Page 1 of 1
Post your PPD here! (looksee, you may have been flagged)
Posted: Sun Jul 27, 2003 9:48 pm
by riffst3r
hullo fellow folders!
i'm trying to figure out points per day based on type of processor... i know i can get this with stats, but i can't get the specs of your folding boxen there... so, post your (average of seven days) PPD as of now, along with the boxen producing 'em.
here's mine:
Code: Select all
Pentium 4 2Ghz Northwood
256MB PC2100
i845 Chipset
DAY1 = 29.40
DAY2 = 39.00
DAY3 = 0.00
DAY4 = 53.29
DAY5 = 53.29
DAY6 = 54.59
DAY7 = 29.39
Average PPD = 36.99
Posted: Mon Jul 28, 2003 12:42 am
by dukla2000
Not sure you can get easily comparible numbers because (at least in my case) depends on what other load is on the system. (e.g. The Sims effectively halts any folding progress and my digital TV card also halves folding progress. Standard 'office' applications also affect things but not nearly so badly.) Also my XP numbers assume Gromacs - the occasional Tinker slugs things as well! Last, the occasional Stanford server glitch can throw you: I have seen a couple of times 1 of my boxen waiting 12 hours for new work!
But my peak numbers are (about):
Athlon XP2700 72 ppd
Athlon XP2100 65 ppd
Athlon XP2000 60 ppd
Athlon 1.4 40 ppd
Pentium 3 1.2 50 ppd
Pentium 2 233 2 ppd
(not all of these are live in boxen today!
)
Where do we *get* thes numbers from?
Posted: Mon Jul 28, 2003 8:12 am
by NeilBlanchard
Hello Riff:
Where do we get these numbers?
My SOTP numbers:
Athlon 2100+ (T'Bred B) @ 1.986mHz = ~50-100 P/D
Athlon 700 "Classic" = ~20 P/D
[Edit] Here are my last seven days -- but I don't know which machines produced them! Going backwards from today;
1 WU = 29.40 (but I know that I'm "sitting" on a 54 pointer because the server is busy...)
2 WU = 107.9
2 = 86.30
1 = 53.3
4 = 137.3
1 =54.6
2 = 109.2
I'm using just two machines for Folding -- the third one is back on my longtime SETI effort, cranking out 8 WU a day!
http://setiathome.ssl.berkeley.edu/stat ... _1107.html
Posted: Mon Jul 28, 2003 9:38 am
by riffst3r
at folding.extremeoverclocking.com user stats, of course
you can just list all the boxen you have folding and the amount of points you make per day for all of them.
for that, find the average of the "last seven days production". or if you know for sure, or even a slight approximate, you could do what dukla2000 did.
Posted: Mon Jul 28, 2003 10:28 am
by Zyzzyx
I'm an SPCR reader, but I've got my own little Folding team going (Legends of Westwood, #13720). I visited folding.extremeoverclocking.com and was able to manually enter my team number. Very cool. But I didn't have access to the user stats for my team. Does that site only have user stats for the top teams, or do you need to request that they be kept?
Zyzzyx
Posted: Mon Jul 28, 2003 10:50 am
by miker
You have to break into the top 100 for that stuff to work.
Posted: Mon Jul 28, 2003 2:01 pm
by Zyzzyx
Ah... gotchya. At our current rate of ~200pts/day, probably won't be reaching top 100. O'ell...
Posted: Mon Jul 28, 2003 2:27 pm
by Wrah
You can allways join SPCR.
Resistance is futile.
Posted: Mon Jul 28, 2003 2:30 pm
by Kostik
Here is a simple "hack" : tell your users to fold for team #31574, and soon they will see their stats
.
Re: Post your PPD here!
Posted: Mon Jul 28, 2003 3:08 pm
by lenny
riff wrote:i'm trying to figure out points per day based on type of processor... i know i can get this with stats, but i can't get the specs of your folding boxen there... so, post your (average of seven days) PPD as of now, along with the boxen producing 'em.
Hi, Riff,
I think you need to break down by work type, operating system and CPU for this to be more meaningful. For example, Gromacs seem to return more points / day than Tinkers. And there's a discussion about Windows code running faster than Linux (hence some people run F@H Windows client under Wine in Linux).
