- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello,
It seems CILK is much easier to use than Open CL.
Do you have any information of the performance of CILK vs. Open CL on the same machine (e.g Intel's Ivy-Bridge) ?
Does it depend on the what I'm trying to do ?
My interest is Signal Processing (FFT, matrix manipulation etc.)
Thanks,
Zvika
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I don't think Clik(tm) Plus (if you mean Intel's implementation) and OpenCL are suited to the same projects.
OpenCL was put forward as a way to support compatibility with applicatations architectures where C++ isn't suitable, not so much as a performance competitive alternative to Intel compilers.
It's in a relative early stage of development. It seems LLVM has only recently been attempting to catch up to gcc in performance. In part, Intel's claim to higher performance than gcc has been based on more aggressive default options, the use of #pragma optimization directives, and the performance libraries provided with Intel C++. The latter would appear to be important for your area of interest.
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page