|
Message-ID: <1b21a725e87dcf7fc6e230170b8549fb@smtp.hushmail.com> Date: Mon, 7 Sep 2015 21:36:37 +0200 From: magnum <john.magnum@...hmail.com> To: john-dev@...ts.openwall.com Subject: Re: SIMD_PARA_SHA256 On 2015-09-07 11:39, Solar Designer wrote: > On Mon, Sep 07, 2015 at 09:31:27AM +0200, magnum wrote: >> On 2015-09-06 21:59, Solar Designer wrote: >>> We should increase SIMD_PARA_SHA256 from 1 to 2 at least for XOP builds. >> Perhaps we should do so for SHA-1 too if built without OpenMP (unless >> that is already the case). > > This is not obvious. For fast hashes, we currently disable OpenMP and > suggest use of --fork, and I expect that --fork will show similar > behavior to OpenMP in terms of which interleaving factors are optimal. > Maybe we should test --fork=8 explicitly and decide based on that. You are right of course. It's a pity you can't "--test --fork" yet. I only committed the x2 for SHA256 with XOP. >> Do you have some gcc-4.9 for bull already built that we could try too? > > Not already built, but I can build it easily. Why 4.9 and not 5.x? 5.x is also interesting. Ubuntu LTE uses gcc-4.8 and latest Ubuntu uses gcc-4.9 so those two versions should ideally be tested too. Older versions than that are of little priority to me. >> I'll do some new AVX and AVX2 tests. FWIW I have yet to see anything calling for a change. > Do you have pre-AVX machines to tune SSE2 and SSE4.1 interleaving on? > I expect similar tradeoffs there, for Core 2 alikes (no HT) vs. early > Core i7 (HT). I could probably find some but again the test matrix quickly grows out of hands. Some day we should just build tests into ./configure (optional and along with an instruction to run on unloaded system and so on). Actually we could just write some instructions instead, for using the already provided means: Run testpara.pl and then supply good values to ./configure. It's not that inconvenient. Come to think of it, testpara.pl should end with printing a complete command line for ./configure, ready for copy+paste. That I'll do right 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.