Intel Announces Knights Mill: A Xeon Phi for Deep Learning

Re:Intel Announces Knights Mill: A Xeon Phi for Deep Learning by marmaduke on Thursday August 18, 2016 @03:01AM (#12310836)

Nope. I have some substantial simulation code written against OpenCL that runs on Intel OpenCL and NVIDIA without modifications, and rather performant on both of them. The only part of the code specific to vendor is the platform selection, which is one line of code.

OpenCL falls down in terms of standard libraries such as cu{dnn,sparse,blas} but if you're writing everything from scratch it's fine.

[ Parent | on HN ]

Re:Intel Announces Knights Mill: A Xeon Phi for Deep Learning by joe_the_user on Thursday August 18, 2016 @12:10PM (#12313685)

I'm an independent developer in the process of choosing a GPGPU library.

I can see simple, comprehensible 20-50 sample code for cuda that does most simple tasks. With OpenCL, I get references to version, boiler-plate, mode with nothing that boils down to simple code.

If you have a simple sample, you should post it here or blog about it.

[ Parent | on HN ]

Re:Intel Announces Knights Mill: A Xeon Phi for Deep Learning by marmaduke on Friday August 19, 2016 @08:05AM (#12319448)

Look at Pyopencl; it makes it quite easy to prototype OpenCL apps in tens of lines of code.

Later, it's straightforward to port to c or c++ if that's your thing, though I find having numpy et al handy even in production code

[ Parent | on HN ]

Re:Intel Announces Knights Mill: A Xeon Phi for Deep Learning by duaneb on Thursday August 18, 2016 @01:00PM (#12314233)

Boilerplate is one thing—this is what transpilers are for, among other solutions—but can OpenCL match CUDA on a performance level?

[ Parent | on HN ]

Re:Intel Announces Knights Mill: A Xeon Phi for Deep Learning by marmaduke on Friday August 19, 2016 @08:06AM (#12319456)

Cl is comparable to Cuda where the cuda code doesn't employ nvidia specific primitives.

[ Parent | on HN ]

Re:Intel Announces Knights Mill: A Xeon Phi for Deep Learning by dagss on Thursday August 18, 2016 @08:15AM (#12311866)

Interesting...so do you use the vector subset designed for CPUs or the wavefront subset designed for GPUs?

[ Parent | on HN ]

Re:Intel Announces Knights Mill: A Xeon Phi for Deep Learning by marmaduke on Friday August 19, 2016 @08:02AM (#12319442)

I write code which maximizes coalesced memory access and plain old 32 bit floats.. nothing special. The respective drivers do a good job of mapping that onto the hardware given sufficient work group sizes.

[ Parent | on HN ]