e Bay... well maybe.
Searching ebay for 3ware escalade cards, you may find some 8 ports or 12 ports IDE controllers. With this, you should be able to connnect 20-40 drives per PC.
Hard to find however, but often not so expensive (IDE cards are no more wanted).
After that, the last issue would be the power supply, but I guess this is just simple DIY.
Reading the article, the goal is nowhere near building a real exaflop computer, but more about thinking about issues (like processor data feeding).
In a year and a half, we shouln't have more than 100 GFlops per socket, which means that you will still need 10 millions of processors (not cores!) to achieve the exaflop computer. No chance to build a cluster that big (at least these years).
The all-times progression of the top500 shows that exaflop computers should arrive around year 2020, definetly not tomorrow. (x10 every ~4 years, 2008:1 PF, 2012:10 PF, 2016:100 PF, 2020:1 EF)
Actually, they combine trash with ethanol to run a diesel engine... The electricity produced is 90% higher than it would be with only ethanol.... is what I understood:)
I used to love NetBSD. The package management is indeed great. BUT :
About stability, it worked well on a 486, but I never managed to run it stable on a brand-new machine with an Athlon (the kernel was always falling in vm_page_fault traps) whereas linux worked with no issue (hence not a hardware issue, more something like an AMD issue?).
About speed, again, unfortunately, linux performed better.
I really wanted to stick to NetBSD, but after 1 year trying to have it functionnal, I installed linux, and this day, all my problems were gone.
I'm afraid that the great thing in NetBSD - which is multiple platforms support - will soon be irrelevant, since linux already supports all the currently-used architectures.
In all case, I hope NetBSD will survive and become more usable. But as said, it needs a lot of work.
Slashdot Mirror
e Bay ... well maybe.
Searching ebay for 3ware escalade cards, you may find some 8 ports or 12 ports IDE controllers. With this, you should be able to connnect 20-40 drives per PC.
Hard to find however, but often not so expensive (IDE cards are no more wanted).
After that, the last issue would be the power supply, but I guess this is just simple DIY.
"an eight-PC Debian cluster"
"[we] wanted to implement a system which is very scalable"
8 PCS ? 64 cores at most. And they call that scalable ? Come on, today's top500 top machines scale on 10'000 cores. They're 15 years late.
Reading the article, the goal is nowhere near building a real exaflop computer, but more about thinking about issues (like processor data feeding).
In a year and a half, we shouln't have more than 100 GFlops per socket, which means that you will still need 10 millions of processors (not cores!) to achieve the exaflop computer. No chance to build a cluster that big (at least these years).
The all-times progression of the top500 shows that exaflop computers should arrive around year 2020, definetly not tomorrow. (x10 every ~4 years, 2008:1 PF, 2012:10 PF, 2016:100 PF, 2020:1 EF)
Actually, they combine trash with ethanol to run a diesel engine ... The electricity produced is 90% higher than it would be with only ethanol. ... is what I understood :)
I used to love NetBSD. The package management is indeed great. BUT :
I really wanted to stick to NetBSD, but after 1 year trying to have it functionnal, I installed linux, and this day, all my problems were gone.
I'm afraid that the great thing in NetBSD - which is multiple platforms support - will soon be irrelevant, since linux already supports all the currently-used architectures.
In all case, I hope NetBSD will survive and become more usable. But as said, it needs a lot of work.
DCC is indeed really similar to what's patented here. IRC is still far from it, since it's centralized (all communication goes through the server)
DCC (in fact CTCP) specification was published in '94. Net2Phone patent was done in 2000.
Skype will have no problem proving Net2Phone's patent is irrelevent.