[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20140112101337.GA9134@openwall.com>
Date: Sun, 12 Jan 2014 14:13:37 +0400
From: Solar Designer <solar@...nwall.com>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] escrypt memory access speed (Re: [PHC] Reworked KDF available on github for feedback: NOELKDF)
On Sun, Jan 12, 2014 at 01:14:53PM +0400, Solar Designer wrote:
> Xeon Phi 5110P, using 2 GB, r=8 p=240, 240 threads, Salsa20/2, abusing
> scalar units for computation (no SIMD code implemented yet), no
> prefetches, 10 hash computations:
>
> real 0m 5.00s
> user 17m 54.14s
> sys 0m 18.27s
>
> 2*3*10*2^30/10^9/5.00 = 12.88 GB/s
Tried running 3 instances in parallel, 80 threads for each, 6 GB total
memory usage:
[user@...er-mic0 user]$ cat>r
OMP_NUM_THREADS=80 LD_LIBRARY_PATH=. time ./tests &
OMP_NUM_THREADS=80 LD_LIBRARY_PATH=. time ./tests &
OMP_NUM_THREADS=80 LD_LIBRARY_PATH=. time ./tests &
[user@...er-mic0 user]$ . r
[...]
real 0m 20.85s
user 8m 39.12s
sys 0m 12.79s
real 0m 20.85s
user 8m 44.29s
sys 0m 11.24s
real 0m 20.85s
user 8m 40.44s
sys 0m 13.44s
No luck - this is even slower:
2*3*10*3*2^30/10^9/20.85 = 9.27 GB/s
Alexander
Powered by blists - more mailing lists