Hello. I'm trying to get pocl to work on a custom multicore processor.
As far as i understand pocl expands all work items of a work group into a single binary (either by expanding the code or with loops) in the call_pocl_workgroup (or in pocl-workgroup script alternately) and then it is executed serially. This is done by setting pocl::LocalSize.addValue(local_x) and respectively for y and z (or with the opt's -local-size parameter in the script). My intention is to be able to override this functionality and avoid expanding the code and run a unique thread for each work item. My question is which is the cleanest way to do this, with minimal modification in the code. Any guidelines or hints would be appreciated. Thanks for your time, Panayiotis Apostolou. ------------------------------------------------------------------------------ Managing the Performance of Cloud-Based Applications Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. Read the Whitepaper. http://pubads.g.doubleclick.net/gampad/clk?id=121051231&iu=/4140/ostg.clktrk _______________________________________________ pocl-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/pocl-devel
