Date: Sun, 15 Mar 2015 08:55:12 +0800 From: Lei Zhang <zhanglei.april@...il.com> To: john-dev@...ts.openwall.com Subject: Re: [GSoC] building JtR for MIC I guess some other users of the MIC card are using a lot of space in their home directory. I'll tell my lab colleagues to delete unneeded files, and see if that helps. As for MIC intrinsics, I do have them in DES_bs_b.c. It's very likely there're still things to fix in the autoconf system. And I just found another problem with it: when CFLAGS is set manually, the generated Makefile won't include '-O2' (or '-O3') option in it and results in unoptimized code. That's why the benchmark ran even slower last time. I'll try to get scrypt to work first and see how the rest of the benchmark run. Lei > On Mar 14, 2015, at 11:21 PM, Solar Designer <solar@...nwall.com> wrote: > > On Sat, Mar 14, 2015 at 10:59:46PM +0800, Lei Zhang wrote: >> This time I got exclusive use of our MIC card and ran john-jumbo's benchmark again. >> Here's the result: >> ---------------------------------------------------------- >> Will run 240 OpenMP threads >> Benchmarking: descrypt, traditional crypt(3) [DES 64/64]... (240xOMP) DONE >> Many salts: 11702K c/s real, 48709 c/s virtual >> Only one salt: 4253K c/s real, 17728 c/s virtual > > The "64/64" suggests you're running scalar 64-bit code, rather than the > 512-bit vector code. > >> It's better than last time. But descrypt is still too slow and scrypt is still crashing. >> >> I checked the status of the MIC card, and found the free memory space left is no more than 3GB. This is reasonable since MIC has no external disk and the filesystem resides entirely on it's RAM. > > You need to free up more space there. A freshly booted MIC card > typically has over 7 GB free. Clearly, you can fit your JtR jumbo > binary and libraries in there and still have over 7 GB free. > > If there are other people's files in there that you can't delete, then > maybe you should move to using our MIC card - we can setup you access. > >> As for descrypt, I don't know why it's so slow. It's supposed to share the same code as john-core, right? Or maybe there's something wrong with configuration. I'm still checking. > > Yes, something is wrong with the configuration, preventing the MIC > intrinsics from being enabled. > > Or maybe your tree doesn't even have the MIC intrinsics in DES_bs_b.c > yet? magnum merged my changes from core into jumbo just a few days ago. > Have you updated to the latest jumbo tree? > > Alexander
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.