Thread: Recent WU's higher credits?
January 17th, 2011, 02:03 AM #1
Recent WU's higher credits?
Has anyone else seen a decent increase in the WU credits? In the past two days my bigger, faster rigs have been getting jobs with 787 points as the lowest and 1482 points is common. This is not a rig I have set for bigadv nor am I running any of the beta clients. Just curious.
On another note, is it just me or does folding REALLY frag an HDD? Since I noticed the degradation I have been watching and fragmented sections are huge on a 320GB sata drive.
Has anyone ever encountered HFM.net? I downloaded it and am running my jobs on a dual GTX 295 machine, I used the same user name and PW to create the account but 5-6K points but the points seem to be going to Harlan357?? Any input appreciated.
Running into the opposite of issue No 1 on a GTS 250 rig. Getting jobs with 29.9 PPD worth 74 credits? Are they losing it at Standard or am I Due for a cleanup and re-install?
January 17th, 2011, 03:54 AM #2
Can't really answer any of that myself, as I haven't really been that active with F@H for either team since this spring, hoping to make a bounce back soon with the GTX 470 which should be here tomorrow or Tuesday, Still have the GTX 285, spare 9600GSO, GT 220 and if I can fix the unstable GTX 260 coming my way, as well as that 9800GTX+ from another friend with the one cap that fell off the board (he completely forgot to mail it out to me awhile back).
But if what your saying is the case with Folding, Fragging the Hard drive, might make sense to have it have its own Hard drive, separate from the OS Drive, this way you can defrag the HDD for Folding on a regular basis and not have to worry about the main Windows/Linux drive as much.i7 940//Corsair H60//EVGA X58 SLI LE//6GB Corsair Vengeance 1600MHz//2x EVGA GTX 560 Ti FPB SLI//NZXT Hale82 850W//CM 690 II Advanced//Win7 64//WD 74GB V-raptor, 750GB Black, 1.5TB Green
TechIMO Folding@home Team #111 - Crunching for the cure!
January 17th, 2011, 06:45 AM #3
- Join Date
- Nov 2006
- Portsmouth, UK
- Blog Entries
I cant say ive seen a change in ppd, i would double check your GTX295 clients and make sure its folding in your name, I prefer hfm to fahmon. I still think those GTS250's dont like folding well, i would probably sell it and get a new card.
January 17th, 2011, 07:12 AM #4
I've numbered the questions posed to enable me to answer more easily.
1) I can't tell whether you're referring to CPU or GPU units, but I can't really comment too much on either, as the vast majority of my Folding is bigadv WUs. I still crunch on WKS02's 250 and I've read that others get the sort of points that you refer to, but then it would be for GPU WUs. It would also be handy to know what hardware this is Folding on, as you might be getting GPU3 WUs, which are more likely to have bonus points available - which have deadlines.
2) Folding itself doesn't really frag a drive, as such. It will be true that there will be some fragmentation in the folder/s that have Folding client/s in, but it would be limited to the "Work" folder - and that would be cleared out every time a WU is finished and submitted... Obviously, you're unlikely to know (without using proper HDD tools) where the fragmentation is, as the default free ones will only analyse the entire partition, but (if you defrag it) you shouldn't see much fragmentation outside of the "Work" folders.
3) Yes, I've got HFM on just about all of my [M$] computers, and it works perfectly. The person who wrote the code for it was Ryan Harlamert (online/Folding tag: harlam357) and the default user that it uses will be showing as his details. You can change that in the "Stanford User ID:" and "Extreme Overclocking User ID:" sections of the Preferences box, under the "Web Settings" tab - where you can also set the "Stanford Team ID:", then if any of the user and team details listed in the config files don't tally up, it will flag it up with orange as the background colour.
4) Without knowing more about what WUs you're getting, I couldn't say - but it is possible that it is trying to pull advmethod WUs for a GPU that can't complete them in time to get the bonus points...check when the deadline is listed as being.
From the two attached pics, you can see that: one is where I'm showing the user/team details area so that you can change it to you're Folding name; and the other is where my PPD for WKS04's SMP client (showing in the details area, as it is the item selected) is only expected to get the base points [due to getting a BSOD and having a power outage, but this should be sorted, as I've got my HFM client set to show the effective rate so it is guaging the overall time...which (including downtime) takes it to a couple of hours after the deadline - but will be sorted as it isn't taking an hour to get each frame done] rather than the ~35k PPD it was last night. Once the average TPF (Time Per Frame) is brought back down, it will show that it will complete in time to earn the bonus points, and the PPD will increase from the 2200 PPD and credit of 8955 to being somewhat higher.....but I can't say what it would be, as there's too long left to run, and it hasn't calculated it properly again yet..
January 17th, 2011, 10:16 AM #5
Thanks for the response. I have an ATI 4970 with a cap that fell off as well. Tried soldering it back myself with no luck but didn't push it, didn't want to chance of frying a trace or something worse.
I have a couple of new drives in the cabinet so I may go ahead and dedicate a drive to Folding ( see the response from NLM below yours too )
Finally made it above 3 million points. Plan to build a BIG folding rig within the next month so I can break 4 mil faster.
