Madshrimps Forum Madness

Madshrimps Forum Madness (https://www.madshrimps.be/vbulletin/)
-   Site & Forum Feedback - Folding@Home (https://www.madshrimps.be/vbulletin/f5/)
-   -   yatta! (https://www.madshrimps.be/vbulletin/f5/yatta-5818/)

BlackRabbit 18th June 2004 17:13

yatta!
 
No, not these funny men...

I'd like to make a F@H yatta :)^
I was thing of 5-6 mobos in 1 array, should be nice..

Only: I have no idea how to make one.. Well, I've got a few ideas, but those plans are pretty basic and I don't know they'll fulfill my needs.

I've seen a few pics featuring a harddisk for every mobo, but I'd like to use none/1 HD for all the systems. The link in this has been down for a long time bw.

greetos,
BlackRabbit

BlackRabbit 18th June 2004 17:19

Ow and btw: is it possible to run some kind of a folding-proxy?
Something like a local folding server, providing enough WUs for 5 systems to work 1 or 2 weeks?

TeuS 18th June 2004 17:22

can't you boot Linux over network?

exotic root (DHCP/TFTP) should do the trick

BlackRabbit 18th June 2004 17:27

Quote:

Originally posted by TeuS
can't you boot Linux over network?

exotic root (DHCP/TFTP) should do the trick

I was also thinking of linux :)

But I'll need to learn a bit about network booting, is kind of a dark area for me :)

BlackRabbit 19th June 2004 10:13

Ow yes: order these in order of importance for fast folding, please:

- CPU Mhz
- FSB Mhz
- RAM timings
- RAM Mhz
- number of lights in your case

jmke 19th June 2004 10:57

- CPU Mhz

is the only that counts

BlackRabbit 19th June 2004 12:19

And how about AMD Mhz vs Intel Mhz?

jmke 19th June 2004 12:35

clock for clock, P4 vs AXP = AXP wins by quite a lot

AMD @ 2.3ghz
P4 @ 3ghz

Intel TestRig: p251_gnra_dd_epsilon_erf_e0.5_c18 total time: 4:08:20
AMD TestRig: p251_gnra_dd_epsilon_erf_e0.5_c18 total time: 4:25:00


A64 vs AXP clock for clock, you will see the A64 pull ahead if I'm not mistaken, no hands on experience there (yet;))


Speedwise

P4 << AXP < A64


more comparisons:

protein name - time per frame - total time - name foldstation

p257_pfold_p254_r0 0:01:38 2:43:20 21/05/2004 AMD TestRig
p257_pfold_p254_r0 0:01:44 2:53:20 24/05/2004 Intel TestRig
p524_BBA5_pf 0:05:23 8:58:20 15/05/2004 Intel TestRig
p524_BBA5_pf 0:05:24 9:00:00 26/05/2004 AMD TestRig
p543_BBA5_ext 0:05:25 9:01:40 27/05/2004 AMD TestRig
p543_BBA5_ext 0:05:22 8:56:40 28/05/2004 Intel TestRig
p568_BBA5_N 0:02:51 4:45:00 14/05/2004 AMD TestRig
p568_BBA5_N 0:02:49 4:41:40 15/05/2004 Intel TestRig
p909_vill_str0 0:04:23 7:18:20 17/05/2004 AMD TestRig
p909_vill_str0 0:04:43 7:51:40 21/05/2004 Intel TestRig

this one is nice ->
p910_vill_str0_rf 0:04:12 7:00:00 23/05/2004 AMD TestRig
p910_vill_str0_rf 0:04:12 7:00:00 25/05/2004 Intel TestRig

BlackRabbit 20th June 2004 00:34

Quote:

Originally posted by jmke
clock for clock, P4 vs AXP = AXP wins by quite a lot

And how about HT?
A P4 can do 2 WU at the same time, right?

jmke 20th June 2004 00:43

let me cook up some more stats, BRB ;)


All times are GMT +1. The time now is 08:16.

Powered by vBulletin® - Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO