SOLVED:Where are my points going?

Moderators: Site Moderators, PandeGroup

SOLVED:Where are my points going?

Postby djohnston » Fri Nov 25, 2011 12:46 am

A couple months ago, I lost the motherboards of my two main computers. While waiting for replacements, my PPD average dropped like a rock to zero. I finally got my PPD average up to about 2500 about a week ago. All six clients are still running 24/7, except for my PentiumIII which lost wireless connection from Nov 21st throught Nov 24th. It contributes probably less than a tenth of total points. All have the same configurations. Yet, my PPD dropped to around 900 this morning, and has since "risen" to about 1200. That's about half of what it was a week ago. I've gone through the logs as far back as they go to copy the completed WUs. No work units have failed to complete.

Code: Select all
User name: djohnston (Team 50619)
User ID: 225CD54549D835CC

Client 1
Project: 7600 (Run 15, Clone 8, Gen 15)
Number of Units Completed: 1132
Project: 6892 (Run 946, Clone 4, Gen 15)
Number of Units Completed: 1133
Project: 6892 (Run 892, Clone 5, Gen 42)
Number of Units Completed: 1134
Project: 6892 (Run 892, Clone 5, Gen 43)
Number of Units Completed: 1135
Project: 6892 (Run 892, Clone 5, Gen 44)
Number of Units Completed: 1136
Project: 7600 (Run 38, Clone 18, Gen 20)
Number of Units Completed: 1137
Project: 6892 (Run 615, Clone 3, Gen 20)
Number of Units Completed: 1138
Project: 6892 (Run 809, Clone 19, Gen 32)
Number of Units Completed: 1139

Client 2
Project: 7600 (Run 9, Clone 7, Gen 16)
Number of Units Completed: 1122
Project: 7600 (Run 19, Clone 5, Gen 18)
Number of Units Completed: 1123
Project: 6892 (Run 230, Clone 14, Gen 22)
Number of Units Completed: 1124
Project: 6892 (Run 913, Clone 19, Gen 23)
Number of Units Completed: 1125
Project: 6892 (Run 680, Clone 9, Gen 27)
Number of Units Completed: 1126
Project: 6892 (Run 680, Clone 9, Gen 28)
Number of Units Completed: 1127
Project: 6892 (Run 680, Clone 9, Gen 29)
Number of Units Completed: 1128
Project: 6892 (Run 680, Clone 9, Gen 30)
Number of Units Completed: 1129
Project: 6892 (Run 945, Clone 16, Gen 27)
Number of Units Completed: 1130
Project: 7600 (Run 32, Clone 15, Gen 16)
Number of Units Completed: 1131

Client3
Project: 7026 (Run 3, Clone 84, Gen 55)
Number of Units Completed: 852
Project: 6892 (Run 620, Clone 8, Gen 16)
Number of Units Completed: 853
Project: 7600 (Run 26, Clone 60, Gen 10)
Number of Units Completed: 854
Project: 7003 (Run 2, Clone 56, Gen 49)
Number of Units Completed: 855
Project: 6872 (Run 841, Clone 3, Gen 41)
Number of Units Completed: 856
Project: 7016 (Run 3, Clone 25, Gen 35)
Number of Units Completed: 857
Project: 6892 (Run 618, Clone 10, Gen 23)
Number of Units Completed: 858
Project: 7015 (Run 0, Clone 9, Gen 52)
Number of Units Completed: 859
Project: 7013 (Run 3, Clone 0, Gen 54)
Number of Units Completed: 860
Project: 7004 (Run 3, Clone 7, Gen 57)
Number of Units Completed: 861
Project: 6892 (Run 799, Clone 14, Gen 22)
Number of Units Completed: 862
Project: 6892 (Run 71, Clone 18, Gen 15)
Number of Units Completed: 863
Project: 6892 (Run 598, Clone 2, Gen 33)
Number of Units Completed: 864
Project: 7600 (Run 29, Clone 2, Gen 14)
Number of Units Completed: 865
Project: 7600 (Run 31, Clone 38, Gen 11)
Number of Units Completed: 866
Project: 6892 (Run 661, Clone 15, Gen 25)
Number of Units Completed: 867
Project: 6892 (Run 481, Clone 8, Gen 36)
Number of Units Completed: 868
Project: 6892 (Run 481, Clone 8, Gen 37)
Number of Units Completed: 869
Project: 7600 (Run 8, Clone 6, Gen 16)
Number of Units Completed: 870

