linuxrouter
Omnipotent Enthusiast
- Total Posts : 8043
- Reward points : 0
- Joined: 2008/02/28 14:47:45
- Status: offline
- Ribbons : 102
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/10 14:24:39
(permalink)
farinhir
Ok, I was thinking it would be the lower level (non-bigadv) WUs that I would be able to do. The question is, with my current setup (E8400@ 3.73Ghz), would it be worth me changing over from notfred's client, or will I just be shooting myself in the foot?
Thanks in advance for the help.
Several people have reported improved PPD with this virtual image so it might be worth a try. You can leave your existing Notfred install in place. Then try out this client with a single work unit to see how it performs on your system. The web config has a oneunit option that you can enable to run a single work unit. The other option would be to grab the Windows SMP client and run the SMP2 work units directly via Windows.
post edited by linuxrouter - 2010/02/10 14:27:33
CaseLabs M-S8 - ASRock X99 Pro - Intel 5960x - XSPC WC - EVGA 980 Ti Hybrid - Samsung 950 512GB - EVGA 1600w TitaniumAffiliate Code: OZJ-0TQ-41NJ
|
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/10 15:46:31
(permalink)
farinhir
Ok, I was thinking it would be the lower level (non-bigadv) WUs that I would be able to do. The question is, with my current setup (E8400@ 3.73Ghz), would it be worth me changing over from notfred's client, or will I just be shooting myself in the foot?
Thanks in advance for the help.
If you use the 1.0 image, you'll see higher PPD than notfred's. 1.1 image uses the newer 6.29 client which is slower than the 6.24 for a2 units... Still, even the 1.1 image is faster than notfred's for a1 and a2 units in my experience... notfred's image cannot run a3 units, which tend to be a bit faster than a1 units but slower than a2 units... I believe your E8400 @ 3.73GHz would behave similar to half of my Q6600 @ 3.70GHz...
|
farinhir
Superclocked Member
- Total Posts : 212
- Reward points : 0
- Joined: 2009/06/22 03:17:19
- Status: offline
- Ribbons : 1
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/10 16:04:36
(permalink)
Thank you for the replies. So it seems that I should use the 1.0 image if I read that correctly. So would I be using "./fah6 -smp 2 -advmethods -forceasm -verbosity 9" as my command line parameters? I have currently cleared notfred's WUs and I am now trying to figure the setup out. I am not sure what options I should activate. I do know not to touch the bigadv switch, but other than that I do not know. In the webconfig I did use the "MPICH_NO_LOCAL," as I hear that is supposed to help dual cores. I have also been pointed to http://www.overclockers.c...howthread.php?t=632715 to get the SMP2 setup under windows instead of using the linux images. Would that be a better option?
post edited by farinhir - 2010/02/10 16:09:13
|
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/10 16:15:38
(permalink)
I think you should be fine with those settings, except that you may need to use "-smp 4" instead of "-smp 2"; it will still run on two cores but I think the 6.24 client is expecting at least 4 smp cores to be input... Edit: I think VMware images would net higher PPD than the Windows client... 1.1 image contains the 6.29 client which runs SMP2 (with advmethods) = a3 units...
post edited by theGryphon - 2010/02/10 16:17:37
|
farinhir
Superclocked Member
- Total Posts : 212
- Reward points : 0
- Joined: 2009/06/22 03:17:19
- Status: offline
- Ribbons : 1
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/10 16:47:10
(permalink)
OK. I have it setup on image 1.1 with the settings I posted above but with -smp 4. Everything seems to be running, I just do not know how well yet. I was getting about 17:40 per frame with smp 2 on a p6014. I just do not know if smp 4 will help or hurt yet. Thank you for the help. Edit: I have changed back to "-SMP 2" from "-smp 4" because it was taking over 25 minutes to complete a frame.
