![]() |
|
|
#12 | |
|
Aug 2002
Quebec, Canada
29 Posts |
Quote:
Anyway, 30 was just a number thrown out there. I'd have to sit down and do real calculations, but they would be conditional on knowing how much the electrical/AC work would cost. My feeling would be it would be more around 40 computers, assuming I'd be holding back some of the cost to pay the electricity. 112G of P4 goodness. |
|
|
|
|
|
|
#13 |
|
Mar 2003
Melbourne
5·103 Posts |
You guys are too frugal. Use a bit of imagination... :)
http://www.vapochill.com/ Buy a couple of these cases and push a couple of cpus over 4GHz. :) Then have about 20 of the usual suspects. (P4 2.4s etc..) Personally with $25k US would go a long way here is Oz hehe. I'd have to go with a diskless farm. Not buying HDDs saves a lot of money and power, not to mention avoiding keeping the neighbours up with the noise :) As for a server, I'd have to go with dual Xeon board...why? Well it has PCI-X 64bit 66/100 MHz slots, grab a IDE RAID controller that can run at those speeds and you have awesome HDD throughput. Motherboard: http://www.tyan.com/products/html/thunderi7505.html http://www.supermicro.com/PRODUCT/Mo...7505/X5DAE.htm http://www.abit.com.tw/page/en/serve...station+Boards http://www.intel.com/design/servers/...e7505vb2_srvr& Controller (SATA): http://www.adaptec.com/pdfs/2410SA_datasheet.pdf http://www.promise.com/product/produ...112&familyId=2 Of course you'd have to grab a decent network switch with GigEthernet ports. -- Craig |
|
|
|
|
|
#14 |
|
Aug 2002
4768 Posts |
Asus P4P800 with P4 2.8C with HSF.....$338
Samsung PC3500 DDR 256MB $45ea x2.....................................$90 Western Digital 30GB HDD......................$39 Generic AGP video card...........................$11 Sparkle 300W PSU..................................$21 Total......................................................$499 Shipping and taxes vary, approx.............$50 Say $550 for each unit. Possible 45 machines not including electricity. One 20M LL test per machine= over 112.5 p90 CPU years per 10 days. |
|
|
|
|
|
#15 |
|
Oct 2002
Lost in the hills of Iowa
7008 Posts |
It's too bad the P4-based Celerons are so SLOW at Prime - I seem to recall them being SLOWER than Athlons per clock, as opposed to the actual P4s that are significantly faster per clock.
Prime of the few cases where the real P4 is significantly MORE cost-effective than Celerons.... |
|
|
|
|
|
#16 |
|
"Patrik Johansson"
Aug 2002
Uppsala, Sweden
52×17 Posts |
First USD 25000 = SEK 193540. All prices in SEK, including tax.
Code:
ASUS P4S800 837.50 overclocked to 3.14 GHz (got that) or ASUS P4P800 1118.60 overclocked to 3.22 GHz (guessing) P4 2.40C 800 MHz 1847.26 Kingston 256 MB PC3200 CAS 2.5 597.11 Maxtor, DiamondMax Plus 8 30GB 511.98 Compucase Miditower CI-6A21 300W 800.00 Papst case fan 120 mm 169.00 Total SEK 4762.85 or 5043.95 Shipping 200 120 W power for three years (3150 kWh) 2800 Code:
D-LINK DES-1008D 10/100 DUAL SPEED 8-PORT SWITCH 3 x 406.25 Network cables, 1 m 21 x 56.25 Network cables, 2 m 5 x 68.75 Network cables, 5 m 1 x 93.75 Network total SEK 2837.50 I just have to make sure they are split up on different (10 A) fuses. |
|
|
|
|
|
#17 |
|
Aug 2002
Quebec, Canada
29 Posts |
Do the USB ports on the ASUS P4S800 work under WIN XP? The P4S333 has problems is why I'm asking.
|
|
|
|
|
|
#18 | |
|
Aug 2002
Dawn of the Dead
5·47 Posts |
And guess what boart he chose ... asus p4s333 ... I suspect this is the reason, as it is impossible to make the machine fail in torture or any other load intensive process. I suspect the boart has far more to do with this ... look at his results, certain machines almost always fail, looks like a QC issue.
Best is to perform thorough doublecheck testing before deployment and at periodic intervals afterwards. I'm at 3000, stock vcore, only because I haven't bothered to push for more, all this monitor unplugging sucks. So far, 9 of 9 doublechecks have matched ... Quote:
|
|
|
|
|
|
|
#19 | |
|
Aug 2002
Dawn of the Dead
5·47 Posts |
Dont know about xp, I used a usb louse while installing w2k pro, usb is usb isn't it ... a veritable disease. If I could disable it ...
Quote:
|
|
|
|
|
|
|
#20 |
|
Sep 2002
Austin, TX
10001100012 Posts |
to save on cooling... the machine can be biult in Alaska. During the winter months maybe the cpus could get about 4ghz...
|
|
|
|
|
|
#21 |
|
Nov 2002
4310 Posts |
My slightly older comment wasn't about different systems, it was about having the same board and cpu but going for something that has been out slightly longer, and thus possibly less expensive than 'state of the art'.
According to the benchmarks, a P IV 3.2 GHz could pull a DC in 'bout 41 hours. A 2 GHz model would take 'bout 65 hours, or 'bout 50% longer (close enough for government work). 3.2 is $387 according to pricewatch, and the 2.0 is $117. So the faster chip is over three times as expensive. If you look at processors alone you could average a DC every 22 hours or so on three slower machines. |
|
|
|
|
|
#22 | |
|
"Patrik Johansson"
Aug 2002
Uppsala, Sweden
52×17 Posts |
Quote:
|
|
|
|
|
![]() |
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| A dream, will stay a dream ( new Nvidia Quadro) | firejuggler | GPU Computing | 0 | 2018-03-28 16:02 |
| Theoretical Experiment Design | c10ck3r | Homework Help | 7 | 2015-02-03 08:54 |
| Theoretical Evaluation of a GPU | tapion64 | PrimeNet | 10 | 2014-04-09 22:21 |
| Theoretical Model | aymanmikhail | Math | 16 | 2011-09-06 01:50 |
| Maximum theoretical MPG | TimSorbet | Lounge | 9 | 2008-07-14 22:45 |