Date: Sun, 22 Apr 2012 13:04:26 +0200 From: magnum <john.magnum@...hmail.com> To: john-dev@...ts.openwall.com Subject: Re: New RAR OpenCL kernel On 04/22/2012 04:17 AM, Claudio André wrote: > >> This is similar to what I get on a mobile E2 "Loveland" GPU. Are the >> GPU's this slow, or do I have a major problem? This is about half the >> speed of one CPU core. I do get warnings about register spill on that >> one though. >> > > I suppose you don't have a profiler for AMD. So attached an output. I have this sprofile thingy. I just did not consider it useful for a toy GPU like mine (I may be wrong of course). Maybe it's time to read some docs again. > It is not science, and i'm a confused user myself. But. > - ALUBusy (very low). > - ALUPacking (low). Would both these figures by closer to 100 in a dream scenario, or what? By the way my previous version of rar got an "occupancy" of 0.01 or so (lol) in nvidia profiler. We'll see if there is any change now. magnum
Powered by blists - more mailing lists
Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.