post edited by farinhir - 2010/02/10 17:03:29
|
leexgx
New Member
- Total Posts : 14
- Reward points : 0
- Joined: 2010/02/11 04:01:15
- Status: offline
- Ribbons : 0
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/11 04:04:22
(permalink)
you need to fix the fourms as it not very friendly with Opera i have to click on this line before i can type norm you can click any where in the box and on top of that it let me post an blank post is there an way to invoke the backup script every 1 hr, the other backup script on the other sticky is for backing up to an windows PC but i just want to one thats in this image to run auto every 1 hr, as i just lost 2.5 days of work doing an bigadv Project due to no auto backup i am no linux guru at all best guess would be something like this work (------- No Quote button ?? ---) #!/bin/sh if [ $# -ne 1 ] then echo "need to supply an hour value" else HOUR_INPUT=$1 if [ $HOUR_INPUT -lt 0 -o $HOUR_INPUT -gt 23 ] then echo "need to supply an hour value between 0 and 23" else fahbackup.sh echo "Backup at $HOUR_INPUT hour completed at $DEST" fi fi (---------------) that above i just thought up bit of an modification to the windows backup script, all this does is call the backup script every 1hr (that you can set) or it should do, i not really an programmer at all just thought the above mite work, if some one can post an script or cronjob that run every hr be nice if you could add option in the Web page to make 1hr backups (add in VM 1.2 ) as i just lost an 2.5 days of work doing an adv work unit, i have disable the Langouste now as well as that may have prevented the work unit been sent this morning
post edited by leexgx - 2010/02/11 04:54:03
|
farinhir
Superclocked Member
- Total Posts : 212
- Reward points : 0
- Joined: 2009/06/22 03:17:19
- Status: offline
- Ribbons : 1
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/11 04:14:09
(permalink)
Hmm. So I am wondering what would happen if I left the -advmethods switch off? I am trying to figure out how to make sure I am either doing the a1 or a2 cores. It seems that the a3 cores give a significant drop in PPD for me. at least for now. Maybe when I can get bonus points it will be better. Yes I use HFM.NET.
|
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/11 07:45:56
(permalink)
farinhir
OK. I have it setup on image 1.1 with the settings I posted above but with -smp 4. Everything seems to be running, I just do not know how well yet. I was getting about 17:40 per frame with smp 2 on a p6014. I just do not know if smp 4 will help or hurt yet.
Thank you for the help.
Edit:
I have changed back to "-SMP 2" from "-smp 4" because it was taking over 25 minutes to complete a frame.
With the 1.1 image you really should use "-smp 2"; 6.29 client likes that you specify the exact number of cores. 6.24 (image 1.0 or lower) doesn't mind...
|
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/11 07:47:30
(permalink)
farinhir
Hmm. So I am wondering what would happen if I left the -advmethods switch off? I am trying to figure out how to make sure I am either doing the a1 or a2 cores. It seems that the a3 cores give a significant drop in PPD for me. at least for now. Maybe when I can get bonus points it will be better. Yes I use HFM.NET.
In HFM settings, make sure you check the "Calculate bonus credit and PPD" option, it'll take care of it...
|
DRSpalding
New Member
- Total Posts : 15
- Reward points : 0
- Joined: 2009/12/01 13:41:18
- Status: offline
- Ribbons : 0
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/11 10:47:11
(permalink)
Just an FYI. I upgraded from the v0.4 and v0.3 images on two machines to v1.1 on both. My observations, all running -bigadv of course: The PPD is slightly lower than before. - Lost about 100 PPD (raw) on the v0.3 machine running a stock i7-920. 7900 -> 7800 PPD.
- Lost about 200 PPD (raw) on the v0.4 machine running a o/c i7-920 d0 at 3.9GHz. 10,900 -> 10,700.
- On the o/c machine, it is also running an nVidia (PNY) GTX275 and the GPU client. It is also somewhat more affected by the v1.1 VM image than it was in v0.4. I am a little less sure of this sensitivity because of the influx of new projects for the GPU lately that really have scrambled my observations.
The sensitivity to outside processes is higher. The machine (the o/c one) I use to browse and email from would typically drop from 10,900 to about 10,500 or so PPD (raw). It is now dropping to about 10,100. OTOH, it seems to be absolutely rock solid in breaking into a running WU and having the checkpoint files be valid now. I guess that this alone makes it likely worthwhile for the small PPD hit. As an aside, I had noticed before that when Win7 shut down due to an update that required restarting, VMWare would restore the session 100% of the time, right where it left off, w/o issues. It does not accomplish exactly the same thing, except that if you have to shut down the machine for non-Folding issues/updates/etc., then it seems to work just fine to stop the VM that way instead of letting the FAH client try to do it cleanly.
