I think I found a way to get TA back on #1 for the dailies

[mindless]

http://cgi.ebay.com/aw-cgi/eBayISAPI.dll?ViewItem&item=426080142

Anybody got $35,000 I can borrow forever?

 

[dvch]

Let's see 600+ active members in TA rc5 would be about $60 each.

If someone wants to start a funddrive for it the rc5crackheads pledge $120 towards it's purchase.

 

[mindless]

Id pay $60 assuming I didnt have to supply its power, couldnt afford much over that though

 

[br0wn]

it is cheaper if you build a BEOWULF cluster.
Cheaper and more powerful.

The Cray is capable of 16 GigaFlops, and costs
$35,000.

Building BEOWULF cluster is around $650 per GigaFlop,
so for only $10,400 you will get 16 GigaFlops.

 

[mindless]

/me thinks TA needs a nice large Beowulf, especially considering the cost per Gflop.

If I could just aquire one that did say 64 Gflops and costed me around $41,000 would anyone like to foot the bill?

 

[Joker81]

I don't know but one of the current buyers is new and the other only has 4 remarks. I would be suspicious of selling it. And its only day one of the sale. This is an excellent deal though 4GB of Solid State memory back in 1993 would have probably cost 500,000+ (just a guess no flames please). How would you organize 16 seti clients BTW your screen would be full of the command lines even minimized but still on the desktop to get pririty. Since each processor can do 1 Gigaflop/s it would take 3000(50min)(running beta 2.7) sec Max to complete each Seti Work unit 24hours /50min * 16 = 460.8 units a day for one computer ahhhhhh!!!! I would need about 120 p3 667's although thats still cheaper but hey how many people on your block can say Yep i have one computer that can do 460 units a day(slyly as you take a sip from your beer). They would say you mean in RC5 right. "nope" in seti. in one year you could have 166510.08 units under your belt(taking 99% uptime). Not bad at all. maybe they would do a trade for my old dual celeron board and 366's.

Joker

 

[Marnix]

Hehe it won't help you getting 3,8 million blocks a day
Kidding... Peace to TA, honour to DPC

 

[LANMAN]

LMAO Paulson!!

Personally I just want to get RC5 wrapped up! Over a thousand days and only 30% completed... whew...

--LANMAN

 

[Dutchman2000]

Paulson: If dnet would have allowed it we would have flushed everything in one or two days

 

[RaySun2Be]

What I find interesting is that the place that is selling the Cray is in essence replacing it with a huge Beowulf type cluster.

http://www.psc.edu/publicinfo/news/2000/terascale-08-03-00.html

Now instead of buying the Cray, how about we recruit whoever is/will be in charge of that "crack Rack"?

Hmmmm, let's see, do we have anyone in the Pittsburg area that could talk to them?

Anyone involved with the project that could talk about using their idle CPU time?

Hmmmmmm, who do we know at COMPAQ that might be involved with this project.

Dan, good buddy, ol' chum. How's it going. You wouldn't just happen to know anything (or who) about this project would you?

Why spend money on 16 Gigaflops when we could be going after Teraflops!

Or do both! hehe

(Although I wouldn't sneeze at 16Gflops added to the team! )

 

[MWalkden]

Now the price has dropped for 16 GigaFlops! The crackheads still pledge the first $120.00 towards such a feat! Anybody else????

 

[RaySun2Be]

Another interesting thing about the new cluster are the projects they are going to us the cluster on:

http://www.psc.edu/publicinfo/tcs/

Thunderstorm Simulation and Protein Folding sound familiar, wasn't there some topics on those as being distributed proejcts?

Hmmmmmm

 

[amok]

I don't see the point in anyone doing protein folding sims right now. IBM's Blue Gene will be able to do in about a week the same amount of work that that cluster would take 2 years to complete (I believe Blue Gene is supposed to be completed sometime in the next two years)...

Even with Blue Gene's unearthly amount of processing power, it will still take them 8 months to a year for each protein. Just how much do these people hope to accomplish?