lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  PHC 
Open Source and information security mailing list archives
 
Hash Suite for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 7 Mar 2014 07:26:17 +0400
From: Solar Designer <solar@...nwall.com>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] TigerKDF paper and code ready for review

Bill,

Another correction:

On Fri, Mar 07, 2014 at 06:45:52AM +0400, Solar Designer wrote:
> BTW, why do you mention 32 KiB there?  Your random lookups arena is
> only 16 KiB, and you can't make it larger than that on current CPUs with
> 32 KiB L1d cache and 2 hardware threads/core, without incurring much
> extra L1 cache thrashing.  TigerKDF's total memory usage is irrelevant.

Of course, my "TigerKDF's total memory usage is irrelevant" implies "as
long as the total memory usage is low enough that a sufficiently high
number of concurrent instances fit in global memory", which is the case
we should be considering in this context.  32 KiB is low enough (it's
twice lower than Litecoin's 128/2 = 64 KiB, considering that Litecoin is
commonly mined with 2x TMTO).  In that case, the small random lookups
arena's size is relevant, whereas the total memory usage is not (yet).

Alexander

Powered by blists - more mailing lists