post edited by DRSpalding - 2010/02/11 11:38:18
|
linuxrouter
Omnipotent Enthusiast
- Total Posts : 8043
- Reward points : 0
- Joined: 2008/02/28 14:47:45
- Status: offline
- Ribbons : 102
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/11 16:42:14
(permalink)
leexgx
is there an way to invoke the backup script every 1 hr, the other backup script on the other sticky is for backing up to an windows PC but i just want to one thats in this image to run auto every 1 hr, as i just lost 2.5 days of work doing an bigadv Project due to no auto backup
i am no linux guru at all best guess would be something like this work
(------- No Quote button ?? ---)
#!/bin/sh
if [ $# -ne 1 ]
then
echo "need to supply an hour value"
else
HOUR_INPUT=$1
if [ $HOUR_INPUT -lt 0 -o $HOUR_INPUT -gt 23 ]
then
echo "need to supply an hour value between 0 and 23"
else
fahbackup.sh
echo "Backup at $HOUR_INPUT hour completed at $DEST"
fi
fi
(---------------)
that above i just thought up bit of an modification to the windows backup script, all this does is call the backup script every 1hr (that you can set) or it should do, i not really an programmer at all just thought the above mite work, if some one can post an script or cronjob that run every hr
The backup script should backup the data every hour to /mnt/winshare provided the cronjob was setup. The cron schedule is included with the scripts. Here is a thread that covers the backup script. There is also a new version of the backup script. http://forums.evga.com/tm.aspx?m=130442 Depending on how the VM was setup, there may also be a backup in /var/fah in a directory called backup.0 through 4. leexgx
be nice if you could add option in the Web page to make 1hr backups (add in VM 1.2 ) as i just lost an 2.5 days of work doing an adv work unit, i have disable the Langouste now as well as that may have prevented the work unit been sent this morning
I will look into setting something up for that. Langouste does actually prevent the main client from sending the work unit. What it does is make a copy of the finished work unit to /tmp/langouste and it sends the work unit from there with a second fah6 process. This is done so that the main client can begin downloading a new work unit. Eventually you will see a message in the folding log that a work unit was removed from the queue and this indicates that the work unit sent out successfully via the second process.
CaseLabs M-S8 - ASRock X99 Pro - Intel 5960x - XSPC WC - EVGA 980 Ti Hybrid - Samsung 950 512GB - EVGA 1600w TitaniumAffiliate Code: OZJ-0TQ-41NJ
|
linuxrouter
Omnipotent Enthusiast
- Total Posts : 8043
- Reward points : 0
- Joined: 2008/02/28 14:47:45
- Status: offline
- Ribbons : 102
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/11 17:06:54
(permalink)
DRSpalding
Just an FYI. I upgraded from the v0.4 and v0.3 images on two machines to v1.1 on both. My observations, all running -bigadv of course:
The PPD is slightly lower than before.
- Lost about 100 PPD (raw) on the v0.3 machine running a stock i7-920. 7900 -> 7800 PPD.
- Lost about 200 PPD (raw) on the v0.4 machine running a o/c i7-920 d0 at 3.9GHz. 10,900 -> 10,700.
- On the o/c machine, it is also running an nVidia (PNY) GTX275 and the GPU client. It is also somewhat more affected by the v1.1 VM image than it was in v0.4. I am a little less sure of this sensitivity because of the influx of new projects for the GPU lately that really have scrambled my observations.
The sensitivity to outside processes is higher. The machine (the o/c one) I use to browse and email from would typically drop from 10,900 to about 10,500 or so PPD (raw). It is now dropping to about 10,100.
OTOH, it seems to be absolutely rock solid in breaking into a running WU and having the checkpoint files be valid now. I guess that this alone makes it likely worthwhile for the small PPD hit.
As an aside, I had noticed before that when Win7 shut down due to an update that required restarting, VMWare would restore the session 100% of the time, right where it left off, w/o issues. It does not accomplish exactly the same thing, except that if you have to shut down the machine for non-Folding issues/updates/etc., then it seems to work just fine to stop the VM that way instead of letting the FAH client try to do it cleanly.
Thanks for sharing your observations. I have noticed a performance difference from one -bigadv project to another. There are three projects, 2681-2683 that each perform differently. When you were comparing the performance, was that with the same project number? The main component that affects the folding performance is the Linux kernel. I am working on several new kernel builds for different architectures with more aggressive optimization and I am hoping these builds will help improve folding performance.
post edited by linuxrouter - 2010/02/11 17:08:08
CaseLabs M-S8 - ASRock X99 Pro - Intel 5960x - XSPC WC - EVGA 980 Ti Hybrid - Samsung 950 512GB - EVGA 1600w TitaniumAffiliate Code: OZJ-0TQ-41NJ
|
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/11 19:13:29
(permalink)
Hey LR, I have a question: Does the ramdisk option work exactly the same as TheCrazyCanucks's scripts? Where does it create the ramdisk backup?
