Using a pinned memory in OpenCL

[alhussain779]

Hello,

I am using  Intel® FPGA SDK for OpenCL. The board I am using is from Terasic DE10-PRO SX.

In my application, I produce data to a buffer in the memory. Then I send this data to the FPGA using the function clEnqueueWriteBuffer and process it there. After that, I send the results back to the host using clEnqueueReadBuffer. I would like to enhance the process of sending the data back and forth between the host machine and the FPGA.

My understanding of the function clEnqueueWriteBuffer is that it copies the data from the provided pointer to a pinned memory, then sends the data from the pinned memory to device. Is this correct?

If yes, is there a way to write data directly to a pinned memory?

In addition,  is there a better way to transfer data between the FPGA and host machine?

Best regards,

Alhussain

[aikeu]

Hi alhussain779,

I will get back to you when I have relevant info to your queston.

Thanks.

Regards,

Aik Eu

[alhussain779]

Hi @aikeu ,

Thanks for your reply. I would appreciate your help

Regards,

Alhussain

[aikeu]

Hi alhussain779,

For better way to transfer data, would suggest perhaps to try burst transfer on global memory (can refer to section 8.2) OR using non blocking channels (can refer to section 10.1.3), for more details in the link below

https://www.intel.com/content/www/us/en/docs/programmable/683521/22-4/introduction-to-pro-edition-best-practices.html

Thanks.

Regards,

Aik Eu

[aikeu]

Hi alhussain779,

Do let me know if you have further question from my previous reply.

Thanks.

Regards,

Aik Eu

[aikeu]

Hi alhussain779,

We do not receive any response from you to the previous question/reply/answer that I have provided. This thread will be transitioned to community support. If you have a new question, feel free to open a new thread to get the support from Intel experts. Otherwise, the community users will continue to help you on this thread. Thank you

Thanks.

Regards,

Aik Eu