topic Ty for your fast reply. in OpenCL* for CPU

Questions about Intel's opencl platform

Nikos_B_ — Fri, 05 Jun 2015 23:47:03 GMT

I'm new to Intel's opencl platform, but have read good things about its performance.

I need to know if I am compatible with your platform, b4 venturing in too deep.

I use Ubuntu 14.04 x64. My card is a radeon Curacao Pro R9 270 2 GB, and my CPU is AMD FX 8320 8-core with 8 GB memory. Unfortunately, I cannot find a working opencl debugger from AMD (CodeXL 1.7 doesn't work and i cannot find earlier releases). So I was wandering if your platform could generate code for this setup, and if Code Builder could debug it.

Ubuntu is an unsupported

Robert_I_Intel — Fri, 05 Jun 2015 23:53:11 GMT

Ubuntu is an unsupported platform, but you maybe able to install Intel's OpenCL runtime for the CPU, as other people succeeded doing that. As to whether it would work on AMD processors, again I am not sure, since this is an unsupported platform :) It will definitely not utilize your GPU: for that you need to go to AMD website.

Ty for your fast reply.

Nikos_B_ — Sat, 06 Jun 2015 00:08:11 GMT

Ty for your fast reply.

I have to use Linux for my work. I have read that Intel supports CentOS. I definitely need the GPU. I probably would have to buy an Intel card for that. Which one would you recommend similar to the radeon one i already have? I need it for DSP, real-time spectrum analyzer.

CentOS or SLES are supported

Robert_I_Intel — Sat, 06 Jun 2015 00:47:13 GMT

CentOS or SLES are supported platforms. Intel does not have a discrete GPU, what we have is an accelerator Xeon Phi:

http://www.intel.com/content/www/us/en/processors/xeon/xeon-phi-detail.html

I am not familiar with Xeon Phi product line and how it compares to the competition: you will need to research this yourself.

Interesting. Intel uses HPC,

Nikos_B_ — Sat, 06 Jun 2015 08:02:36 GMT

Interesting. Intel uses HPC, while AMD uses HSA. I could live with a Xeon Phi, although it represents a departure from what I am doing. I am concerned to use another general purpose processor, over many discrete number-crunching GPUs. Intel's performance metrics show a benefit of ~3.5x over a 2 CPU setup. That would roughly mean a max 7x benefit over a single multi-core CPU. I have clocked my application running 10x faster than running from my 8-core CPU alone.

I am not sure about the performance with nvidia cards. I imagine that the metrics with nvidia cards, use nvidia's openCL.

However, the lack of AMD's debugging environment is critical. It is good to have more options. I will have to think it over.

Thanks again.