I wanna use the Auto-start option with ramdisk but couldn't figure out the best way. I'm using TCC's hourly backup cron jobs...
|
farinhir
Superclocked Member
- Total Posts : 212
- Reward points : 0
- Joined: 2009/06/22 03:17:19
- Status: offline
- Ribbons : 1
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/12 04:03:03
(permalink)
theGryphon
farinhir
Hmm. So I am wondering what would happen if I left the -advmethods switch off? I am trying to figure out how to make sure I am either doing the a1 or a2 cores. It seems that the a3 cores give a significant drop in PPD for me. at least for now. Maybe when I can get bonus points it will be better. Yes I use HFM.NET.
In HFM settings, make sure you check the "Calculate bonus credit and PPD" option, it'll take care of it...
I had the bonus calculation on, that is why I mentioned HFM. It was saying I would make about 1300 PPD if I were awarded the bonus points on a project 6014. I have changed over to the "image 1.0" and am running the "./fah6 -smp 4 -advmethods -forceasm." This got my a standard A1 1920 WU. the speed I see here is comperable to my speed in notfreds (maybe slightly faster). I think I will stick to this setup and see what WUs I get sent. Besides, I need to complete 10 appropriate WUs before I can even get the bonus points. I just got my passkey the other day.
|
slowb00st
ACX Member
- Total Posts : 273
- Reward points : 0
- Joined: 2009/10/24 15:39:18
- Location: Puerto Rico
- Status: offline
- Ribbons : 1
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/12 07:05:45
(permalink)
So I updated to 1.1 the other day from 0.4 and started a bigadv and everything went like it should except when it was time to send it. It could not connect to the server got a bunch of these for about an hour: [06:15:25] + Attempting to send results [February 12 06:15:25 UTC] [06:15:25] - Couldn't send HTTP request to server [06:15:25] + Could not connect to Work Server (results) [06:15:25] (171.67.108.22:8080) [06:15:25] + Retrying using alternative port [06:15:25] - Couldn't send HTTP request to server [06:15:25] + Could not connect to Work Server (results) [06:15:25] (171.67.108.22:80) [06:15:25] - Error: Could not transmit unit 01 (completed February 12) to work server. [06:15:25] Keeping unit 01 in queue. [06:15:25] Project: 2681 (Run 12, Clone 18, Gen 51) and then it just suddenly got this one and I lost the WU  [07:00:20] + Attempting to send results [February 12 07:00:20 UTC] [07:00:20] - Couldn't send HTTP request to server [07:00:20] + Could not connect to Work Server (results) [07:00:20] (171.67.108.25:8080) [07:00:20] + Retrying using alternative port [07:00:20] - Couldn't send HTTP request to server [07:00:20] + Could not connect to Work Server (results) [07:00:20] (171.67.108.25:80) [07:00:20] Could not transmit unit 01 to Collection server; keeping in queue. [07:00:50] Project: 2681 (Run 12, Clone 18, Gen 51) [07:00:50] - Error: Could not get length of results file work/wuresults_01.dat [07:00:50] - Error: Could not read unit 01 file. Removing from queue. [07:01:20] + -oneunit flag given and have now finished a unit. Exiting.- Preparing to get new work unit... [07:01:20] Cleaning up work directory EDIT: The good and funny news is that checking in extremeoverclocking.com stats I did get the bigadv in  !
post edited by slowb00st - 2010/02/12 07:40:53
i7 860 @ 4.3 GHz || Gigabyte GA-P55A-UD7 || GSkill Ripjaw 8GB 2000MHz & Fan || Swiftech XT CPU Block || MCP355 w/ XSPC Res Top || MCR320-QP || 2 x BlackIce GTX 120 || WD 750GB Caviar Black RAID 0 || Corsair HX1000w || CoolerMaster ATCS 840 || OCZ Agility 60GB SSD RAID 0 || Lamptron FC-5 || 2 x Asus EAH5850 DirectCU TOP w/ EK Nickel/Acetal Water Blocks || EVGA 8800GTX physx
|
hotwired
ACX Member
- Total Posts : 342
- Reward points : 0
- Joined: 2010/01/30 08:48:49
- Status: offline
- Ribbons : 0
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/12 08:42:07
(permalink)
920 @ 20x200 @1.3
fpo/batch#3938B024
kingston @ 2000
cosair hx 1000 ps
gtx480 875/2001
|
bogmali
New Member
- Total Posts : 93
- Reward points : 0
- Joined: 2008/08/19 14:30:59
- Status: offline
- Ribbons : 0
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/12 08:56:29
(permalink)
Hey LR-I had an experience (on two different rigs) where I had to CTRL C and then power down the rig to install some needed updates. When I try run the image and have it pick up where it left off, it gave me "could not connect to server" and "attempt # XX failed and no other work to do". When I restarted the client it started a new WU and I ended up losing the previous WU's which were at 88% and 92% respectively. I am constantly checking EOC to see if I received credits for those 2 WUs but after 3 days, I gave up hope. Anybody else had to deal with the same thing? How did you fix it? Thanks again.
