1 / 2

Performance Comparison between OpenCL and OpenMP on Multi-core CPUs

Performance Comparison between OpenCL and OpenMP on Multi-core CPUs. Is there a performance difference?. Vectorize your code !. YES, it is. But be careful with data-dependent branches!. Figure out your memory access patterns !. Before. Use zero-copies !. After.

agatha
Download Presentation

Performance Comparison between OpenCL and OpenMP on Multi-core CPUs

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Performance Comparison between OpenCL and OpenMP on Multi-core CPUs Is there a performance difference?

  2. Vectorize your code ! YES, it is ... But be careful with data-dependent branches! Figure out your memory access patterns ! Before Use zero-copies ! After Check if fine(-grain) works fine ! Optimize the math ! Before After Before Row-major is CPU friendly, column-major is NOT. Transposing OR undoing the transposition works! After Using fast-math compiler option works! Make sure accuracy is still OK! Replacing D2H and H2D with zero-copy works! Before After Fine grain parallelism doesn't work for all apps ... ... but increasing workload granularity might fix it! Jie Shen j.shen@tudelft.nl ... but it DOESN'T have to be ...

More Related