[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 20 Jun 2012 17:44:50 +0530
From: SAYANTAN DATTA <std2048@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: bf-opencl
On Wed, Jun 20, 2012 at 1:44 AM, Solar Designer <solar@...nwall.com> wrote:
> On Wed, Jun 20, 2012 at 12:04:13AM +0530, SAYANTAN DATTA wrote:
> > How do I compile program written in IL? Is there any documentation
> > regarding AMD IL?
>
> Yes:
>
>
> http://developer.amd.com/sdks/AMDAPPSDK/assets/AMD_Intermediate_Language_(IL)_Specification_v2.pdf
>
> (506 pages). I think there's a lack of documentation on GCN, though.
>
> I think we should start by reviewing code generated from OpenCL source,
> much like we review x86-64 assembly code generated by gcc.
>
But as far as I know IL codes are independent of ASICs. So it shouldn't
matter whether its GCN or not.However ISA is ASIC dependent. So, did you
mean ISA instead of IL?
> BTW the bf_kernel is stable upto 1225Mhz core and 1300MHz memory.
Does this mean it was no longer stable at 1250 MHz core? Are you
> referring to my test from yesterday, where I got ASIC hang at 1250 MHz
> after 3.5 minutes? If so, that does not tell me much because I also got
> a system reboot after 15 minutes running the same code at stock clocks.
>
> BTW, changing the memory clock back and forth did not affect bf-opencl
> performance at all, which confirms that it indeed does not use global
> memory much (in the current code revision). That's good.
>
> Alexander
>
I think two or more john builds collided on the same card resulting in
crash after 15 min. Otherwise the implementation is perfectly stable at
stocks(I tested using pw-fake-unix hashes for 28 minutes). And yes the
card is unstable after 1225Mhz core.
Content of type "text/html" skipped
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.