|
leftygof
FTW Member
- Total Posts : 1755
- Reward points : 0
- Joined: 2007/11/08 17:16:59
- Location: Amarillo, TX
- Status: offline
- Ribbons : 5
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/12 09:58:40
(permalink)
suspend the image....don't control-C
|
linuxrouter
Omnipotent Enthusiast
- Total Posts : 8043
- Reward points : 0
- Joined: 2008/02/28 14:47:45
- Status: offline
- Ribbons : 102
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/12 11:13:27
(permalink)
theGryphon
Hey LR, I have a question: Does the ramdisk option work exactly the same as TheCrazyCanucks's scripts? Where does it create the ramdisk backup?
I wanna use the Auto-start option with ramdisk but couldn't figure out the best way. I'm using TCC's hourly backup cron jobs...
Yes, the web interface calls the same script to create the ramdisk. The system will make a backup on shutdown or reboot and then restore that data the next time you boot up. The scripts from TheCrazyCanuck will do an hour-by-hour backup and stores those backups in /mnt/winshare which should be mounted to a Windows shared folder.
CaseLabs M-S8 - ASRock X99 Pro - Intel 5960x - XSPC WC - EVGA 980 Ti Hybrid - Samsung 950 512GB - EVGA 1600w TitaniumAffiliate Code: OZJ-0TQ-41NJ
|
linuxrouter
Omnipotent Enthusiast
- Total Posts : 8043
- Reward points : 0
- Joined: 2008/02/28 14:47:45
- Status: offline
- Ribbons : 102
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/12 11:16:26
(permalink)
farinhir
I had the bonus calculation on, that is why I mentioned HFM. It was saying I would make about 1300 PPD if I were awarded the bonus points on a project 6014. I have changed over to the "image 1.0" and am running the "./fah6 -smp 4 -advmethods -forceasm." This got my a standard A1 1920 WU. the speed I see here is comperable to my speed in notfreds (maybe slightly faster). I think I will stick to this setup and see what WUs I get sent. Besides, I need to complete 10 appropriate WUs before I can even get the bonus points. I just got my passkey the other day.
Only the core a2 or a3 work units apply to the 10-work unit requirement. If you grab the latest 6.29 client, you can run the SMP2 work units which do qualify while using a passkey.
CaseLabs M-S8 - ASRock X99 Pro - Intel 5960x - XSPC WC - EVGA 980 Ti Hybrid - Samsung 950 512GB - EVGA 1600w TitaniumAffiliate Code: OZJ-0TQ-41NJ
|
linuxrouter
Omnipotent Enthusiast
- Total Posts : 8043
- Reward points : 0
- Joined: 2008/02/28 14:47:45
- Status: offline
- Ribbons : 102
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/12 11:17:59
(permalink)
slowb00st
So I updated to 1.1 the other day from 0.4 and started a bigadv and everything went like it should except when it was time to send it. It could not connect to the server got a bunch of these for about an hour:
[06:15:25] + Attempting to send results [February 12 06:15:25 UTC]
[06:15:25] - Couldn't send HTTP request to server
[06:15:25] + Could not connect to Work Server (results)
[06:15:25] (171.67.108.22:8080)
[06:15:25] + Retrying using alternative port
[06:15:25] - Couldn't send HTTP request to server
[06:15:25] + Could not connect to Work Server (results)
[06:15:25] (171.67.108.22:80)
[06:15:25] - Error: Could not transmit unit 01 (completed February 12) to work server.
[06:15:25] Keeping unit 01 in queue.
[06:15:25] Project: 2681 (Run 12, Clone 18, Gen 51)
EDIT:
The good and funny news is that checking in extremeoverclocking.com stats I did get the bigadv in !
Do you have langouste enabled? That could explain these messages. Also I read on folding forum that there was a server outage in one of the threads.
