10 Oct
2008
10 Oct
'08
2:58 a.m.
Dave,
I am sure you guys are really smart, but may have overlooked something or want to listen.
I will make a guess, that this is gene sequencing related.
Data small, search field large, and iterations many.
Have you considered a FPGA for this?
Tell me what you are doing, I might be able to help or present some hardware speedup
alternatives.
Randy
From: mcguire at neurotica.com
Date: Fri, 10 Oct 2008 01:52:13 -0400
To: cctech at classiccmp.org
CC: cctalk at classiccmp.org
Subject: Re: NEC SX-4B on Ebay
On Oct 9, 2008, at 11:26 PM, William Donzelli wrote:
Eight, actually. My employer is about to be, and there are quite
a few others. With design cycles measured in years rather than
weeks, the supercomputer world doesn't consider things "obsolete" as
quickly as the PC world. Indeed, the code is being developed on a 25-
year-old super, but admittedly only because it's available and
compatible.
Besides, SV1ex machines are still pretty pricey, and T90s are
pretty difficult to find, and WAY out of any reasonable price range.
-Dave
--
Dave McGuire
Port Charlotte, FL
_________________________________________________________________
Get more out of the Web. Learn 10 hidden secrets of Windows Live.
http://windowslive.com/connect/post/jamiethomson.spaces.live.com-Blog-cns!5…
The code
is being developed on a (slow but cheap to run) YMP-EL,
and it
will eventually run on a (fast but expensive to run) SV-1.
Who is actually still using an SV1 in production? It is a ten year
old super.