Project 5801 issues. [Should be Offline]

[anko1]

Still pulling 5801s. Rinsing like crazy.

[lukeszabo01]

Hi there, this has been a problem for quite a while now and I don't understand why the 5801 project has not been pulled, as I have two of my three cards still wanting to try and run that project. I have deleted everything but the cudart.dll and client config file and it still downloads and runs the 5801 project on two of my cards, one of them has gotten the 5800 project and is running.

The part I don't understand is that you guys are now aware there is something very wrong with the 5801 work unit, but you keep sending it to us. Please pull if from the servers so we can continue to use working projects and help you in your research. Thanks.

[lukeszabo01]

I have a solution for myself, I just need to know if this would work, will it work if I place the 5800 project files into my other cards, or will I just process the same data 3 times and get 1 lot of credit for it?

[Insidious]

You'll only get credit for the first one that completes.

[lukeszabo01]

Bugger, was hoping that would work so I could make my cards keep doing something. I can't believe it is still on the servers being handed out, surely common sense says pull it from the servers until it can be sorted.

[Tigerbiten]

I think its mostly pulled as I've got none now.
I think whats left are work-units that are being re-issued and have slipped back through.
Keep going rinse and repeat.

Luck .............

[shatteredsilicon]

If this doesn't get pulled/sorted very soon, I'm going to be tempted to post out a bash wrapper (useful to Linux or cygwin users) that detects when this project is downloaded and preemptively kills the client, dumps the offending unit and restarts the client. The downside of this is that it will enable anyone to make a trivial modification to said wrapper to systematically auto-dump (as opposed to doing it manually) any WU they don't like (high vs low PPD WUs, anyone?) on any client they run. That would be bad, right?

[Insidious]

They're not pulled. 3 clients "paused" in the last 20 minutes.

I'm just going to let them sleep since that's apparently what Pande Group wants.

-Sid

[(_KoDAk_)]

core 1.18 not help ((((((( still UM ((((

[lukeszabo01]

I have been "rinsing and repeating" for about 30-40mins, I think there comes a time when you just throw away clothes that require that much cleaning, lol.

[zxy]

after worldwide massive server problem,STU kept pulling no-one-can-finish 5801 project, has it test the 5801 WU before pulling to us?i think they became more and more irresponsible.

[Saleen219]

I'm still getting them on the home PC which is my best folder (quad core and also folding on the gpu)

[theo343]

Please remove this WU ASAP, im getting them on all my 7 GPUs! It goes to UM on all computers after about 20 seconds.

Code: Select all

[23:11:41] Project: 5801 (Run 9, Clone 173, Gen 0)
[23:11:41]
[23:11:41] Assembly optimizations on if available.
[23:11:41] Entering M.D.
[23:11:47] Working on p5801_supervillin_e1
[23:11:47] Client config found, loading data.
[23:11:48] Starting GUI Server
[23:12:03] mdrun_gpu returned
[23:12:03] NANs detected on GPU
[23:12:03]
[23:12:03] Folding@home Core Shutdown: UNSTABLE_MACHINE
[23:12:07] CoreStatus = 7A (122)
[23:12:07] Sending work to server
[23:12:07] Project: 5801 (Run 9, Clone 173, Gen 0)
[23:12:07] - Read packet limit of 540015616... Set to 524286976.
[23:12:07] - Error: Could not get length of results file work/wuresults_00.dat
[23:12:07] - Error: Could not read unit 00 file. Removing from queue.
[23:12:07] EUE limit exceeded. Pausing 24 hours.

Btw, what does this line mean? [23:12:03] NANs detected on GPU

[osgorth]

Well, I guess it's *one* way of dealing with the server issues.. Release a new project that kills all GPU folders.

All my GPUs are now dead, and I'm going to keep them that way until this and the server issues has been fixed.

[theo343]

Well, I guess it's *one* way of dealing with the server issues.. Release a new project that kills all GPU folders.

I was thinking the same thing. Release a WU that makes all clients fail until it pauses for 24 hours to get the WU and collection serveres back on theire feet, but I surly hope not thats a plan of action. Would be much easier to ask everyone to back off for 48 hours hehe.