Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAD=7XjMuY9Hk7b3S62s+HNL2FVtnr8fxirFjFi_wuN8Ah7UfmA@mail.gmail.com>
Date: Wed, 3 Feb 2016 22:20:24 -0500
From: Frank Wang <frankw@....edu>
To: john-users@...ts.openwall.com
Subject: Re: Doing raw SHA512 hashes

Thanks for the information! I actually need to hash about 64 bytes at most.
I'll try to separate out the code and see how far I get.

Frank

On Wed, Feb 3, 2016 at 3:39 PM, magnum <john.magnum@...hmail.com> wrote:

> On 2016-02-02 23:49, Frank Wang wrote:
>
>> I am trying to run about 10 million raw SHA 512 hashes in my program and
>> offloading it to a GPU will make it substantially faster. It seems like
>> John the Ripper has this functionality, but it seems to be intertwined
>> with
>> the password cracking capabilities. Are there any SHA 512 OpenCL
>> implementations that are easy to separate and integrate into my program?
>> Possibly that already exist in John the Ripper?
>>
>
> I would think so. Our current GitHub code contains (among other
> implementations) generic SHA512 macros in src/opencl_sha512.h which is then
> used in src/opencl/office2013_kernel.cl and src/opencl/
> pbkdf2_hmac_sha512_kernel.cl.
>
> These are raw single-block functions though, for performance reasons.
> Perhaps you need to hash data longer than 112 bytes, eg. with generic
> "init-update-finish" style functions? I don't think we have any such thing
> for OpenCL SHA512 right now but it's easily accomplished with generic
> Merkel-Damgard wrappers. If you know that stuff it's easy.
>
> 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.