Folding @ Home

[Cipher7071]

I may be way out of my league, but here goes anyway. I remember reading not so long ago that research is being done on what I believe was referred to as "stackable" chip technology. While it was acknowledged that there is presently a thermal problem, it does appear to have potential benefits in use of available space, length of wires, etc.

Besides, even if the thermal problem isn't solved, laptops with the new chips might find an alternative market as hot-seats for northern deer hunters.


P.S. You guys are making me want to run out and upgrade my computer, never mind that don't I actually have use for that kind of processing power.

[ December 16, 2003, 17:48: Message edited by: Cipher7071 ]

[Krsqk]

Originally posted by Cipher7071:
P.S. You guys are making me want to run out and upgrade my computer, never mind that don't I actually have use for that kind of processing power.

It's got me--the call of the Never-Ending Upgrade! 3.0 gigahertz! Wait, no, 5.0 gigahertz! Arrgh, 8.8 terahertz! Oh, look, a quantum computer! Gotta have it! AAAAAAAAAAAAAAAAAAAAAAAAAAH!

[Thermodyne]

Yes, there is some research into stacking chips on die. But with the heat issue continuing to plague the 90 nanometer chips, stacking is a ways off yet. Early 90nano chips are reported to leak more than 50% of the power that they consume. Tandem chips on die is closer, but that is a side by side arrangement. With x86, quad chips on board is about the performance/price limit. But for some applications, 8 chip systems serve well. Another arrangement is blades, where “systems on a card” are stacked in a high speed buss arrangement. The larges one of these that I have personally seen had 24 blades/CPU’s.

For folding or crunching seti, stacks are some times used. They are striped down systems that are served by a complete system w/hard drive. The rest of the stack is just main Boards, CPU and ram with a NIC. By eliminating the additional hard drives and cases, a low cost high performance system can be built. If all goes well and I can get a grip on Linux, I’ll have a 17.4GHz stack running by new years. The cost will be about $83 a GHz. Output should be 8 times what I get from my FX51 which cost way more than the whole stack will.

[Thermodyne]

Congrats to Rdouglass on breaking 2500. It’s a little harder than SETI


IF any of you have some spare CPU time, we could sure use your help. Check us out http://folding.stanford.edu/cgi-bin/teampage?q=34149
We will break into the top 100 teams some time tomorrow morning, and we are in the top 25 teams based on production. Also, we are two months old today! No team has climbed as fan and as fast as we have. But we are way down the list as far as membership goes. We could sure use any help that you guys can offer. And remember, it is for a good cause.

[ December 17, 2003, 00:02: Message edited by: Thermodyne ]

[Thermodyne]

Well, the team made the top 100 today, out of 34000+ teams. And nine weeks ago it was only an idea some guys at Sharky's had. Sure wish some more of you would come over and help out.

http://folding.extremeoverclocking.c...?&pagenumber=1

[ December 18, 2003, 21:04: Message edited by: Thermodyne ]

[Bill Door]

I suppose I should have posted sooner, but I'm on as Ste.
As of 22:52 GMT, 18-12-2003 I have done 12 WU and am 91% through my latest work unit (2275/2500)
I recieved the WU on 15:08:45 PST, 15-12-2003

I've got an AMD Athlon XP 2000+ which Windows XP clocks at 1.66Ghz. and 512MB of RAM.

It isnt fast but it's stays switched on all day. The real question is, dare I leave it switched on over the Christmass break...

[narf poit chez BOOM]

shouldn't we be the SPACE EMPIRES forum?

i mean, what's this sharky forum anyway?