CaseLabs M-S8 - ASRock X99 Pro - Intel 5960x - XSPC WC - EVGA 980 Ti Hybrid - Samsung 950 512GB - EVGA 1600w TitaniumAffiliate Code: OZJ-0TQ-41NJ
|
linuxrouter
Omnipotent Enthusiast
- Total Posts : 8043
- Reward points : 0
- Joined: 2008/02/28 14:47:45
- Status: offline
- Ribbons : 102
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/12 11:20:10
(permalink)
bogmali
Hey LR-I had an experience (on two different rigs) where I had to CTRL C and then power down the rig to install some needed updates. When I try run the image and have it pick up where it left off, it gave me "could not connect to server" and "attempt # XX failed and no other work to do". When I restarted the client it started a new WU and I ended up losing the previous WU's which were at 88% and 92% respectively. I am constantly checking EOC to see if I received credits for those 2 WUs but after 3 days, I gave up hope. Anybody else had to deal with the same thing? How did you fix it? Thanks again.
Do you have the latest 1.1 image? It has a different file system that should be more resistant to failure with folding work units. Also, I agree with Lefty regarding suspending the image. This method may be safer for shutting down the system.
CaseLabs M-S8 - ASRock X99 Pro - Intel 5960x - XSPC WC - EVGA 980 Ti Hybrid - Samsung 950 512GB - EVGA 1600w TitaniumAffiliate Code: OZJ-0TQ-41NJ
|
bogmali
New Member
- Total Posts : 93
- Reward points : 0
- Joined: 2008/08/19 14:30:59
- Status: offline
- Ribbons : 0
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/12 12:57:56
(permalink)
Yup, I'm using ver. 1.1 and it just did it again with my other rig after a BSOD (some IRQ error). Hopefully I get credit for the WU that was at 87%.
|
slowb00st
ACX Member
- Total Posts : 273
- Reward points : 0
- Joined: 2009/10/24 15:39:18
- Location: Puerto Rico
- Status: offline
- Ribbons : 1
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/12 13:54:02
(permalink)
linuxrouter
slowb00st
So I updated to 1.1 the other day from 0.4 and started a bigadv and everything went like it should except when it was time to send it. It could not connect to the server got a bunch of these for about an hour:
[06:15:25] + Attempting to send results [February 12 06:15:25 UTC]
[06:15:25] - Couldn't send HTTP request to server
[06:15:25] + Could not connect to Work Server (results)
[06:15:25] (171.67.108.22:8080)
[06:15:25] + Retrying using alternative port
[06:15:25] - Couldn't send HTTP request to server
[06:15:25] + Could not connect to Work Server (results)
[06:15:25] (171.67.108.22:80)
[06:15:25] - Error: Could not transmit unit 01 (completed February 12) to work server.
[06:15:25] Keeping unit 01 in queue.
[06:15:25] Project: 2681 (Run 12, Clone 18, Gen 51)
EDIT:
The good and funny news is that checking in extremeoverclocking.com stats I did get the bigadv in !
Do you have langouste enabled? That could explain these messages. Also I read on folding forum that there was a server outage in one of the threads.
Have to check on that, but it is most likely it. Thanks again...
i7 860 @ 4.3 GHz || Gigabyte GA-P55A-UD7 || GSkill Ripjaw 8GB 2000MHz & Fan || Swiftech XT CPU Block || MCP355 w/ XSPC Res Top || MCR320-QP || 2 x BlackIce GTX 120 || WD 750GB Caviar Black RAID 0 || Corsair HX1000w || CoolerMaster ATCS 840 || OCZ Agility 60GB SSD RAID 0 || Lamptron FC-5 || 2 x Asus EAH5850 DirectCU TOP w/ EK Nickel/Acetal Water Blocks || EVGA 8800GTX physx
|
bogmali
New Member
- Total Posts : 93
- Reward points : 0
- Joined: 2008/08/19 14:30:59
- Status: offline
- Ribbons : 0
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/12 16:44:08
(permalink)
Will the back-up script work after a BSOD or a power failure? I've lost three WU's now because the client does not pick up where it left off but pull a new WU instead and start from scratch.
|
linuxrouter
Omnipotent Enthusiast
- Total Posts : 8043
- Reward points : 0
- Joined: 2008/02/28 14:47:45
- Status: offline
- Ribbons : 102
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/12 16:49:49
(permalink)
bogmali
Will the back-up script work after a BSOD or a power failure? I've lost three WU's now because the client does not pick up where it left off but pull a new WU instead and start from scratch.