Client 4
Project: 6884 (Run 646, Clone 4, Gen 67)
Number of Units Completed: 177
Project: 6888 (Run 95, Clone 9, Gen 23)
Number of Units Completed: 178
Project: 6871 (Run 641, Clone 1, Gen 41)
Number of Units Completed: 179
Project: 6882 (Run 384, Clone 4, Gen 93)
Number of Units Completed: 180
Project: 6892 (Run 910, Clone 13, Gen 24)
Number of Units Completed: 181
Project: 6892 (Run 382, Clone 9, Gen 28)
Number of Units Completed: 182
Project: 6892 (Run 771, Clone 10, Gen 26)
Number of Units Completed: 183
Number of Units Completed: 184


Client 5
Project: 6892 (Run 827, Clone 3, Gen 19)
Number of Units Completed: 677
Project: 6892 (Run 827, Clone 3, Gen 20)
Number of Units Completed: 678
Project: 6872 (Run 664, Clone 0, Gen 48)
Number of Units Completed: 679
Project: 6888 (Run 320, Clone 8, Gen 13)
Number of Units Completed: 680
Project: 7600 (Run 2, Clone 30, Gen 8)
Number of Units Completed: 681
Project: 6892 (Run 804, Clone 14, Gen 7)
Number of Units Completed: 682
Project: 6884 (Run 680, Clone 1, Gen 114)
Number of Units Completed: 683
Project: 6892 (Run 586, Clone 18, Gen 22)
Number of Units Completed: 684
Project: 6872 (Run 17, Clone 8, Gen 15)
Number of Units Completed: 685
Project: 6871 (Run 112, Clone 8, Gen 28)
Number of Units Completed: 686
Project: 7600 (Run 36, Clone 91, Gen 10)
Number of Units Completed: 687
Project: 6888 (Run 143, Clone 4, Gen 50)
Number of Units Completed: 688
Project: 6888 (Run 143, Clone 4, Gen 51)
Number of Units Completed: 689
Project: 6892 (Run 450, Clone 16, Gen 14)
Number of Units Completed: 690
Project: 6892 (Run 594, Clone 19, Gen 25)
Number of Units Completed: 691
Project: 7001 (Run 0, Clone 155, Gen 18)
Number of Units Completed: 692
Project: 10085 (Run 1, Clone 13, Gen 56)
Number of Units Completed: 693
Project: 6892 (Run 271, Clone 17, Gen 26)
Number of Units Completed: 694
Project: 6892 (Run 271, Clone 17, Gen 27)
Number of Units Completed: 695
Project: 6892 (Run 271, Clone 17, Gen 28)
Number of Units Completed: 696
Project: 6892 (Run 271, Clone 17, Gen 29)
Number of Units Completed: 697
Project: 6892 (Run 271, Clone 17, Gen 30)
Number of Units Completed: 698
Project: 6892 (Run 271, Clone 17, Gen 31)
Number of Units Completed: 699
Project: 6892 (Run 271, Clone 17, Gen 32)
Number of Units Completed: 700
Project: 6892 (Run 271, Clone 17, Gen 33)
Number of Units Completed: 701
Project: 6892 (Run 271, Clone 17, Gen 34)
Number of Units Completed: 702
Project: 6892 (Run 271, Clone 17, Gen 35)
Number of Units Completed: 703
Project: 6892 (Run 271, Clone 17, Gen 36)
Number of Units Completed: 704
Project: 6892 (Run 271, Clone 17, Gen 37)
Number of Units Completed: 705
Project: 7010 (Run 1, Clone 115, Gen 66)
Number of Units Completed: 706
Project: 7600 (Run 2, Clone 23, Gen 16)
Number of Units Completed: 707
Project: 6892 (Run 819, Clone 3, Gen 34)
Number of Units Completed: 708
Project: 6892 (Run 758, Clone 2, Gen 24)
Number of Units Completed: 709
Project: 6892 (Run 285, Clone 9, Gen 25)
Number of Units Completed: 710
Project: 6892 (Run 169, Clone 9, Gen 37)
Number of Units Completed: 711
Project: 6892 (Run 169, Clone 9, Gen 38)
Number of Units Completed: 712
Project: 6892 (Run 169, Clone 9, Gen 39)
Number of Units Completed: 713
Project: 6892 (Run 169, Clone 9, Gen 40)
Number of Units Completed: 714
Project: 6892 (Run 169, Clone 9, Gen 41)
Number of Units Completed: 715