The information is available from FAHlog.txt - you'd probably want a script to process it though. Or search the web for someone who's already done it in PERL
Posted: Tue Jul 29, 2003 12:55 am
by riffst3r
yeah.... i've heard the linux client is about 30% slower (or was it that the windows client is 30% faster?)... but i've also read somewhere, that with gromacs, there's hardly any difference.
i just wanted a quick and dirty calculation so i'll get some kind of idea which cpu's are folding efficient.
Posted: Tue Jul 29, 2003 2:12 am
by riffst3r
alright, i'm going to list out our 'one active processor' folders, and if your name happens to be on the list, then please reply with your boxen's specs and os. i've also taken the time to find out if your WUs are gromacs, tinker or both. if its both, then i'll include a ratio of 'tinker : gromac'. it'll be a lot easier this way, i guess.
from bottom up:
Code: Select all
1) Siili : ................. 18.7 PPD (TINKER)
2) Graham_Astor : .......... 43.1 PPD (GROMAC)
3) xsoverload : ............ 25.6 PPD (TINKER)
4) Landroval : ............. 19.7 PPD (TINKER)
5) MaYak311 : .............. 22.1 PDD (TINKER/GROMAC 1:3)
6) FRODO : ................. 10.3 PPD (TINKER/GROMAC 2:1)
7) ahao : .................. 35.5 PPD (GROMAC)
8) vortex222 : .............. 4.0 PPD (TINKER/GROMAC 1:1)
9) Aristide1 : ............. 37.9 PPD (TINKER/GROMAC 1:4)
10) netguru : .............. 13.1 PPD (TINKER/GROMAC 1:2)
11) CHHAS : ................ 10.2 PPD (GROMAC)
12) monkiman : ............. 30.9 PPD (TINKER/GROMAC 3:6)
13) BillyBuerger : ......... 50.0 PPD (TINKER/GROMAC 3:6)
14) Adam_Mills : ........... 29.0 PPD (GROMAC)
15) Silicon_Acoustics : .... 61.3 PPD (TINKER/GROMAC 3:7)
16) Goatie : ............... 21.0 PPD (TINKER/GROMAC 1:2)
17) raygun : ............... 22.3 PPD (TINKER/GROMAC 3:6)
18) Fingle : ............... 53.7 PPD (TINKER/GROMAC 3:5)
19) rollomooro : ........... 40.8 PPD (GROMAC)
20) changed : .............. 34.3 PPD (TINKER/GROMAC 3:4)
21) Mynci : ................ 31.8 PPD (TINKER/GROMAC 2:4)
22) Kostik : ............... 31.4 PPD (GROMAC)
23) Radeonman : ............ 42.3 PPD (TINKER/GROMAC ?:?)
24) XtalMan : .............. 35.5 PPD (GROMAC)
25) DaShiv : ............... 55.5 PPD (TINKER/GROMAC 1:7)
26) herosformula : ......... 64.6 PPD (TINKER/GROMAC 1:8)
27) rpc180 : ............... 57.8 PPD (GROMAC)
28) jinu117 : .............. 42.8 PPD (TINKER/GROMAC 2:7)
so.. all of you please post your box/os details here. and anyone else who'd like to contribute is most welcome to do so
i can't believe i'm actually researching research.
dukla2000, thanks for your info
NeilBlanchard, that's a very impressive 90PPD for about 2.7Ghz of amd power
Posted: Tue Jul 29, 2003 1:02 pm
by aphonos
riff wrote:alright, i'm going to list out our 'one active processor' folders, and if your name happens to be on the list, then please reply with your boxen's specs and os. i've also taken the time to find out if your WUs are gromacs, tinker or both. if its both, then i'll include a ratio of 'tinker : gromac'. it'll be a lot easier this way, i guess.
If you really want to research researching you could search through the forums for system specs on quite a few folks on your list.
Many of them have them listed in their sig and a few have systems in the gallery. (of course that doesn't mean that those are the systems they are using to fold.)