Yes because it makes a hour-by-hour backup. However, if you are seeing BSOD and other stability issues, you may want to run some more system stress testing. System crashes can potentially cause other issues such as issues with the Linux file system. It would be better to get the system running stable first before running bigadv.
CaseLabs M-S8 - ASRock X99 Pro - Intel 5960x - XSPC WC - EVGA 980 Ti Hybrid - Samsung 950 512GB - EVGA 1600w TitaniumAffiliate Code: OZJ-0TQ-41NJ
|
linuxrouter
Omnipotent Enthusiast
- Total Posts : 8043
- Reward points : 0
- Joined: 2008/02/28 14:47:45
- Status: offline
- Ribbons : 102
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/12 16:58:13
(permalink)
For those using Langouste for the first time, I thought it might be handy to see what the process is for sending out the work unit. Once the work unit is finished, you should see something like this in the folding log:
[20:18:35] + Attempting to send results [February 12 20:18:35 UTC]
[20:18:35] - Couldn't send HTTP request to server
[20:18:35] + Could not connect to Work Server (results)
[20:18:35] (171.67.108.22:8080)
[20:18:35] + Retrying using alternative port
[20:18:35] - Couldn't send HTTP request to server
[20:18:35] + Could not connect to Work Server (results)
[20:18:35] (171.67.108.22:80)
[20:18:35] - Error: Could not transmit unit 06 (completed February 12) to work server.
[20:18:35] Keeping unit 06 in queue.
[20:18:35] Project: 2682 (Run 6, Clone 22, Gen 5) [20:18:35] + Attempting to send results [February 12 20:18:35 UTC]
[20:18:35] - Couldn't send HTTP request to server
[20:18:35] + Could not connect to Work Server (results)
[20:18:35] (171.67.108.22:8080)
[20:18:35] + Retrying using alternative port
[20:18:35] - Couldn't send HTTP request to server
[20:18:35] + Could not connect to Work Server (results)
[20:18:35] (171.67.108.22:80)
[20:18:35] - Error: Could not transmit unit 06 (completed February 12) to work server. [20:18:35] + Attempting to send results [February 12 20:18:35 UTC]
[20:18:35] - Couldn't send HTTP request to server
[20:18:35] + Could not connect to Work Server (results)
[20:18:35] (171.67.108.25:8080)
[20:18:35] + Retrying using alternative port
[20:18:35] - Couldn't send HTTP request to server
[20:18:35] + Could not connect to Work Server (results)
[20:18:35] (171.67.108.25:80)
[20:18:35] Could not transmit unit 06 to Collection server; keeping in queue.
[20:18:35] - Preparing to get new work unit...
[20:18:35] Cleaning up work directory
[20:18:35] + Attempting to get work packet
[20:18:35] Passkey found
[20:18:35] - Connecting to assignment server
[20:18:36] - Successful: assigned to (171.67.108.22).
[20:18:36] + News From Folding@Home: Welcome to Folding@Home
[20:18:36] Loaded queue successfully.
[20:35:52] Project: 2682 (Run 6, Clone 22, Gen 5)
Langouste prevents the main client from sending out the work unit so that the client will begin to download a new one. Where it says "Loaded queue succesfully" is where the new download begins. At the same time, Langouste copies the entire folding directory to /tmp/langouste using a helper script. This helper script will fire off a second copy of fah6 in the /tmp/langouste directory. Here is what that log should show (/tmp/langouste/langouste-helper-3021.log):
/dev/shm/langouste/3021/clientdir
processing work/wuresults_06.dat
unit number '06'
launching fah... Note: Please read the license agreement (fah6 -license). Further
use of this software requires that you have read and accepted this agreement. 8 cores detected --- Opening Log file [February 12 20:19:36 UTC] # Linux SMP Console Edition ###################################################
############################################################################### Folding@Home Client Version 6.29 http://folding.stanford.edu/ ###############################################################################
############################################################################### Launch directory: /dev/shm/langouste/3021/clientdir
Executable: ./fah6
Arguments: -send 06 -bigadv -smp 8 [20:19:36] - Ask before connecting: No
[20:19:36] - Proxy: localhost:8080
[20:19:36] - Machine ID: 1
[20:19:36] A potential conflict was detected: Process 1933 is currently running and may also be a client with Mach. ID 1.
