Folding Forum

I lose an undue number of GPU WUs "just because." My machine will run for days without a problem, then occasionally the GPU client will EUE multiple times due to an "unstable machine." When I restart the client and get a new WU, it continues on for several more days.

The biggest problems are when the machine is unattended, since I travel a lot. It will be a welcome change when the new GPU client better handles errors and stops blaming "unstable machine" for all EUEs...

Qinsp wrote:

7im wrote:

theteofscuba wrote:...

i've had to abandon some wus in the past every now and then when configuring and reconfiguring f@h clients.

Most times the client can simply be udpated, or copied and moved, and not touch the WU. Under what conditions does reconfiguring a fah client require you to abandon a WU?

If you are running XP, and attempt to change the configuration, you can lose the job. Win7 seems to be stable. XP is broken. Any change can (and often is) lethal because the CONFIG code itself is bad. It's not the job, it's that you cannot restart the client without removing everything first. The only way I've been able to restart is by wiping out the whole install and reinstalling.

I have never had any problems stopping a client and running configuration mid-WU with XP. Editing the config file itself is always dangerous, but properly running a console with -config or -configonly works perfectly- CPU, GPU, or SMP. I don't use the systray clients, but I don't recall hearing any issues with configuration.

There were times when I was running WinXP platforms and GPU2 clients (Win7 and GPU3 now) when making changes to the config file on the fly would destroy a work unit. I eventually learned to shut down the client, make the configuration change, and to then restart the client. I did not lose work units that way.

bruce wrote:Nobody disputes that an occasional WU will be lost, but let's be realistic. What percentage of the WUs that you've been assigned were lost because of a hardware swap where you forgot to run -oneunit?

An allowable loss rate of 20% of the SMP WUs is HUGE compared to the number of WUs that are actually lost -- unless someone is intentionally cherrypicking.

You won't get an argument from me on the 20%, I doubt my lost WUs amount to even 1% let alone 20. My comments WRT a -release switch were directed more towards doing better science by letting the system know as early as possible that it needs to reassign a WU.

While we're on the subject though.... any changes to the 20% grace for lost units should be done on a sliding scale. IMO, 20% is a reasonable figure to use when a folder is first starting out. If we were to use something like 2%, any small glitch could quickly put that folder into a hole that would be very difficult to dig out of. The last thing we want to do is something that will discourage new folders. (Even the current 8 complete units has some new folders gnashing their teeth while they wait a few days for the bonus points to show up... ). A scale based on total WUs submitted (i.e. 20% for <25 WUs, 10% for <50 WUs, 5% for <75 WUs, etc....) would have the effect of discouraging long time folders from cherry picking, while not putting new folders at an unfair disadvantage.

edit: It should also be noted that the recent state of WUs being made available has made cherry picking mostly moot..... doesn't do any good to attempt cherry picking if the only WUs available for days on end are 6701/6702s....

bruce wrote:Nobody disputes that an occasional WU will be lost, but let's be realistic. What percentage of the WUs that you've been assigned were lost because of a hardware swap where you forgot to run -oneunit?

An allowable loss rate of 20% of the SMP WUs is HUGE compared to the number of WUs that are actually lost -- unless someone is intentionally cherrypicking.

The allowable lost rate is not 20% if you get a true-blue known bad WU. It crashes 5 times, so the real lost rate is 4% if my math is right. Only "user problem" rate is 20%.

Qinsp wrote:

7im wrote:

theteofscuba wrote:...

i've had to abandon some wus in the past every now and then when configuring and reconfiguring f@h clients.

Most times the client can simply be udpated, or copied and moved, and not touch the WU. Under what conditions does reconfiguring a fah client require you to abandon a WU?

If you are running XP, and attempt to change the configuration, you can lose the job. Win7 seems to be stable. XP is broken. Any change can (and often is) lethal because the CONFIG code itself is bad. It's not the job, it's that you cannot restart the client without removing everything first. The only way I've been able to restart is by wiping out the whole install and reinstalling.

When the CONFIG crashes, it won't let you make anything but a blank config file, and that won't finish the job.

Yes, if there were detailed instructions on the work-arounds, you could save the job. But most the "fixes" posted on the forum do not work. It has nothing to do with renaming the executable, it crashes either way. It does not affect 100% of XP machines apparently, but on this one I'm typing from, it will crash 100% of the time like clockwork. Config == RIP WU.

Only once have I had to completely remove a client to get it running again, and I've been folding since 2003. Deleting a config file, or deleting the WU is about as drastic as I have to go. I can't remember the last time a v6 client crashed on me either. And while SMP used to lose WUs regularly 4 years ago, but not recently.

I don't know what you are doing that is so very dangerous to your computer that so many WUs or Configs get corrupted, but I can't remember the last time one of my config files died. Your client issues are the exception to the rule, not the common place occurance. Please take more care.

If the client keeps giving you unstable-machine errors, there's a distinct chance that you might actually have an unstable machine, perhaps a faulty GPU or RAM.

For people stopping in the middle of WUs to change configuration, there's a "Pause when done" option; you can change things between WUs instead.

If the machine is unstable, why does it Fold SMP+GPU for 2 weeks before it gets to its "unstable" phase?

Because it is not 100% stable

Hmmm... If it's stable except for a few F@H WUs, maybe it's the WUs that are unstable...

jrweiss wrote:Hmmm... If it's stable except for a few F@H WUs, maybe it's the WUs that are unstable...

Drop the OC back to default settings and run it for a while if it does not loose any WU's at default then you have your answer. My bet would be not stable.

7im wrote:...

I don't know what you are doing that is so very dangerous to your computer that so many WUs or Configs get corrupted, but I can't remember the last time one of my config files died. Your client issues are the exception to the rule, not the common place occurance. Please take more care.

Well, you were right. I should have listened. I feel so guilty.

I should have seen the signs, since I've been working on computers since the CP/M days, but I suppose I was in denial. It was good computer when it was young, but as it got older, odd things started to happen. Things started to disappear off my desk. I would often find letter openers, scissors, or even razor blades on my chair. I would look at the History in my web-browser and it had listings for Satanic websites. My Music Folder was filled with heavy metal rock and vulgar rap songs.

I blamed it on the employees. If only I had listened to you.

This morning there was a dead hooker in the alley behind our shop. Next to the body was a CMOS battery...

The computer wouldn't boot. I opened the case, and the CMOS battery was missing.

I don't know what to do. I raised that computer from a chip, and I would feel like I betrayed it if I send it to Warranty. I put the battery back in and it still works great. But it still won't let me config FAH without a fight.

It was good computer when it was young, but as it got older, odd things started to happen.

I don't know what to do. I raised that computer from a chip

Were you careful to have it imprint on its peers?
It may be confused of its identity, causing it to challenge you for leadership of the herd.

Classic mistake that needs to be avoided with wolves and dairy bulls.
http://www.grandin.com/behaviour/princi ... dents.html

You may have to cull it. Once they decide to go after you, it is the only safe choice.

Grandpa_01 wrote:Drop the OC back to default settings and run it for a while if it does not loose any WU's at default then you have your answer. My bet would be not stable.

I don't OC.

Have you tested the memory