Migrating to SYCL
One-stop forum for getting assistance migrating your existing code to SYCL
43 Discussions

Is DPCT able to migrate <<<...>>> syntax in CUDA?

Damon
Employee
1,132 Views

I did the migration on https://github.com/dusty-nv/jetson-utils/blob/master/cuda/cudaWarp-affine.cu using DPCT.

 

Found that it cannot migrate the kernel with  <<<...>>> syntax in CUDA.

 

Just convert from

// launch kernel
const dim3 blockDim(8, 8);
const dim3 gridDim(iDivUp(width,blockDim.x), iDivUp(height,blockDim.y));

gpuPerspectiveWarp<<<gridDim, blockDim>>>(input, output, width, height,cuda_mat[0],                     cuda_mat[1], cuda_mat[2]);

 

to 

 // launch kernel
 const sycl::range<3> blockDim(1, 8, 8);
 const sycl::range<3> gridDim(1, iDivUp(height, blockDim[1]),
                                     iDivUp(width, blockDim[2]));

  gpuPerspectiveWarp<<<gridDim, blockDim>>>(input, output, width, height,
                                              cuda_mat[0], cuda_mat[1], cuda_mat[2]);
 
My question is whether the dpct could migrate the gpuPerspectiveWarp line or not?
 
Thanks
0 Kudos
1 Solution
Damon
Employee
1,118 Views

Forget this TOPIC.

I'm new but got the way to migrate it correctly.

 

Sorry.

View solution in original post

0 Kudos
2 Replies
Damon
Employee
1,119 Views

Forget this TOPIC.

I'm new but got the way to migrate it correctly.

 

Sorry.

0 Kudos
ShwethaS_Intel
Moderator
1,096 Views

Hi,

Thanks for reaching out to us.


Glad to know that your issue is resolved. If you need any

additional information, please post a new question as this

thread will no longer be monitored by Intel.


Best Regards,

Shwetha




0 Kudos
Reply