Send the WU I don't like this Windoz SMP

F

fenderltd

[H]ard|DCer of the Month - July 2007
Joined
Feb 10, 2004
Messages
2,432
[18:16:07] Loaded queue successfully.
[18:16:07] - Preparing to get new work unit...
[18:16:07] + Attempting to get work packet


[18:16:07] - Connecting to assignment server
[18:16:07] + Attempting to send results
[18:16:07] - Successful: assigned to (0.0.0.0).
[18:16:07] + News From Folding@Home: Welcome to Folding@Home
[18:16:07] Work Unit has an invalid address.
[18:16:07] - Error: Attempt #1 to get work failed, and no other work to do.
Waiting before retry.
[18:16:08] - Couldn't send HTTP request to server
[18:16:08] + Could not connect to Work Server (results)
[18:16:08] (171.64.65.64:8080)
[18:16:08] - Error: Could not transmit unit 05 (completed April 11) to work server.


[18:16:08] + Attempting to send results
[18:16:20] - Couldn't send HTTP request to server
[18:16:20] + Could not connect to Work Server (results)
[18:16:20] (171.65.103.100:8080)
[18:16:20] Could not transmit unit 05 to Collection server; keeping in queue.
[18:16:21] + Attempting to get work packet
[18:16:21] - Connecting to assignment server
[18:16:22] - Successful: assigned to (0.0.0.0).
[18:16:22] + News From Folding@Home: Welcome to Folding@Home
[18:16:22] Work Unit has an invalid address.
[18:16:22] - Error: Attempt #2 to get work failed, and no other work to do.
Waiting before retry.
[18:16:46] + Attempting to get work packet
[18:16:46] - Connecting to assignment server
[18:16:52] + Could not connect to Assignment Server
[18:16:53] - Successful: assigned to (0.0.0.0).
[18:16:53] + News From Folding@Home: Welcome to Folding@Home
[18:16:53] Work Unit has an invalid address.
[18:16:53] - Error: Attempt #3 to get work failed, and no other work to do.
Waiting before retry.
[18:17:26] + Attempting to get work packet
[18:17:26] - Connecting to assignment server
[18:17:32] + Could not connect to Assignment Server
[18:17:32] - Successful: assigned to (0.0.0.0).
[18:17:32] + News From Folding@Home: Welcome to Folding@Home
[18:17:32] Work Unit has an invalid address.
[18:17:32] - Error: Attempt #4 to get work failed, and no other work to do.
Waiting before retry.
[18:18:22] + Attempting to get work packet
[18:18:22] - Connecting to assignment server
[18:18:23] - Successful: assigned to (0.0.0.0).
[18:18:23] + News From Folding@Home: Welcome to Folding@Home
[18:18:23] Work Unit has an invalid address.
[18:18:23] - Error: Attempt #5 to get work failed, and no other work to do.
Waiting before retry.
 
I'm getting the same stupid thing. I am seriously considering quitting folding because of this stupid SMP client.
 
I'm getting it too. Apparently it is not just the SMP client because there is one guy on the FAH forums who said he has it with two regular console clients as well. I guess we just have to wait it out and hope our WUs are submitted before their deadlines.

Lyquist, you are very impatient. You have to remember this is a beta client. If you don't want to mess with the errors that are likely to be there with a beta client, go install two regular console clients and wait for a stable SMP client.
 
I'm also getting the same error. I read something the other day on the FAH forums about this regarding when it gives this error and assigns you to 0.0.0.0. Admin said only thing you can do is wait for Stanford to get the SMP servers back online. They did extend the deadlines for some WU's due to this from what I remember reading.

Just re-checked there. Seems to be the most popular problem at the moment.

http://forum.folding-community.org/forum60,smp-for-windows-beta.html

After reading some more at the FAH forums it seems that the problems are probably on Stanfords end. Looking at the server stats page nearly half of the servers are either down or in reject mode.
 
Lyquist said:
I'm getting the same stupid thing. I am seriously considering quitting folding because of this stupid SMP client.
Click to expand...

Why would you quit folding? If the SMP client is giving you the hassle then just go back to the standard console client. I have over 90 CPUs reporting and a whopping 3 of them run the SMP client. Now when Stanford gets the bugs worked out (remember this is still in early beta) then I will switch to more SMP but as it sits now it just isn't stable enough and requires too much hassle to monitor every day so I keep right on crunching just for smaller point gains but the science wins either way
 
Tormond, you never anwsered my PM I sent you last night.

