[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 28 Jun 2013 10:56:20 +0200
From: Katja Malvoni <kmalvoni@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: Parallella: bcrypt
Hi Alexander,
On Thu, Jun 27, 2013 at 5:15 PM, Solar Designer <solar@...nwall.com> wrote:
> Katja,
>
> On Thu, Jun 27, 2013 at 04:54:31PM +0200, Katja Malvoni wrote:
> > 26c: 905f 4806 lsl r20,r20,0x2
> > 270: 2456 lsl r1,r1,0x2
>
> We should be able to avoid needing these instructions if you pick the
> version of the BF_ROUND macro that's intended for archs without scaled
> index on loads. The crypt_blowfish.c file in musl doesn't include it,
> so you'll need to take it from our separate crypt_blowfish distribution.
>
> In fact, here it is:
> [...]
> Another optimization to try is unrolling more rounds. The loop in
> musl's BF_encrypt() unrolls only two rounds, but it has that "#if 0"
> block with all 16 unrolled - would the code still fit if you change it
> to "#if 1"? Perhaps it would.
>
With other version of BF_ROUND macro and with "#if 1" execution speed on
one core is 39.131000 ms and 288.555000 ms on all cores.
Yaniv, what is the most efficient way to measure execution time of BF_crypt
call on Epiphany core? Using timer in clk mode? CTIMER0 is 32-bit register?
Katja
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.
