View Full Version : I got a DG2360!
TollhouseFrank
12-08-05, 10:19 PM
that's right! the full 1294 pointer!
It's done 2 frames so far on my A64 3200+...
so far at 284 ppd... with a little over 1 hour 5 minutes on each frame!
Guys from F(OC)LD... we are gonna see at least a 200 PPD drop in our points till this thing lands from my machine.
that's right! the full 1294 pointer!
It's done 2 frames so far on my A64 3200+...
so far at 284 ppd... with a little over 1 hour 5 minutes on each frame!
Guys from F(OC)LD... we are gonna see at least a 200 PPD drop in our points till this thing lands from my machine.
:eek: :D I didn't even know there was 1294 pointer until u said so
What system do u have folding it (specs) ? What flags is your client using? I WANT ONE! lol
TollhouseFrank
12-08-05, 11:29 PM
its sorta luck of the draw. You HAVE to have -advmethods on to have a chance to get it.
With P4's.. if you have -advmethods and BP's on... you'll still get mostly QMD's. This unit is not considered a BP (its got a small memory footprint... it just takes forever to fold, thus the huge points). With A64's... with -admethods and BP's on... you get a mix of 364 pointers, GbGromacs, and DGromacs like this one. With AMD's and all other cpu types set to -advmethods WITHOUT BP's... you are more likely to get one of these it seems, but not by much. If you are running without -advmethods (like my wife's sempron so her machine can get a steady diet of 600 pointers which no longer require -advmethods).... you will never see one of these.
its sorta luck of the draw. You HAVE to have -advmethods on to have a chance to get it.
With P4's.. if you have -advmethods and BP's on... you'll still get mostly QMD's. This unit is not considered a BP (its got a small memory footprint... it just takes forever to fold, thus the huge points). With A64's... with -admethods and BP's on... you get a mix of 364 pointers, GbGromacs, and DGromacs like this one. With AMD's and all other cpu types set to -advmethods WITHOUT BP's... you are more likely to get one of these it seems, but not by much. If you are running without -advmethods (like my wife's sempron so her machine can get a steady diet of 600 pointers which no longer require -advmethods).... you will never see one of these.
Wow u really know your stuff huh? That's cool. Thx for the info bout that also.
I also didn't know that u don't need -advmethods to get 600 pointers ne more, hmm may be the family rig can benefit from that.
Thx again Frank
TollhouseFrank
12-08-05, 11:48 PM
no probs man. I didn't know the thing about 600 pointers myself until about 3 weeks ago after scrounging HARD around a lot of forums for why 600 pointers were suddenly gone. It's just that they no longer required -advmethods
TollhouseFrank
12-09-05, 12:11 AM
to give you an idea of how long this sucker folds.... remember... i'm doin' frames at 1hour 5 minutes a frame. 100 frames.
THis particular DGromac?
Project 2360
Credit: 1294
Downloaded: December 9, 01:41
Prefered Deadline August 03, 02:41 (NEXT AUGUST?!?!)
Final Deadline November 29, 01:41 (A FULL YEAR??!?!?!)
ghettocomp
12-09-05, 04:40 AM
:eek: :eek:
I don't think so...
Give it a week, should be done by then :D
threeme2189
12-09-05, 04:57 AM
thats still a long time to fold.
AlabamaCajun
12-09-05, 07:30 AM
How much ram and OC do you have on that A64. I got one last week took 34-36 hours on my 3700 with DDR500. You should see it in under 48hours that will be a little over 600 ppd. :burn:
DGromacs produce between 150 and 160ppd/GHz on A64s. No way to get 600 ppd out of them.
TollhouseFrank
12-09-05, 10:30 AM
ok... am 13% complete with this one now. Here are the PPD i am getting according to FAHMon:
Min. Time / Frame : 1h 05mn 33s - 284.27 ppd
Avg. Time / Frame : 1h 06mn 35s - 279.85 ppd
and ETA on finishing date? 4Days 26 Minutes
ok... am 13% complete with this one now. Here are the PPD i am getting according to FAHMon:
Min. Time / Frame : 1h 05mn 33s - 284.27 ppd
Avg. Time / Frame : 1h 06mn 35s - 279.85 ppd
and ETA on finishing date? 4Days 26 Minutes
Do u run that comp 24/7? If so then it shouldn't be all that bad IMO (my comp isn't on 24/7 so yeah)
TollhouseFrank
12-09-05, 10:58 PM
yeah... its been 24/7 for about 4 months now.
AlabamaCajun
12-10-05, 12:29 AM
DGromacs produce between 150 and 160ppd/GHz on A64s. No way to get 600 ppd out of them.
I could not find the one I did before but I went back and checked a 2355 went through yesterday. It's 683pts and was completed in 1day:17 hours and some minutes for just under 400 PPD. This Rig has a memory bandwidth of 7.2G running at 261Mem, 242FSB X11 at 2.664G. 48-50C. (Yes thats an bump up divider on the memory, OCZ :D ).
TollhouseFrank
12-10-05, 12:33 AM
thing is..... i don't have my rig OC'd. its not folding stable if i oc it.
AlabamaCajun
12-10-05, 01:20 AM
thing is..... i don't have my rig OC'd. its not folding stable if i oc it.
Bummer! Don't you just hate when some rigs won't OC. My XP won't even run stable 2 meg over stock. The OC is not as much a factor as the tight latancies of OCZ ram.