Good luck with the 470!
January 17th, 2011, 10:29 AM #6
Great to hear from you and as always, you have provided some great data points.
The only WU's I was referencing were GPU ( dual GTX 295 ) and I HAVE noticed the orange highlights so now I will rectify that.
Have learned a bit about the work folder. If I get an EUE I have discovered that if I delete the files in the work folder I can avoid the 24 hours in the penalty box....
All of the WU's I am getting have deadlines. I didn't realize that indicated GPU3 WU's.
One question. Under Stanford Team ID I thought we were Team 111. Is this another team you are on?
As always, thanks for the education.
January 17th, 2011, 01:32 PM #7
You get the deadlines on all WUs, but if you're running with the advmethods flag set (either manually, or within the config file) then you'll have a MUCH sorter period to get them done in....which is where the bonus points come in... I should have clarified that earlier..
Now that WKS04 has done a few more frames (% of the WU) it is now showing as expected to complete within the time, and will now earn me around 55k points, but that's due to increase with every frame that goes by, as it will be shortening the time before it will complete: by taking less than the (earlier) hour, meaning that the completion time is brought forward by the amount shorter than the average, and by bringing the average down too..
If you take another look at the image of my HFM instance, you'll see that the ETA was after the [initial] deadline, so didn't get anything for a bonus...
Well spotted, I do indeed Fold for another Team... I'm up to ~6.5 million over at OCn, which means that I've broken into the top 80 on there,.. I've still got Folding clients for here (Team 111) too, but they're ones I don't have any access to any more, so I can't change them or do anything else with them.. I'm still active here, and am prolly more so here than over there, as I can navigate easily around TechIMO with the BlackBerry (and the cr@p dial-up speeds I get with it) and there it just takes forever and a day (or more) just to load each page - and it then loses the previous page from the cache...
If you delete the contents of the Work folder, then you might as well also wipe the "queue.dat" file that will be in the main instance folder (where the Work folder sits, not in the Work folder) as then it won't start to flap and whinge when you start it off again...
January 17th, 2011, 08:27 PM #8
All good information. Especially about wiping the "queue.dat" file,,,
Where can I access the config file? I have two folding instances highlighted in Orange that inform me there are two instances folding with the same Machine ID. I knew how to set up the original config file but,,,have never learned where to access it again to make an changes.
January 18th, 2011, 02:03 AM #9
There are two ways, one will be the "config.txt" (or possibly "config.cfg", I can't recall now, it could even be "client.cfg") which you can open with your preferred text editor, but the better (for those who aren't 100% about things) way to do it would be to open a CLI instance, navigate it to the Folding instance's [main] folder, then run the "Folding@Home....exe" file. For you, from the one I set up for you would be to:
Start --> Run --> type 'cmd' [Enter]
'cd C:\F@H\GPU1' [Enter]
'dir' [Enter] (note down the name of the file that starts "Folding@Home"
Type that file name out, followed by a space and '-configonly' [Enter]
Run through the questions and update them as required, most of the ones that would be checked for by HFM wouldn't be in the advanced section - though I think that the instance number is.*
Once you've done, you should stop the F@H instance/s that don't tally with your details, as per what you put into HFM, then start them again. Once HFM gets to the specified update point (which you can set within the HFM preferences) it should update with the new details.
* = IDK whether dual-core GPUs should have two instances of HFM running against it or just one, and I can't remember whether HFM will list all clients [that are listed in it] that have the same instance number on a computer, or just the repeated ones (where the first is ignored) as I haven't seen that in a LONG time.
January 18th, 2011, 09:59 AM #10
This is all good. I will try this today after work.
On the dual CPU question, HFM lists both cores. They are numbered correctly and not in play here. HFM tells me that GPU1 & GPU3 are running with the same ID and, running the same WU.
This should fix that.
January 18th, 2011, 10:15 AM #11
Things may have changed since the CPU clients were installed, as it is better to have a single SMP client (with a passkey) for all available cores than a single/simple F@H client per core...as you'd be able to earn the bonus points for completing the WUs early - presuming you then opted to run with the advmethods flag..
With regards to the multiple GPU instances, you'd be best off waiting for someone else to reply - as I don't have any experience with multi-core'd cards; my 250 and 275 are only single-core GPUs....
January 18th, 2011, 01:37 PM #12
- Join Date
- Nov 2006
- Portsmouth, UK
- Blog Entries
Users Browsing this Thread
There are currently 1 users browsing this thread. (0 members and 1 guests)
By Irv in forum General Tech DiscussionReplies: 1Last Post: August 8th, 2006, 10:58 AM
By Blitzkreig75 in forum Distributed ComputingReplies: 2Last Post: April 20th, 2006, 01:11 AM
By vikeor in forum Processors, Memory, and OverclockingReplies: 14Last Post: February 22nd, 2005, 10:55 AM
By RicheemxX in forum Distributed ComputingReplies: 4Last Post: December 2nd, 2003, 11:49 PM
By ShawnD1 in forum Graphics Cards and DisplaysReplies: 1Last Post: October 2nd, 2003, 01:40 PM