunchKernel(function,
32, 32, 1,
32, 32, 1,
0, null,
kernelParams, null
);
英文:
When I call the function, cuLaunchKernel(), my program crashes with the error, CUDA_ERROR_LAUNCH_OUT_OF_RESOURCES.
Is this because I'm passing 29 parameters to my kernel?
If so, is there any work around? I am pretty certain that I need these parameters for calculations.
Pointer kernelParams = Pointer.to(
Pointer.to(new int[] {n}),
Pointer.to(new int[] {xRes}),
Pointer.to(new int[] {yRes}),
Pointer.to(new double[] {camX}),
Pointer.to(new double[] {camY}),
Pointer.to(new double[] {camZ}),
Pointer.to(new double[] {camforX}),
Pointer.to(new double[] {camforY}),
Pointer.to(new double[] {camforZ}),
Pointer.to(new double[] {camupX}),
Pointer.to(new double[] {camupY}),
Pointer.to(new double[] {camupZ}),
Pointer.to(new double[] {fov}),
Pointer.to(new double[] {aspectRatio}),
Pointer.to(in14),
Pointer.to(in15),
Pointer.to(in16),
Pointer.to(in17),
Pointer.to(in18),
Pointer.to(in19),
Pointer.to(in20),
Pointer.to(in21),
Pointer.to(in22),
Pointer.to(new double[] {meshposX}),
Pointer.to(new double[] {meshposY}),
Pointer.to(new double[] {meshposZ}),
Pointer.to(out0),
Pointer.to(out1),
Pointer.to(out2)
);
cuLaunchKernel(function,
32,32,1,
32,32,1,
0,null,
kernelParams,null
);
答案1
得分: 1
这是我的天真回答：
正如评论中建议的那样，问题与块大小有关。减小块大小可以解决错误。在减小块大小后，我收到了错误消息：CUDA_ERROR_ILLEGAL_ADDRESS。减小网格大小解决了那个问题。
英文:
Here is my naive answer:
As suggested in comments, the problem was to do with the block size. Reducing the block size fixed the error. After reducing the block size, I received the error: CUDA_ERROR_ILLEGAL_ADDRESS. Reducing the grid size solved that problem.
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库，让每个人都能够通过互相帮助和分享经验来进步。
评论