• State Accepted Answer
  • Locked Locked
  • Replies 5 replies
  • Subscribers 136 subscribers
  • Views 6103 views
  • Users 0 members are here
Options
Related
How was your experience today?
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

OpenCL doubleN shuffle in MaliT604

Diego

Hello all,

I am working in Mali T604 and I am using double precision. I enabled this extension with:

#pragma OPENCL EXTENSION cl_khr_fp64 : enable

All double precision operations work and vector double precision operations also work but shuffle function is unknown for double.

    uint4 mask = (uint4)(3, 2, 1, 0);
    double4 a;
    double4 r = shuffle(a, mask);

Build error

If I do the same with single precision there is no problem.

    uint4 mask = (uint4)(3, 2, 1, 0);
    float4 a;If I do the same with single precision there is no problem.
    float4 r = shuffle(a, mask);

Thanks!

Parents
  • 0

    Hi Diego,

    The OpenCL 1.1 Spec states the following:

    The size of each element in the mask must match the size of each element in the result.

    What this means is that in the case for the 'double', your mask should instead be of type 'ulong4'

    Can you please try with this change and let us know if it fixes the problem?

    Kind Regards,

    Michael McGeagh

Reply
  • 0

    Hi Diego,

    The OpenCL 1.1 Spec states the following:

    The size of each element in the mask must match the size of each element in the result.

    What this means is that in the case for the 'double', your mask should instead be of type 'ulong4'

    Can you please try with this change and let us know if it fixes the problem?

    Kind Regards,

    Michael McGeagh

Children