anyone else got the winSMP hanging sometimes?

I

Imitation

2[H]4U
Joined
Jun 18, 2004
Messages
2,536
So the last couple of days my SMP has just hung a couple times at random. I closed it and reopened it and it would go again. But it's happened 3 times now, I'd hate to miss deadlines because it can't keep itself going properly :) I haven't even been on the computer when folding hung. Anyone else got this issue?
 
Did it just stop finishing frames but looked like it was still running? Did you check task manager to see if the cores were idle? I get that occasionally as my wireless isp sometimes drops out, therefore it just hangs. Its a known issue with stanford that if your network/ip address changes the client will go into an idle state.

 
Yeah, mine hangs all the time. My desktop usually doesn't, but my lappy does and ends up with expired WU's. I swapped the lappy back to two regular cores.
 
I noticed that when it hangs, it is usually when the network was having a hiccup. The SMP is sensitive to network issues so a single drop of the connection, even for a few seconds will stall fah.

 
Yeah I think the processes had actually died, but I didn't check this last time, I just noticed that it was still stuck on frame 47 and closed and restarted. The first couple times though I'm pretty sure that the processes had died because my cpu temp was about 36, instead of the usual 40-42 I get while folding.

UPDATE: I just came home for lunch and the window is still here, but the processes are definitely gone. This is very discouraging because it was running great for about 3 weeks and now its having issues.
 
I have a desktop and a laptop running c2d / SMP and they both have stalled on me today.

I stopped and started them again, and all seems fine for now.

On my desktop, it finished a WU, sent it off okay, grabbed a new WU and then just kinda sat there for hours doing nothing. No CPU resources being used tells me that it is stuck or stalled.

Weird.
 
SpoogeMonkey said:
<snip> Its a known issue with stanford that if your network/ip address changes the client will go into an idle state.

Click to expand...

Note that this comment is directed toward Windows SMP client --- Once I started looking at all of my stalled situations with this client, they definitely occurred when I had an IP lease renewal. (But every IP lease renewal does not seem to cause a stall.... that's where I get confused...)

I've had lot's of stalls with the Win SMP client, and a few with the linux SMP client. I had been so confident of the linux client that I put it on a headless boxen..... MISTAKE !!! .. After my ppd tanked for 4 days, I finally got in touch with that boxen and guess what.... it was stalled.....:rolleyes: So at this point I definitely like the SMP client for points, but will only put it on a box that I can check in with every day or 2

 
Yeah so it stalled again last night or this morning. I didn't manage to sit down to it before work to check it. All the other fah processes are still in the task manager, but the 4 core processes are gone. This is getting frustrating.
 
I've had the client window itself hang, but the cores keep running. The client software will stop at a certain percentage complete, say 50%, so I'll shut it down and restart it and it'll say 55%. I've never experienced the cores themselves shutting down.

I hope the latter doesn't happen as I wouldn't want to miss a deadline, the SMP client just pumps out so much production that I don't want to have to switch back to the dual standard client setup.
 
Yeah it happened again last night. I dropped my cpu multi to 8x and set the fsb to 400 to yield me 3.2 instead of 3.5, maybe somehow my machine has become slightly unstable at that speed. You know, now that I think about it, this didn't start happening until after that last windows update.
 
It doesn't happen every time, but occasionally it will stall out if I unplug the cable to my 2nd NIC on my work PC. I described the issue to the fine folks over on Stanford's site and they got sufficient information to "look in to it", so hopefully they release a new Beta in a week or two with some updated info.

The 2nd NIC connects to a hardware box sitting on my desk (which doesn't have internet access), and my 1st NIC is the cable that gives me access to the Internet. *SHRUG*

The major symptom you will notice with this situation is that the four threads will go MIA on you... like they JUST did not 20 seconds ago for me. *SIGH*

I think they will show up again over time, but I don't like chancing it so I close it and start it up again whenever I see this happening.

A quick pic of my Task Manager and the time:



Hope this helps someone identify a 'hung' WinSMP so they can restart it! Ctrl-C and re-opening it usually solves the problem and it MIGHT have to do with the 8-hour attempt to transmit completed WU's to Stanford. *SHRUG* Your guess is about as good as mine ATM.


202276
 
That's exactly what's been happening to me with the screenshot. No more work processes. But I haven't been messing with my network at all. It happens when no one is home. I don't know if its comcrap or my router or on stanford's end.
 
I think you should limit your bit torrent pr0n downloads.




:D
 
Mine stalled last night about 30 minutes after I went to sleep. I lost about 14 hours of work:mad: I have to keep my temperature monitor up to be sure that it is running. Thankfully, it will stil make the deadline on this one.

 
Yeah mine is still hanging, it appears to be random as far as I can tell. I've went all the way down to stock speeds and it's still hanging so i really doubt its my machine. Either windows is screwing it up or the client is having some issues.
 
For what it's worth, I have completed 34 WinSMP units on my C2D E6600 @ 3.33 with zero problems. 14 units done on a X2 I have running WinSMP and no issues there either.
 
I have to reboot one computer every 24 to 36 hours due to some network issues. (Specifically, a Linksys router that hangs after too many connections have been made.) Right after I unplug the ethernet from my C2D, the SMP stops running. That is all it takes... no ethernet, then no SMP. The SMP just stops... but there is no error message. The CPU usage goes down to 18% or so from 100%.
 
And if you leave it hanging too long it kills your wu and gets a new one. That really sucks butt. Especially when you are 90% done. :( I noticed this last Saturday when I lost ip and it sat idle for 5 hours. Not even close to missing the deadline.
 
Imitation said:
Yeah so it stalled again last night or this morning. I didn't manage to sit down to it before work to check it. All the other fah processes are still in the task manager, but the 4 core processes are gone. This is getting frustrating.
Click to expand...

ive had this issue at home with my dual opteron machine
 
Anyone got the official support forums link? I might go over there to see if they have any more words of wisdom.
 
That's kinda strange though, my internet connection has a dynamic IP and I didn't have any problems with the SMP client until a week or so ago.

I had my new boxen hang long enough for it to fetch a new WU, I luckily made it to my main machine in time to restart the client as it hand been stalled for more than six hours. I'm sure that comes close to deleting the WU and getting another one.

I think I'm going to switch my new boxen over to GPU to see how that goes, it has an X1900XTX and it should work pretty well with the clocks cranked to 3D as it isn't in a case.
 
Did the Windows OS Critical Updates effect the network connections in some way... thus messing up the SMP's ability to run smoothly?



 
Pocatello said:
Did the Windows OS Critical Updates effect the network connections in some way... thus messing up the SMP's ability to run smoothly?



Click to expand...

It's possible if a update is affecting the networking stack. If you don't want to risk this, disable automatic updates and run windows update yourself so you can restart if it stall.

However, I had a laptop running windows smp and it's ok with the automatic updates. The only time it stalled is when the router was having a hiccup.

 
Yeah mine ran great for about a month and then last week it starts this hanging crap about 2-3 times a day last weekend. Thanks for the link killer.
 
You must log in or register to reply here.
Back
Top