The program will now exit. Upon restart, this check will not be done --
You may wish to check that no client is currently running in
/usr/local/fah before restarting. Please press any key to exit.
work/wuresults_06.dat
re-launching fah... Note: Please read the license agreement (fah6 -license). Further
use of this software requires that you have read and accepted this agreement. 8 cores detected --- Opening Log file [February 12 20:19:36 UTC] # Linux SMP Console Edition ###################################################
############################################################################### Folding@Home Client Version 6.29 http://folding.stanford.edu/ ###############################################################################
############################################################################### Launch directory: /dev/shm/langouste/3021/clientdir
Executable: ./fah6
Arguments: -send 06 -bigadv -smp 8 [20:19:36] - Ask before connecting: No
[20:19:36] - Proxy: localhost:8080
[20:19:36] - Machine ID: 1
[20:19:36]
[20:19:36] Loaded queue successfully.
[20:19:36] Attempting to return result(s) to server...
[20:19:36] Project: 2682 (Run 6, Clone 22, Gen 5) [20:19:36] + Attempting to send results [February 12 20:19:36 UTC]
[20:43:20] + Results successfully sent
[20:43:20] Thank you for your contribution to Folding@Home.
[20:43:20] + Number of Units Completed: 45 Folding@Home Client Shutdown.
ls: cannot access work/wuresults_06.dat: No such file or directory
unit 06 sent!
all done
The last several lines show that the work unit sent out. Also in the above log, I have my langouste helper script set to use /dev/shm instead of /tmp/langouste since I prefer to have the work unit copied to RAM instead of disk, but the default is /tmp/langouste. Finally in the main folding log, you will eventually see this:
[20:36:07] Entering M.D.
[21:04:26] pleted 2500 out of 250000 steps (1%)
[21:32:26] Completed 5000 out of 250000 steps (2%)
[22:00:30] Completed 7500 out of 250000 steps (3%)
[22:28:32] Completed 10000 out of 250000 steps (4%)
[22:56:35] Completed 12500 out of 250000 steps (5%)
[23:24:38] Completed 15000 out of 250000 steps (6%)
[23:34:40] Project: 2682 (Run 6, Clone 22, Gen 5)
[23:34:40] - Error: Could not get length of results file work/wuresults_06.dat
[23:34:40] - Error: Could not read unit 06 file. Removing from queue.
[23:52:41] Completed 17500 out of 250000 steps (7%)
[00:20:45] Completed 20000 out of 250000 steps (8%)
The two lines with the Error shows that the work unit was removed from the queue since the work unit sent via another directory and another instance of the fah6 client.
post edited by linuxrouter - 2010/02/12 17:01:05
CaseLabs M-S8 - ASRock X99 Pro - Intel 5960x - XSPC WC - EVGA 980 Ti Hybrid - Samsung 950 512GB - EVGA 1600w TitaniumAffiliate Code: OZJ-0TQ-41NJ
|
kougar
FTW Member
- Total Posts : 1003
- Reward points : 0
- Joined: 2006/05/08 10:11:19
- Location: Texas
- Status: offline
- Ribbons : 3
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/13 00:38:14
(permalink)
leftygof
suspend the image....don't control-C
Yes. If you need to shut down the VM suspend the thing from the VMware control panel, so far I've not encountered any problems doing this with my Bigadv machine, despite the huge amount of RAM and load that needs to be halted and saved to disk. I was worried about syncing issues but I have yet to see any.
|
bogmali
New Member
- Total Posts : 93
- Reward points : 0
- Joined: 2008/08/19 14:30:59
- Status: offline
- Ribbons : 0
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/13 01:48:22
(permalink)
OK got multiple replies already about "suspending" the client instead of CTRL C, that could be what I am doing wrong. So, how do you actually suspend it? I remember doing it before by accident but I have not a clue how. Thanks for help fellas.
|
linuxrouter
Omnipotent Enthusiast
- Total Posts : 8043
- Reward points : 0
- Joined: 2008/02/28 14:47:45
- Status: offline
- Ribbons : 102
Re:VMWare 3.0 && bigadv folding (Image available for testing)
2010/02/13 23:52:41
(permalink)
bogmali
OK got multiple replies already about "suspending" the client instead of CTRL C, that could be what I am doing wrong. So, how do you actually suspend it? I remember doing it before by accident but I have not a clue how. Thanks for help fellas.
VM -> Power -> Suspend should do the trick.
CaseLabs M-S8 - ASRock X99 Pro - Intel 5960x - XSPC WC - EVGA 980 Ti Hybrid - Samsung 950 512GB - EVGA 1600w TitaniumAffiliate Code: OZJ-0TQ-41NJ
|