Posted: Tue Jul 29, 2003 1:05 pm
by POLIST8
AMD 2k+ CPU
1GB DDR 333
120GB 7200.7
120GB ATA V
KX7-333 Mobo
Radeon 7500
Posted: Tue Jul 29, 2003 3:08 pm
by frodo
The machine that has generated the truly phenomenol 10.3 PPD is:
Duron 650 OCed to 848
256MB
Windows 2000
I let it do whatever it wants (Tinker, Gromacs) because I figure both need to be done and I know that I'm not a major points contender anyway. Although I was one of the first to sign up for the team.
However the new machine just arrived (Intel 2.8c, 1GB) and it will be running the client in the next day or so, here comes a little more power.
Posted: Tue Jul 29, 2003 3:37 pm
by riffst3r
aphonos, thanks a bunch for that tip... i haven't really thought about it like that
POLIST8, i'll wait a few more days till i calculate your PPD, right now its about 27, but i have a feeling its gonna be a lot higher than that.
frodo, thanks a lot
Posted: Wed Jul 30, 2003 4:55 pm
by CharlieChan
These boxes are all calculated using EMIII,
gromac
P341 - P3 1G win2k - 31 ppd
P362 - P3 866 redhat 9/wine - 32.95 ppd
tinker
P678 - K7 750 ltsp client - 15.8 ppd
P638 - P3 733 ltsp client - 10.58 ppd
P638 - P3 700 ltsp client - 9.74 ppd
P638 - duron 1.3G ltsp server - 27 ppd
Looks like running linux/wine gives the most ppd
.
But need to verify by running the same project.
When running tinker core the AMD has the advantage.
It would be interesting to compare win2k and linux with tinker core.
Posted: Wed Jul 30, 2003 5:29 pm
by herosformula
whoa, I have the highest PPD on your list.
P4 2.53 GHz (not overclocked); winXP home; folding@home text client
Posted: Thu Jul 31, 2003 3:37 am
by riffst3r
CharlieChan, thanks a lot for those numbers, they call for some very detailed analysis. for the linux vs windows... i thought the windows client was faster because it had a stronger fortran base... and i'm not sure if using wine under linux would change anything. but your numbers prove the exact opposite.
herosformula, i was expecting yours to be amd based.. i'm very surprise
My actual average
Posted: Sat Aug 02, 2003 7:58 pm
by NeilBlanchard
Hello:
Now that I'm switched one fast machine over to SETI full time, me remaining 2.7gHz (two machines) is averaging 64+ P/D. And that is not counting the TWO units that have not been uploaded....
Posted: Sat Aug 02, 2003 8:23 pm
by Kostik
riff wrote:
Code: Select all
<...>
22) Kostik : ............... 31.4 PPD (GROMAC)
<...>
so.. all of you please post your box/os details here. and anyone else who'd like to contribute is most welcome to do so
Athlon XP 1800+.
512Mb DDR 2100.
Win XP pro, console client.
Link in my sig for details.
Considering that I sometimes use my computer for relevant stuff (like playing UT 2003), I'd say it's folding 20/24 hours.
Posted: Sun Aug 03, 2003 9:36 am
by herosformula
[i]herosformula, i was expecting yours to be amd based.. i'm very surprise[/i]
Ever since the SSE version of gromacs, it has been moving along fairly quick. When I get one of those older tinkers, everything slows down considerably (2 times as long 1/3 the points)
Posted: Thu Aug 07, 2003 6:48 pm
by aristide1
Aristide1
PIII 1.2GHz, 512MB RAMM (half used during folding)
W2K, CLI version.
When I get one of those older tinkers, everything slows down considerably (2 times as long 1/3 the points)
Tinkers suck.
Posted: Thu Aug 21, 2003 8:08 pm
by rpc180
stock P42.2Ghz 400 FSB. 512mb PC2100 Ram at 2-2-2-2-7. Windows XP Professional / FireDaemon Service Program with FAH Text Client. Runs 24/7.
Posted: Thu Aug 21, 2003 8:10 pm
by rpc180
stock P42.2Ghz 400 FSB. PC2100 Ram at 2-2-2-2-7. Windows XP Professional / FireDaemon Service Program with FAH Text Client. Runs 24/7.
Posted: Fri Aug 22, 2003 11:00 am
by herosformula
The EOC stats page currently shows that my computer has turned in 158 points in the last 24 hours - boggle.
Posted: Sun Sep 14, 2003 3:21 am
by riffst3r
thanks everyone
i think i got enough to pick out which hardware to setup on.