TollhouseFrank
12-10-05, 01:22 AM
well.. i have Corsair Value Select. I know it can OC like mad with the right voltage. I also know my 3200+ can OC from 2.2Ghz up to nearly 2.8 or 2.9 if i had the right board... but I have a board with no pci/agp lock... and i'm VERY limited in my OCing options.
darkcrystal
12-10-05, 06:19 PM
im working on a dg 868 pointer right now :) about 35 mins per frame when im doing nothing, just under 40 when im doin stuff. got about 53 hours left. heres the kicker:
i DONT have -advmethods tag on :) and i dont have BIGwu enabled, either. im up to 314 ppd for this bad boy :) should be interesting to see my 24 hour avg after i finish this one up
It's been over a week and the 2360 on the sempron still isn't done yet :D
TollhouseFrank
12-12-05, 07:01 AM
LOL. I'm on my 4th day or so of this 1k pointer. 20 hours to go according to FAHMon
@Frank,
Corsair Value Select is great ram for the money. It's overclockability is highly variable. Corsair like most of the major manufacturers changes chip manufacturers regularly based on price. It all meets the stated spec but beyond that it's a crap shoot depending on just how good the least expensive ram chips on the market were.
TollhouseFrank
12-12-05, 10:54 AM
@Frank,
Corsair Value Select is great ram for the money. It's overclockability is highly variable. Corsair like most of the major manufacturers changes chip manufacturers regularly based on price. It all meets the stated spec but beyond that it's a crap shoot depending on just how good the least expensive ram chips on the market were.
ah. That i didn't know.
However.. this RAM i have isn't stable on 1:1 over 210 unless i up it to 2.75
I gleaned this off Anandtech, which by the way is one of the webs best sources of information. Corsair wouldn't give Anandtech any Value Select to test and, as I recall, asked not to be included in the value ram roundup. They felt they couldn't win even if they happened to be the best in the test as the next batch may not be as good and people would complain when they didn't get the expected results. If they lost, then people wouldn't buy it. I've used it a lot and never had a problem with it. When you think about it, if value ram oc'd reliably to a high oc, why buy expensive ram?
Edit: Found Quote from Corsair and Anandtech on Anandtech
"Our policy is not to send Value Select parts for review for a variety of reasons. After a quick discussion here we decided to stick with that policy and sit this one out." When we asked for clarification, Corsair expressed concern that we would be testing performance and overclocking of their Value products when memory chips change frequently on these products. The concern is that the overclocking performance we would report in a roundup today might not be what a customer would get tomorrow because chips change frequently in value products. Corsair recommended that customers looking for overclocking and enthusiast level performance should purchase Corsair products that are geared to performance and overclocking, where results are consistent.
TollhouseFrank
12-13-05, 12:20 AM
ah. So THAT"s why. Thanks for that info ChasR.
Man... it's like you have the whole internet stuffed in your head!
TollhouseFrank
12-13-05, 08:54 AM
*crying*
NO!!!!!!!!!!!!!
[07:31:08] Completed 440000 out of 500000 steps (88)
[08:01:08] Timered checkpoint triggered.
[08:31:09] Timered checkpoint triggered.
[08:37:22] Writing local files
[08:37:22] Completed 445000 out of 500000 steps (89)
[09:07:23] Timered checkpoint triggered.
[09:16:12] - Autosending finished units...
[09:16:12] Trying to send all finished work units
[09:16:12] + No unsent completed units remaining.
[09:16:12] - Autosend completed
[09:30:14] Quit 101 - Fatal error: NaN detected: (ener[13])
Now i KNOW My rig is stable. *sigh*........... This just sucks. I fold all week hoping for a big dump, and now i'm probably gonna get 5-6 points for it.
nope... 1,161 points!
I guess Stanford knows these are unstable or somethinmg.... and give us points for what we folded.
Misfit138
12-13-05, 09:15 AM
*crying*
NO!!!!!!!!!!!!!
[07:31:08] Completed 440000 out of 500000 steps (88)
[08:01:08] Timered checkpoint triggered.
[08:31:09] Timered checkpoint triggered.
[08:37:22] Writing local files
[08:37:22] Completed 445000 out of 500000 steps (89)
[09:07:23] Timered checkpoint triggered.
[09:16:12] - Autosending finished units...
[09:16:12] Trying to send all finished work units
[09:16:12] + No unsent completed units remaining.
[09:16:12] - Autosend completed
[09:30:14] Quit 101 - Fatal error: NaN detected: (ener[13])
Now i KNOW My rig is stable. *sigh*........... This just sucks. I fold all week hoping for a big dump, and now i'm probably gonna get 5-6 points for it.
nope... 1,161 points!
I guess Stanford knows these are unstable or somethinmg.... and give us points for what we folded.
NaN!!!!!!!!Damn you, NaN!!!!!!!!!!! If I remember correctly, you might be right, there seems to be a little disclaimer after that error that states that the error may be a result of the simulation itself.
I'm having NaN trouble with one of my Bartons...still working on it. Folding is certainly the ultimate stability test!
Nexus Realized
12-13-05, 01:00 PM
That sucks Frank, but 1,100 is a nice payload. :) :thup:
I could not find the one I did before but I went back and checked a 2355 went through yesterday. It's 683pts and was completed in 1day:17 hours and some minutes for just under 400 PPD. This Rig has a memory bandwidth of 7.2G running at 261Mem, 242FSB X11 at 2.664G. 48-50C. (Yes thats an bump up divider on the memory, OCZ :D ).
400/2.66=150ppd/GHz which is in the expected range for a DGromac on an A64.
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.