Heehee, it hasn't finished any jobs yet, but it has figured out my PPD. 20,378 total for this system. :twisted:
Printable View
Heehee, it hasn't finished any jobs yet, but it has figured out my PPD. 20,378 total for this system. :twisted:
Give it a couple days to settle - apparently they send out different WU's at first - my GTS450 (I think I might have referred to it as a GTS250 a couple times in this thread lol but it's a 450) was getting around 13k ppd alone initially - after about 3 days it dropped and now sits at around 8k - quite a drop. It's also worth turning off the CPU folding and seeing what you get on just the GPU's as the GPU3 client does use a chunk of CPU cycles - freeing up the CPU to JUST work the GPU3 clients can make a big difference - possibly more than you're getting on the CPU at the moment - check both then choose whichever works for you.
My system gets about 11k now, going with one GPU2 and one GPU3 only running. If I start CPU I get very little from the CPU but the GPUs lose almost 3k points so I end up worse off WITH CPU running than I am without...
I read up about this:
A note on PPD - this will vary from project to project!
bonus points - it's up to 50% extra if you have a passkey, but only after 10 SMP WUs
Yeah which I'm not likely to ever qualify for with this rig - I'm running it purely on GPU which wont qualify lol - bah humbug and all that
Under SMP settings, select 'Use specified number of cores' and enter 3 (or total#-1). Voila! You're now SMP folding with 1 core dedicated to gpu3 assistance and other misc tasks.
To add to this, in the Tracker, if the PPD is yellow it's an estimate, when it's green it's actual. I've noticed the yellow estimates are pretty random sometimes...both high and low.Quote:
Originally Posted by dr.walrus
I'm getting 2801 PPD from SMP on my Q6600 @3.0GHz, and 3034 PPD on my 6870 @ 955/1140, so my CPU is doing well enough. I may toss in that spare 4870 just to fold with, but we'll see
I heard about that but idn if it'll help. Lets find out - right now, GPU2 and GPU3 only, the system is claiming 11,001 ppd - see what happens when I kick in SMP on 3 cores only...
Just for the hell of it, I just started my sons PC running SMP on a Athlon x2 5000 (2.6 ghz? not sure)
See what that pushes out...
I've been getting really varied results from my CPU. As some points I have gotten upwards of 2000 PPD on my Phenom II X3, but recently it has been more like 350-600 PPD, which is about 1/2 what my 1.8 Ghz core 2 duo imac gets.
1) Depends on the WU itself
2) f@h is prioritised so low that any other work on the PC will trash your performance. Running full screen (non-hd)TV, internet, remote desktop reduces me from ~14000 to ~2500, even though those things are very non-processor intensive
3) Are you using the SMP or the normal client? I'm assuming SMP, but that figure seems low on an X3
I'm using the SMP. I have had nothing else running for a few hours, and am getting 386 PPD. On the few previous WU's, I got in the low 500's.
Dropped into the 9,000's so killed the SMP client - I'll let the GPU's do the work on this machine I think heh
My son's PC is barely getting 1000 ppd so I don't think that'll be running long - might leave the software on there to work in the background when he chooses to have the pc on, but 1000ppd isn't really worth the cost of running the PC 24/7
On a seriously positive note - in the last 24 hours we produced almost 75000 points - a rate which is accelerating all the time. At that production rate, we're already in the top 200 teams based on production...
That does seem awfully low, Technochicken. Granted, it's not an exact comparison, but my x6 1090T has been getting around 7,500-8,000 PPD. Given that, I would guess the x3 710 should do around 35-40%.
In college I ran a bunch of little P3's folding with a different project, but once I moved away from there and started paying my own electric bills....yeah, not really worth it. :P IIRC I would get less than 100 PPD out of each of them.
Way to go guys :D
:banana:
Lol... i love how everyone uses all cpu meter....
GAAK!
Just looked at my stats page.
It said I haven't completed any for TBCS.
(No, I haven't folded for the other guys in a long time)
???
if you didnt finish the wu in time you may not have gotten credit for it
I think it said it fully transmitted to the F@H server.....
Ah well. Let's see how this next WU goes........
w00tz! We've passed 4,000! :D
EDIT:
Also, I got my backup system up and running, contributing ~5,200 ppd per GTS250, and ..~200 ppd from the CPU... ...yeah... Oh well, 200 is 200, and all that system is doing is folding, so it should put out a solid ~10,600 ppd. :D
so, interestingly... i looked, and if we keep up our 25k+ rate, we could hit pretty high, alot of the higher teams don't have the points rate we have...