Client 6
Project: 6888 (Run 639, Clone 9, Gen 14)
Number of Units Completed: 16
Project: 6892 (Run 469, Clone 5, Gen 25)
Number of Units Completed: 17
Project: 6892 (Run 469, Clone 5, Gen 26)
Number of Units Completed: 18
Project: 6872 (Run 506, Clone 0, Gen 28)
Number of Units Completed: 19
Project: 6892 (Run 312, Clone 19, Gen 16)
Number of Units Completed: 20
Project: 6892 (Run 245, Clone 3, Gen 22)
Number of Units Completed: 21
Project: 6892 (Run 169, Clone 11, Gen 20)
Number of Units Completed: 22
Project: 6888 (Run 445, Clone 9, Gen 26)
Number of Units Completed: 23
Project: 6892 (Run 40, Clone 16, Gen 9)
Number of Units Completed: 24
Project: 6892 (Run 40, Clone 16, Gen 10)
Number of Units Completed: 25
Project: 6892 (Run 251, Clone 11, Gen 27)
Number of Units Completed: 26
Project: 6892 (Run 251, Clone 11, Gen 28)
Number of Units Completed: 27
Project: 6892 (Run 906, Clone 19, Gen 30)
Number of Units Completed: 28


So, what's the deal? I know work units can vary by point averages. But, by more than 200% on the same hardware with the same settings?
Last edited by djohnston on Fri Dec 16, 2011 8:00 am, edited 1 time in total.
Image
djohnston
 
Posts: 85
Joined: Sat Feb 06, 2010 8:33 pm

Re: Where are my points going?

Postby Jesse_V » Fri Nov 25, 2011 12:53 am

How are you measuring ppd? Are you doing it manually or using third-party websites? Perhaps they could be at fault.
Pen tester at Cigital/Synopsys
User avatar
Jesse_V
 
Posts: 2894
Joined: Mon Jul 18, 2011 4:44 am
Location: USA

Re: Where are my points going?

Postby k1wi » Fri Nov 25, 2011 2:34 am

I really don't have enough information to give you on your results given the amount of information, but let me summarize what I am taking out of it.

1. You are folding on 6 uni-proc clients.
2. All clients are identically configured, but may have wildly different specifications.
3. PPD is fluctuating, across days and within days.
4. You have recently folded QRB qualifying work units.

What is unclear

1. How PPD is measured i.e. with a third party app, or with a third party website.
2. Whether you are folding with a passkey enabled or not.
3. Whether your computers are dedicated or not (we know they run 24/7).
4. Whether your internet connection is fast or slow.
5. Whether your clients have been having difficulty connecting to the folding servers.

I can think of a number of scenarios where PPD on uni-proc clients will vary, and vary significantly. Number one suspicion is QRB work units. The ppd of QRB work units can be significantly higher than non QRB qualifying work units, depending on the performance of your machine. Alternatively, without a passkey they can reduce the effective PPD. Are your clients configured with a passkey or not? It is possible that while you were experiencing a PPD in the realm of 2500 PPD you were folding QRB qualifying work units on some or all of your machines on a fairly regular basis. Alternatively, have you not added a passkey to your clients and as a result, has this suppressed your ppd figures when folding these work units?

