Date: Sun, 5 Apr 2015 13:14:24 +0300 From: Solar Designer <solar@...nwall.com> To: john-dev@...ts.openwall.com Subject: Re: [GSoC] John the Ripper support for PHC finalists Agnieszka, On Fri, Apr 03, 2015 at 04:31:28AM +0200, Agnieszka Bielec wrote: > [a@...er run]$ ./john --format=pomelo-opencl -dev=4 --test > Device 4: Intel(R) Many Integrated Core Acceleration Card > Build log: Compilation started > Compilation done > Linking started > Linking done > Build started > Kernel <pomelo_crypt_kernel> was successfully vectorized > Done. > Local worksize (LWS) 64, global worksize (GWS) 64 > Benchmarking: pomelo-opencl, POMELO [POMELO OpenCL (inefficient, > development use only)]... DONE > Raw: 34.2 c/s real, 6400 c/s virtual BTW, I noticed that you're running a pomelo-opencl cracking job on -dev=4 now. (As well as another one on -dev=5.) This is fine. However, I used this opportunity to SSH in to the MIC card and see whether your use of it causes much load there. Not surprisingly, it does not. The load is almost non-existent: [solar@...er ~]$ ssh user@...0 [user@...er-mic0 user]$ uptime 14:09:22 up 1 day, 20:11, load average: 0.16, 0.34, 0.38 At full load, this would report load averages of 240 here. So it's like 1/1000 of full load now. I also tried this: [solar@...er ~]$ micsmc -a | less which shows the load on each core. There's just ~10% of system load on core 60, with everything else idle. You can/should be making sanity checks like these, too. Any idea why your OpenCL kernel isn't making much use of the MIC card? Sure, OpenCL for MIC is in a very poor state now, but we're seeing speeds (and usage) that are not exactly that poor for other kernels. Thanks, 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.