<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Steve Johnson wrote:
<blockquote
cite="mid:c1b74de10901141517j4f1a783atf588e3918a6c41f0@mail.gmail.com"
type="cite">On Wed, Jan 14, 2009 at 2:22 PM, Aaron Grattafiori <span
dir="ltr"><<a moz-do-not-send="true"
href="mailto:aaron@digitalinfinity.net">aaron@digitalinfinity.net</a>></span>
wrote:<br>
<div class="gmail_quote">
<blockquote class="gmail_quote"
style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Yep.
Cracking Encryption and finding MD5 collisions :)<br>
<br>
<a moz-do-not-send="true"
href="http://arstechnica.com/news.ars/post/20081013-company-puts-nvida-gpus-to-work-cracking-wireless-security.html"
target="_blank">http://arstechnica.com/news.ars/post/20081013-company-puts-nvida-gpus-to-work-cracking-wireless-security.html</a><br>
<a moz-do-not-send="true"
href="http://www.darknet.org.uk/2007/11/graphics-cards-the-next-big-thing-for-password-cracking/"
target="_blank">http://www.darknet.org.uk/2007/11/graphics-cards-the-next-big-thing-for-password-cracking/</a><br>
<a moz-do-not-send="true"
href="http://bvernoux.free.fr/md5/index.php" target="_blank">http://bvernoux.free.fr/md5/index.php</a><br>
<a moz-do-not-send="true"
href="http://somken.com/images/md5_crack1billion.jpg" target="_blank">http://somken.com/images/md5_crack1billion.jpg</a>
<- GPU is _MUCH_ faster.<br>
<br>
I don't know if its "popular" yet but.. it sure is cool and I hope for<br>
more open source tools for utilizing your GPU's vector tweaked mhz.<br>
<font color="#888888"><br>
-Aaron<br>
</font><br>
ps. Awesome work Jake (and team) on the SSL certs w/ MD5</blockquote>
<div><br>
</div>
</div>
I wonder if a port of seti at home would be good for this.<br>
</blockquote>
I posed this question to an irc buddy up in Toronto who has done work
with cog. sci., and he<br>
pointed me to a Google tech talk given by one of his professors at the
U. of Toronto...<br>
<br>
<a class="moz-txt-link-freetext" href="http://tinyurl.com/hinton">http://tinyurl.com/hinton</a><br>
<br>
"The Next Generation of Neural Networks"<br>
<br>
The document classification experiments are pretty amazing...maybe
these types of neural nets would work well for spam classification?<br>
And could learning be speeded-up (and more RBM layers added) with the
use of CUDA?<br>
<br>
( I asked said irc buddy, who says "Bayes is good enough". But: I
wonder? )<br>
<br>
-Scott<br>
p.s. Distributed Password Recovery on the CUDA zone:
<a class="moz-txt-link-freetext" href="http://www.nvidia.com/object/cuda_home.html#state=detailsOpen;aid=450f1e45-cf76-43ee-8b6f-fa2d8af8202c">http://www.nvidia.com/object/cuda_home.html#state=detailsOpen;aid=450f1e45-cf76-43ee-8b6f-fa2d8af8202c</a><br>
<br>
</body>
</html>