Finally, what is the distribution of PPD contribution - are they all similarly specced machines (other than the PIII)? If one machine is more powerful than the others, any of the above factors influencing that machine will have a much greater effect on total ppd.

Seeing QRB qualifying work units mixed in with non QRB qualifying work units makes me suspect that it is this that is causing your variation in ppd, but I say that tentatively as there is not a huge amount of information to go on - I look forward to any clarification/additional information.
k1wi
 
Posts: 1322
Joined: Tue Sep 22, 2009 10:48 pm

Re: Where are my points going?

Postby gwildperson » Fri Nov 25, 2011 4:50 am

My uniprocessor clients are slow and often take more than a day to finish some projects. Thus, instead of getting point at somewhat regular intervals, I may have a couple WUs finish on one day and none the next day.

I used to get only non-bonus assignments for FahCore78. Then for a while I got mostly bonus WUs for FahCoreA4 and my PPD went up quite significantly. Recently I've been getting mostly Core78 assignments again and the PPD is more in line with the old days. As old projects are completed and new projects come on-line, this problem will go away, but nobody knows how long that transition will take.
gwildperson
 
Posts: 785
Joined: Tue Dec 04, 2007 8:36 pm

Re: Where are my points going?

Postby djohnston » Fri Nov 25, 2011 7:22 pm

k1wi wrote:
What is unclear

1. How PPD is measured i.e. with a third party app, or with a third party website.
2. Whether you are folding with a passkey enabled or not.
3. Whether your computers are dedicated or not (we know they run 24/7).
4. Whether your internet connection is fast or slow.
5. Whether your clients have been having difficulty connecting to the folding servers.



1. PPD is measured by the third party website, folding.extremeoverclocking.com. As I understand it, we should not use stats from folding@home servers, as that would put undue strain on their limited resources.
2. All clients use my personal passkey.
3. Two "lesser" single core computers are dedicated. They basically do nothing else. The two dual cores run folding non-stop, but are also used to browse the web and do word processing.
4. The internet connections are high speed cable connection, connected to a router. Two computers are ethernet, two are wireless.
5. Except for the one computer that lost wireless connection for three days, I found no indication in the logs that any of the others had trouble connecting to work servers to send results or get new work units.

k1wi wrote:
I can think of a number of scenarios where PPD on uni-proc clients will vary, and vary significantly. Number one suspicion is QRB work units. The ppd of QRB work units can be significantly higher than non QRB qualifying work units, depending on the performance of your machine. Alternatively, without a passkey they can reduce the effective PPD. Are your clients configured with a passkey or not? It is possible that while you were experiencing a PPD in the realm of 2500 PPD you were folding QRB qualifying work units on some or all of your machines on a fairly regular basis. Alternatively, have you not added a passkey to your clients and as a result, has this suppressed your ppd figures when folding these work units?



Sorry, I don't know what a QRB work unit is. As stated above, all clients use the same passkey.

k1wi wrote:
Finally, what is the distribution of PPD contribution - are they all similarly specced machines (other than the PIII)? If one machine is more powerful than the others, any of the above factors influencing that machine will have a much greater effect on total ppd.

Seeing QRB qualifying work units mixed in with non QRB qualifying work units makes me suspect that it is this that is causing your variation in ppd, but I say that tentatively as there is not a huge amount of information to go on - I look forward to any clarification/additional information.



Four of the clients run from two identically specced machines. Each machine is an AMD64 running @ 2.7 gHz with 4 GB of DDR2 RAM. All 4 of the computers run the same 32 bit distro. I will be changing to the 64 bit version when it comes out of testing. None of the clients have RAM limits configured. In addition, the configuration for all clients is to accept big work units, accept advanced units and use all idle time.

gwildperson wrote:My uniprocessor clients are slow and often take more than a day to finish some projects. Thus, instead of getting point at somewhat regular intervals, I may have a couple WUs finish on one day and none the next day.

