LinkBack URL
About LinkBacks399 now
399 now
396 as of this morning.
Annoyingly, one of my work laptops seems to have died :s
Currently got this lot running:
Home
-------
E6600 @ 2.7Ghz - SMP Client
IMac Intel Core - SMP Client
P4 2.4 GHz Laptop - Single Core
AMD XB2400 @ 1.8Ghz - Single Core
Work
------
C2D @ 2Ghz Latop - SMP Client
P4 2.8Ghz Test Server - Single Core
Hopefully this lot will help when a get the first round of units completed
Well the SMP units will be worth 1760 points each and the single core depends on what type of unit it is running but a good estimate is ~300 points each (some can be worth 180 some 350, who can say!). So your looking at ~6200 points for that little lot when they come in, I'm happy for the team all folding is good folding, but sad for me as you will be another person I can't match in folding performance due to complete lack of additional computers...
Team Ranking (incl. aggregate) 399 of 88112
and my smp units all landed
another note: intel > amd.
i started all 16 machines within about 20 minutes of each other. here's how they look right now:
compa000 (opteron): [10:31:53] Completed 1600000 out of 10000000 steps (16 percent)
compa001 (opteron): [10:38:41] Completed 1800000 out of 10000000 steps (18 percent)
compa002 (opteron): [10:39:51] Completed 1500000 out of 10000000 steps (15 percent)
compa003 (opteron): [10:37:42] Completed 1500000 out of 10000000 steps (15 percent)
compa004 (opteron): [10:36:07] Completed 1400000 out of 10000000 steps (14 percent)
compa005 (opteron): [10:31:17] Completed 1300000 out of 10000000 steps (13 percent)
compa006 (opteron): [10:32:50] Completed 600000 out of 10000000 steps (6 percent)
compa007 (opteron): [10:31:27] Completed 1400000 out of 10000000 steps (14 percent)
compb000 (xeon): [10:07:18] Completed 190000 out of 500000 steps (38 percent)
compb001 (xeon): [10:03:13] Completed 180000 out of 500000 steps (36 percent)
compb002 (xeon): [10:05:05] Completed 190000 out of 500000 steps (38 percent)
compb003 (xeon): [10:08:29] Completed 190000 out of 500000 steps (38 percent)
compb004 (xeon): [10:11:17] Completed 190000 out of 500000 steps (38 percent)
compb005 (xeon): [10:10:06] Completed 190000 out of 500000 steps (38 percent)
compb006 (xeon): [10:04:48] Completed 185000 out of 500000 steps (37 percent)
compb007 (xeon): [10:04:09] Completed 185000 out of 500000 steps (37 percent)
all are 2.6ghz chips, so it's not a clock speed thing - xeon is simply faster for floating point code
I'd second the AMD vs Intel at floating point,
My AMD machine is hugely slower that any of the others, it's doing a tiny 1500 frame job at about 3 mins per frame :s
The P4 Home laptop (which i think is actually only 1.8Ghz, got it confused with work laptop (which then broke)) is doing a 5000 frame job at about 30 seconds per frame.
P.S. Sorry Webbywill take me a long time to reach the levels you guys have already achieved though
Doug, to quote tesco - Every little helps!Originally Posted by DougMcDonald
I'd second the AMD vs Intel at floating point,
My AMD machine is hugely slower that any of the others, it's doing a tiny 1500 frame job at about 3 mins per frame :s
The P4 Home laptop (which i think is actually only 1.8Ghz, got it confused with work laptop (which then broke)) is doing a 5000 frame job at about 30 seconds per frame.
P.S. Sorry Webby
True,
Frustrated at the length of time it takes to get some stats on the board though! Am I right in thinking I need to complete a WU, or is it just a periodic update that I missed (maybe 24 hours or something)
Generally you'll start to show once you've completed a WU.
when you complete a WU, it ought to appear on the stanford site. the extremeclocking site has a delay of lots of hours, as it's processed in batches (the time due for the next data dump is at the top left of the page)
That makes sense, my main home PC was up to about 30% of a large SMP WU when i left for work this morning (had been running for about 6 hours)
Hopefully will complete at some point tonight, tomorrow morning at the latest
If thats the E6600 then assuming your in windows (linux is apparently loads faster for SMP) you will be looking at somewhere around 27-28 hours for each SMP unit I would think.
My E6550 at 2.8GHz runs a SMP unit in about 26 hours when it was at stock 2.33GHz it ran at about 30 hours, at 2.5GHz was 28.5 hours and at 2.65GHz 27 hours (all times approx and to the nearest hour).
I may clock it up to 3GHz tonight just to see if it can
Oh and you will soon catch up as I go home (from Uni) tomorrow and will have to share a room with my PC and while its not loud the lights in the keybard and mouse are a bit bright for sleeping next to! So I will drop from 24 hours a day folding to probably 16-18 depending on how much sleep I need! On the plus side I have an old PC at home I can fold on as well, but its going to be slow (old 2GHz Celeron).
Would be interesting to know how the Mac does for SMP - I could get my good lady friend's Macbook Pro and get it folding without her knowledgeor maybe with her knowledge after all I don't want to die!
Yes, it's the 6600. I think it was a half million step unit, and i was doing about 1% per 16 mins.
So i guess, 1600 mins or 26.66 Hours, which matches up to what you're saying.
I could happily push the processor to 3.15Ghz without any issues. Only thing i was seeing was that pushing the FSB above 355-360 seemed to make the PCI-e bus width drop to 1x.
I was told it was a quirk with the P5B and lowering the multiplier and upping the fsb above 400 would sort it, but it doesn't seem to.
The mac is a funny one, i got it folding quite happily, but I'm unsure how to monitor it, the install suggests altering settings such as name and team in the system preferences, which i have done. But no idea how to see progress on it
eBaums World 6 -30,061 41,389 11.23.07, 12am / 17.4 Hours
Woah - we're getting some big names cropping up now!
Yeah but Fermilab are going to go past us in 0 time flat!. 800K points yesterday alone...., if its going at the same rate another 400K to hit in a few hours time. Someone has got a cluster like DirectHex's folding flat out!.
Makes our processing power look rather pitiful.... oh to have a cluster like that to play with. (and i'm sure before Directhex points out, I BET its running linux
TiG
-- Hexus Meets Rock! --
We're still making steady progress though, and that's what f@h is all about - lots of people pulling together rather than a monster cluster pulling in silly numbers. They could just be stress testing for all we know
our integrator hasn't bothered to turn up today, so there should be enough time for another two SMP WUs per node before we rip up the test cluster
There are currently 1 users browsing this thread. (0 members and 1 guests)
Posting Permissions
Reply With Quote