Yeah, that's one thing I've been noticing is that a lot of the teams we're blowing past look like they used to be active (or they wouldn't have as many points as they do), but are stagnating if not completely stopped, going off their average 24-hour rate. Looking at the stats, I don't think we'll see any real competition until we get well past the 1,000. Even there, the concentration of really active teams is fairly low. ..that being said, it will probably take us a little while to get there, simply because of the sheer number of points that teams at that level have. Even if the EOC estimate is right (which it definitely won't be if we keep getting more power at this rate :twisted: ), it'll probably take us a few months to get to that point (current team 1,000 has 5,479,363 points atm, vs our current 735,381).
hmm if we get 50 people to at least do 8000 ppd we can catch up in about 15 days :P
I joined in, folding nice and slowly on my laptop! lol intel t6500 ftw!!
gonna throw my x2 4200 - 4870 in the mix. (maybe someday I will have my i7-960 running, and when not playing games I'll have it folding)
on a side note, would my 8600gts 512mb fare better than my 4870 1gb? nvidia seems to be much better at folding... (my limited HP mATX mobo only has one pci-express slot lol)
You should see between 3 and 4k from the 4870 - the 8600 looks like it'd be under 2k but might be worth trying
i have my gtx 460 folding
man... my 4870 is killing my x2 athlon... lol. getting about 2,500 ppd on it so far... the stupid thing is cooking though, running around 90c so I throttled it to only using 80% of the gpu.
EDIT: my lappy is kicking out about 1400ppd too, not too shabby, better than I expected!
I have to give my 4870 some cooling kudos - it's sitting in the mid-60's temp wise, having been running folding 24/7 for over a week - and the fans are below 60% too.
Gainward HD 4870 Golden Sample Edition 512MB GDDR5 DVI HDCP PCI-E Graphics Card
lol...sometimes I forget what air temps are like...I was upset that one of my 260's was hitting 54C. :P ...forgot to replace the quicky-generic-thermal-paste job with AS5 before reassembling...otherwise it would be in the mid 40's.
Well my brother in-law just converted his desktop to an HTPC since he has a bad ass MSI gaming laptop now. So I will be throwing my 8800 GTS 512 and a GT240 in my HAF X along with my 2 5870s and my Phenom II X4 965 BE @4.01 Ghz Should be good for about 20k PPD
I don't have time to figure out the MP software, but if we can get the clients working right, I will have 160 cpu's to fold on for a couple of days (and possibly double that). I can justify it by calling it a "burn in". The ETA for these systems is the end of April and if I time it right I can have them fold, er uh, Burn IN, for a long weekend.
Here is a link to the hardware:
http://www.supermicro.com/servers/blade/TwinBlade/
10 twin blades = 4 cpu's per chassis.
Right now the order is for 4 chassis but may get bumped to 8!
To get the F@H software working might take access to a MP system, and I can ask my IT guy to setup a VPN tunnel to an MP machine at my desk for someone to check out the client with.
If anyone is interested, let me know.
EDIT: We will be using the E5640's...
WOW That would be pretty awesome!
Well I have tried to install the F@H Tracker and every time it begins downloading the clients it will not start the CPU/SMP client download. I have had this problem on 4 different machines now, 2 of them being on 10Mb connections. Anyone else have problems?
When I set up my second machine, it wouldn't start the CPU client download until after the GPU clients were already downloaded, but other than that, no. Strange.
It would dl both of the GPU clients and would try to start the CPU for about 90 seconds then would throw up a "Unable to download client(s)" error in the log window.
Weird....well, one workaround for that would be to run it once on a system that the download works, then copy that whole directory to other systems. Then you just have to start it and configure it for each system instead of having to re-download it each time.
On another note, I was wondering about something. Back when I was folding with BOINC, they had stats pages for each participant, that showed what projects they were contributing to, how much they had contributed to each project, what systems they were using, recent production stats for those systems, and so on. Is there anything like that with FAH? Mainly, I want some sort of way to verify that all of my systems are actually checking in WUs...heck, a few times, the only way I knew a system had died was because it stopped submitting WUs. :P
I think that we get those stats once we reach the top 800 teams
How big is the directory folder? Someone zip it and email it to cgantt@cjganttphotography.com
I ran across this tool. It's handy for keeping track of your various clients. You just point it at the install directories for all your clients and it checks the FAHlog.txt files every now and then, and figures out the various statusy-stuff.
http://fahmon.net/
What I did for monitoring my other systems is just have the FAH stuff in a shared directory, then mount the share on my main system, and point fahmon at that. ..it does take a while to figure out some of the columns...but I don't think that's any different with the tracker either.