I used to get only non-bonus assignments for FahCore78. Then for a while I got mostly bonus WUs for FahCoreA4 and my PPD went up quite significantly. Recently I've been getting mostly Core78 assignments again and the PPD is more in line with the old days. As old projects are completed and new projects come on-line, this problem will go away, but nobody knows how long that transition will take.


Thanks, gwildperson. I'm getting both. The A4s consist of 1 or 2 million steps. The others consist of 250 thousand steps. We're not supposed to cherry pick the work units, and I never have. Sometimes I wish ALL the units were high yielding ones. After all, there's no guns or glory in all this. We do it for the points, but more importantly, for the beneficial research on disease that this project addresses.

I just wanted to be sure that none of already completed work units were not being counted. I see that's not possible just from the Project, Run, Clone and Gen numbers. I was also questioning why points can vary so wildly from one week to the next when all other factors are the same as before.
djohnston
 
Posts: 85
Joined: Sat Feb 06, 2010 8:33 pm

Re: Where are my points going?

Postby gwildperson » Fri Nov 25, 2011 7:59 pm

djohnston wrote:1. ... I understand it, we should not use stats from folding@home servers, as that would put undue strain on their limited resources.


You should not write a script that queries your official stats every X minutes because that creates a continuous load whether you're actually looking at the data it returns or not. How many times per day do you need to query your stats?

I've never seen a recommendation that we avoid using the official site manually through a browser. Where did you get that information?

The third party sites are nice, and they track your history, but you should use either or both, depending on whether they give you the information you want or not. Avoid saying "should not use" for either.
gwildperson
 
Posts: 785
Joined: Tue Dec 04, 2007 8:36 pm

Re: Where are my points going?

Postby Jesse_V » Fri Nov 25, 2011 8:05 pm

djohnston wrote:1. PPD is measured by the third party website, folding.extremeoverclocking.com. As I understand it, we should not use stats from folding@home servers, as that would put undue strain on their limited resources.


Hmm. I don't know how accurate or reliable that website is, simply because I haven't used it. I use kakaostats.com myself, but that's just my choice. Those third-party websites download information from the F@h servers. So far as I know, your welcome to go to your team page on the F@h website and look at the points yourself. You can compare one day to the next by just writing down the number and comparing that with the numbers the following day. Stanford doesn't want you do download their text file that contains the ENTIRE list of all teams and users, as that file is somewhat large and they don't want that strain. That file is what the third-party sites use, but your team page is perfectly fine. However, the Pande Group hasn't written any PPD information into the stats pages, so if you're watching that you have to go those other sites, or install a F@h client monitor on your computer. (http://fahwiki.net/index.php/Third_Party_Contributions)

djohnston wrote:Sorry, I don't know what a QRB work unit is. As stated above, all clients use the same passkey.


QRB stands for Quick Return Bonus. Certain Work Units, such as those for SMP (and a few for uniprocessor but they're rare) tell the servers to reward you with bonus points for completing a WU quickly. For example, I process SMP WUs, and they have base points, which I get no matter so long as I return the WU before the deadline. However, if I return the WU really fast (like in the third of the time to the Preferred Deadline) then I'm non-linearly rewarded a lot of extra points on top of the base points. They have this system in place to enourage people to complete those hefty WUs quickly, and as a reward for much more of our computing resources. Since the QRB WUs give more points, they could have influenced your PPD, and that's why it was brought up.

djohnston wrote:Thanks, gwildperson. I'm getting both. The A4s consist of 1 or 2 million steps. The others consist of 250 thousand steps. We're not supposed to cherry pick the work units, and I never have. Sometimes I wish ALL the units were high yielding ones. After all, there's no guns or glory in all this. We do it for the points, but more importantly, for the beneficial research on disease that this project addresses.


