NVidia OpenCL 2 beta nested parallelism

Accelerated Computing CUDA CUDA Programming and Performance
1

Hi,
I have experimented with the NVidia OpenCL 2 beta.
When trying to launch kernel from within kernel(aka nested parallelism), the GPU hangs and then the host receives an error of CL_OUT_OF_RESOURCES from the device.

Has anyone experimented/ran/used this feature in the OpenCL 2 beta version?

Regards,
Tomer Gal, CTO at OpTeamizer (Israel)

2

Found the answer in the release notes:
“• The current implementation is limited to 64-bit platforms only”

This specific application I am working on, is unfortunately, 32 bit.

Regards,
Tomer Gal, CTO at OpTeamizer