30 replies to this topic

[Nvyseal]

m.oreilly, on Oct 3 2007, 09:09 AM, said:

it did "it" again! just before completion of a 500 point work unit, the fah client stopped folding, sent the data prematurely (13 hours ahead of estimated, resulting in just 200 points, though it would seem to be an incomplete unit, wouldn't it?), restarting with another work unit. what gives here? i kinda don't want to be running for days/weeks and have the work chopped up like this. anyone else with issues like this? i am going to kill this latest work unit, as i was intending to redo the array, and install x64 vista and mint linux today...

I think your having the same Problems i did, and why i use VMware with LinuxSMP:

Quote

This actually is a pretty stable client IF you don’t use your PC for anything else but folding. After a week of trial and error, and a total of 5280+ points lost because of errors, I mastered the Windows SMP client. You must have a very stable PC, if it’s overclocked and not completely stable, or you have the slightest bug in your hardware drivers, or your RAM timings are just a slight bit off, your work unit will be rejected. If you use your PC for anything else such as gaming, ripping to web surfing, it could cause your client to become unstable and at the end of your workup contain errors in your work unit, which will be rejected. The same is true if your PC requires a reboot in the middle of a work unit or you terminate the work unit early and restart it, even if you have the checkpoints enabled.

Oh, and another thing, Some WU's just dont work, like im on a 2652, and it just keeps rejecting it

[brewin]

Keep up the good work, guys. We'll be breaking into the top 500 in a few days.

[m.oreilly]

brewin, on Oct 3 2007, 11:50 AM, said:

Keep up the good work, guys. We'll be breaking into the top 500 in a few days.

i'm back in the saddle w/ v6

[m.oreilly]

...and now i'm not. my first work unit using v6 was 1/3 done. i decided to install an AV (new installation, vista 32bit). after rebooting, fah had completely vanished from my system! nothing, except a text file in my temp...no exe, nothing. not even any mention in the latest ran apps in the start menu. i don't think i can afford weeks of constant system 24/7 use, and have nothing to show for it. i'm stopping folding as of today.

[KelpFries]

Im proud of myself, if only a little. Just finished my 100th WU!

[KelpFries]

Instead of using the new client, v6, just use the stable client. v6 is a beta client.

[Nvyseal]

v6 is junk, i installed it today, kept crashing, rejecting WU's (Kubuntu) used the -smp The old SMP Client expired October 1st. Set your date back start the SMP then set the date ahead (it gets the date and time from the server anyways)

[WFO]

Nvyseal, on Oct 4 2007, 04:43 PM, said:

v6 is junk, i installed it today, kept crashing, rejecting WU's (Kubuntu) used the -smp The old SMP Client expired October 1st. Set your date back start the SMP then set the date ahead (it gets the date and time from the server anyways)

Thank you NVY for confirming what i asked on the other site.

Just wondering if anyone else is having problems with the new client. Strange things have been happening to me. I'm getting hangs connecting after WUs finish. I have to stop and restart clients on occasion to get results to upload. I've had a couple segmentation faults. Tonight I got sooo lucky. A WU finished. It was at ending MPI. It was taking too long so I took a quick snapshot with VMware Workstation. I then got a core communication error and the WU was deleted. I quickly restored the snapshot and it uploaded. Woohoo 20 hours work saved. Stiil... for me the V6 client sucks the nubes compared to V5.

And a follow up...

[01:04:46] Timered checkpoint triggered.
[01:09:47] Timered checkpoint triggered.
[01:11:43] Writing local files
[01:11:43] Completed 445000 out of 500000 steps (89 percent)
[01:16:44] Timered checkpoint triggered.
[cli_1]: aborting job:
Fatal error in MPI_Sendrecv: Error message texts are not available
[0]0:Return code = 0, signaled with Segmentation fault
[0]1:Return code = 1
[0]2:Return code = 0, signaled with Segmentation fault
[0]3:Return code = 0, signaled with Segmentation fault
[01:18:37] CoreStatus = 1 (1)
[01:18:37] Client-core communications error: ERROR 0x1
[01:18:37] Deleting current work unit & continuing...
[0]0:Return code = 0, signaled with Quit
[0]1:Return code = 0, signaled with Quit
[0]2:Return code = 18
[0]3:Return code = 0, signaled with Quit
[01:23:00] - Warning: Could not delete all work unit files (8): Core returned invalid code

Oh well it's a beta and the points surpass the losses.
Should have some more here soon. My X2-4800 had problems. I came home to a crashed VMware Linux client. When I restarted the WU had suddenly expired @ about 85%. *&^%(%$&

[Nvyseal]

Yup, same problems i was having all day today, i just couldnt get it to run. Im back on v5. If i get another expiration notice, im going to quit folding until they have a stable client to give us. Seems their adding "Bloat" to it. and also why cant they put the -SMP switch setup in the -configonly?? Thats just stupid!

[brewin]

I haven't had any problems with v6, but I'm running Linux natively. Could it be a problem with vmware? Either way, you should report it to the Pande Group. They'll want to know about it. Their forum is down more often than not these days, but maybe if you refresh the page 20 times and then sacrifice a goat, you'll be able to post.

http://forum.folding-community.org/forum57...n-smp-mode.html

[banj0]

For some reason, setting affinity for 2 projects on 2 cores via the -config menu isn't working for me on Vista even though I used '1' and '2' for the machine ID's like I always have. I didn't notice until tonight, but both instances were being run on both cores at the same time. I wonder how much that slowed me down over the last couple days.

I changed it thru tskmgr so hopefully my production will pick up now.

Edited by banj0, 05 October 2007 - 05:01 AM.

[WFO]

brewin, on Oct 4 2007, 09:44 PM, said:

I haven't had any problems with v6, but I'm running Linux natively. Could it be a problem with vmware? Either way, you should report it to the Pande Group. They'll want to know about it. Their forum is down more often than not these days, but maybe if you refresh the page 20 times and then sacrifice a goat, you'll be able to post.

http://forum.folding-community.org/forum57...n-smp-mode.html

And thank you Brew for confirming their forum has turned to sheep dip. I actually thought I had been banned. I haven't been able to post for a month from any computer, browser or IP. Sacrificing a goat? Why didn't I think of that? Here goat. Here goat. bahahahbahaha.

[brewin]

WFO, on Oct 5 2007, 12:27 AM, said:

And thank you Brew for confirming their forum has turned to sheep dip. I actually thought I had been banned. I haven't been able to post for a month from any computer, browser or IP. Sacrificing a goat? Why didn't I think of that? Here goat. Here goat. bahahahbahaha.

Yeah, it's bad. I don't advocate killing small animals. But you do what you gotta do...

But seriously, they do need to get their act together. First they had some weird cookie problem a few months ago that blocked half the members from logging in for weeks and now this.

[KelpFries]

I continue to have problems running the Windows SMP client. The client will fold a WU in approx 16 hours, but 4 times now including one it was folding this morning the client has had an error. The error this morning was "long 1-4 interactions.....Gromacs cannot continue further" after about 54% completion. And bang, gone, approx 9 or 10 hours of folding a 1760 point WU and its gone for the 4th time in the past week. This happened over night with nothing else running on my system to cause the client to result in an error. This is really annoying and its starting to piss me off.