The number of steps is rather insignificant. What matters is the points the WU is worth. Processing occurs in a loop, so the steps count the total number of iterations that must be performed before the WU can be returned. They can have little bearing on how long your computer will take to complete the WU, as you could have a tiny protein with few atoms to work on, or you could have a larger, more complex protein instead, and that could influence the ETA as well. These and other factors make the steps not very meaningful by themselves to us donors.

djohnston wrote:I just wanted to be sure that none of already completed work units were not being counted. I see that's not possible just from the Project, Run, Clone and Gen numbers. I was also questioning why points can vary so wildly from one week to the next when all other factors are the same as before.


The PRCG numbers identify the WU and make it unique. The official stats pages update every hour on the hour, so if an update happens and you've already completed the WU, that may indicate an issue somewhere since you aren't getting credit. If this occurs, a moderator would be very helpful. As for the PPD fluctuations, its puzzling that it changes so much, but not all WUs are the same, and its difficult to make them all behave exactly the same on every hardware architecture. Perhaps there is something more going on here that's the root cause instead.
User avatar
Jesse_V
 
Posts: 2894
Joined: Mon Jul 18, 2011 4:44 am
Location: USA

Re: Where are my points going?

Postby djohnston » Sat Nov 26, 2011 8:02 pm

Jesse_V wrote:
QRB stands for Quick Return Bonus. Certain Work Units, such as those for SMP (and a few for uniprocessor but they're rare) tell the servers to reward you with bonus points for completing a WU quickly. For example, I process SMP WUs, and they have base points, which I get no matter so long as I return the WU before the deadline. However, if I return the WU really fast (like in the third of the time to the Preferred Deadline) then I'm non-linearly rewarded a lot of extra points on top of the base points. They have this system in place to enourage people to complete those hefty WUs quickly, and as a reward for much more of our computing resources. Since the QRB WUs give more points, they could have influenced your PPD, and that's why it was brought up.



Thanks for all the detailed explanations. I didn't know that finishing a WU early could earn bonus points.
djohnston
 
Posts: 85
Joined: Sat Feb 06, 2010 8:33 pm

Re: Where are my points going?

Postby djohnston » Sat Nov 26, 2011 8:14 pm

gwildperson wrote:
You should not write a script that queries your official stats every X minutes because that creates a continuous load whether you're actually looking at the data it returns or not.


Never said I did. I'll rephrase. As I understand it, we should not use a script to gather stats from folding@home servers, as that would put undue strain on their limited resources. The script is run on the server side on extremeoverclocking.com and generates assorted individual and team statistics. I just point to it in a signature I use here and on a Linux forum. You can see it at the bottom of each post I make.

gwildperson wrote:
I've never seen a recommendation that we avoid using the official site manually through a browser. Where did you get that information?


See above.

gwildperson wrote:
Avoid saying "should not use" for either.


Avoid saying? Seriously? What terminology guide should I use so that I don't raise the ire of the censors?
djohnston
 
Posts: 85
Joined: Sat Feb 06, 2010 8:33 pm

Re: Where are my points going?

Postby k1wi » Sat Nov 26, 2011 9:34 pm

If you're interested in monitoring the performance of your computers and their ppd, then I strongly suggest that you look at installing a third party monitoring client to your home network (such as HFM.net) and then link in all your computers to monitor it (you are forced to run v6 however, as it is incompatible with v7). That way you're able to monitor the performance of each machine independently as well as additional information such as machine benchmarks. It'll provide much better information than average sites. The issue with using 3rd party monitoring sites is that they're the aggregate of all your machines and they will fluctuate as you return work units.

For example: if one of your machines gets one of the larger work units that takes several days to many days to complete (such as P1072*), then your ppd will drop towards the end of the work unit (because that machine hasn't returned any points for the last couple of days) and then jump up as it returns several days of points in one return. Have three or four clients pick up work units of that length and the jumps will be even greater. Likewise, a large work unit that returned six days ago will still have an effect on the current reported average by the third party stats site.

On the other hand a monitoring app like HFM.net will give you a point-in-time ppd based on the actual performance of your machines and the speed at which they are completing work units, so regardless of the length of the work unit and the length of time between returns it will give you an accurate report of your current folding performance.

Regarding QRB (bonus work units) - the monitoring app such as HFM.net will help you distinguish between the older Core 74 work units that don't earn bonus points and the A4 work units that do earn bonus points. On my core2duo 2.33GHz machine (fairly low-midrange) I tend to get a 2x premium ppd on those work units - so if your PPD picks up you can refer to HFM.net and see that it's one of your clients folding a QRB work unit. I'm getting quite a few QRB work units these days, it'll just depend on your settings.

Regarding the site's stats - don't use a script to scrape it, feel free to browse it, but you'll get more information easier from the 3rd party sites.

k1wi
k1wi
 
Posts: 1322
Joined: Tue Sep 22, 2009 10:48 pm

Re: Where are my points going?

Postby djohnston » Sun Nov 27, 2011 8:18 am

k1wi wrote:
On the other hand a monitoring app like HFM.net will give you a point-in-time ppd based on the actual performance of your machines and the speed at which they are completing work units, so regardless of the length of the work unit and the length of time between returns it will give you an accurate report of your current folding performance.


Thanks. I downloaded it from code.google.com and will take a look. Do you use it on Linux? Documentation says it requires mono.
djohnston
 
Posts: 85
Joined: Sat Feb 06, 2010 8:33 pm

Re: Where are my points going?

Postby k1wi » Sun Nov 27, 2011 7:52 pm

I haven't run it from linux - the way I have it configured it is running on a Windows 7 64bit machine with the F@H folders of other machines shared and mounted as network drives and a shared drive on my linux webserver for the webpages generation.
k1wi
 
Posts: 1322
Joined: Tue Sep 22, 2009 10:48 pm

Re: Where are my points going?

Postby Amaruk » Tue Nov 29, 2011 8:37 am

djohnston wrote:Documentation says it requires mono.

I have it running using mono. Separate instance on each machine, so can't vouch for how it works over a network.
Image
User avatar
Amaruk
 
Posts: 512
Joined: Fri Jun 20, 2008 3:57 am
Location: Watching from the Woods

Re: Where are my points going?

Postby djohnston » Fri Dec 16, 2011 7:58 am

My apologies to Folding@home administrators. The statistics I was relying on are coming from extremeoverclocking.com. They have been experiencing server outages lately, as well as corruption in their Folding@home MySQL reporting database. My reported daily "average" coming from extremeoverclocking suddenly jumped from 1100 PPD to well over 3000PPD in the space of 36 hours. Clearly, something was wrong at their end, and probably still is.

I didn't think this situation through well enough before posting. I'm now marking this topic as SOLVED, and the event was in no way, shape or form the fault of the Folding@home staff.
djohnston
 
Posts: 85
Joined: Sat Feb 06, 2010 8:33 pm

Re: Where are my points going?

Postby Grandpa_01 » Fri Dec 16, 2011 9:04 am

djohnston wrote:
k1wi wrote:
On the other hand a monitoring app like HFM.net will give you a point-in-time ppd based on the actual performance of your machines and the speed at which they are completing work units, so regardless of the length of the work unit and the length of time between returns it will give you an accurate report of your current folding performance.


Thanks. I downloaded it from code.google.com and will take a look. Do you use it on Linux? Documentation says it requires mono.


I run it on Linux (Ubuntu 10.10) there is a good guide for installing HFM on Linux over at the [H] http://hardforum.com/showthread.php?p=1 ... 1037125479 and it does require mono to launch it. ( sudo mono HFM.exe)
Image
2 - SM H8QGi-F AMD 6xxx=112 cores @ 3.2 & 3.9Ghz
5 - SM X9QRI-f+ Intel 4650 = 320 cores @ 3.15Ghz
2 - I7 980X 4.4Ghz 2-GTX680
1 - 2700k 4.4Ghz GTX680
Total = 464 cores folding
User avatar
Grandpa_01
 
Posts: 1873
Joined: Wed Mar 04, 2009 7:36 am


Return to Linux CPU V6 Client

Who is online

Users browsing this forum: No registered users and 1 guest

cron