Yeah, one way is to basically delete all the work re do the bat and it helped me last night on one machine. Also..... That machine had just submitted the work and got accepted so I didn't have anything to lose! It's beta, it's frustrating, I don't feel it's a reason to quit folding, Beta's are released to the public to test and find bugs... If Stanford said the Windows SMP was a Private Beta, you would be pissed that the beta was private/selected user tested. It's will get you mad, but remember, we are all folding to save lifes.

Ninja edit - Mine just started working again and picked up a project 5:08 pm (CST)
 
Yes. Beta software. If you have issues with it, and you agreed to used it during the beta period, you should report your findings to the software maker so that when the FINAL client comes out, it won't have that problem.
 
In this particular case it was not a problem with the client though. It was Stanford's servers being overwhelmed or dorked up in some way. I discovered I had three machines giving this error, ont one was running the SMP client. Suddenly about 5 minutes ago all three were able to send their results and get new work units. One was down since yesterday, the other two for just a few hours today.
 
xfalconx2 said:
I'm getting it too. Apparently it is not just the SMP client because there is one guy on the FAH forums who said he has it with two regular console clients as well. I guess we just have to wait it out and hope our WUs are submitted before their deadlines.

Lyquist, you are very impatient. You have to remember this is a beta client. If you don't want to mess with the errors that are likely to be there with a beta client, go install two regular console clients and wait for a stable SMP client.
Click to expand...

I agree that I'm not very patient...but I lost about a weeks worth of units because they weren't makeing the deadlines because the service was also running. I didn't think it was supposed to run in service mode. All I know is that my points are way down lately. I had to remove several borgs as well though...so that is part of it.
 
Guess I've been lucky and haven't had any problems so far. The only issue I've seen is the smp client going idle if you loose your ip address (known issue).Since I'm currently on very sucky wireless internet, my ip drops every couple days and I have to reset the router. Kinda sucks to have your box sitting idle for 18 hours if you don't check up on it every morning.
 
Ya about the SMP folding client, Until it becomes a stable smp client program, then I'll run it. But seeing how much trouble and hassle from everyone here that takes to get it to run, I'm not risking any of my computers with it.
 
Scorpionjwp said:
Ya about the SMP folding client, Until it becomes a stable smp client program, then I'll run it. But seeing how much trouble and hassle from everyone here that takes to get it to run, I'm not risking any of my computers with it.
Click to expand...

for me, it's fairly stable but it depends on machies :

-My home C2D computer has issues with the SMP because it refuse to get a new WU after the old one is done. It also lost me 1 week due to EUE or 0x1 errors :( I'm now running Ubuntu64 with the linux 64 bits SMP clients and it's giving me almost 2000 ppd.
-A T7200 laptop at work with XP is chugging SMP fine for 1 week now. Only one time, it stalled so I just restarted the client and it works.
-I also have a C2D E6300 on the bench for testing purposes prior to installation cugging SMP fine but had issues sending/receiving packets, not the fault to the SMP tho because the networking configuration is messed up and I must reinstall XP to make it work, which I will probably do today.

As long as you can watch each computer one time per day, it will be fine. For 24/7 100% unattended setup, it's not yet perfect however.
 
I don't know how you guys are having problems with this client. I've had it on four machines since it came out and have never had a problem.
 
Stanfords been having trouble with the servers again.

My SMP rig took 6 hours to get any work.

I've got two plain CPU clients that cannot download work at this moment.
One of which has been stuck now for 20 hours.
Its been sent to server (171.64.65.65), (171.65.103.106), (171.65.103.160), (171.64.65.58), (171.64.65.58) & (0.0.0.0) and still no work.

Luck ............ :D
 
I am having trouble now with even my console verisons not being able to send the WU after it's completed.
[14:20:26] + Could not connect to Work Server (results)
[14:20:26] (171.65.103.100:8080)
[14:20:26] Could not transmit unit 03 to Collection server; keeping in queue.
Click to expand...

I feel Stanford has some serious issues with their servers...
Last night one of my SMP boxes never sent the WU and downloaded another unit after the 0.0.0.0 error connecting to the assignment server
 
tiger, we posted @ the same time, about the same issue:D

 
I had the same issue with a WU now. I tried forcing it to send it and after 1 hour, it said it failed to send the WU. I pinged the 2 servers it tried and both works...
 
<start>
<rant>
I am about fed up with Stanford right now. 11 clients are not sending WU's and 8 have got work and NOT send results. It's great that all the hard work we do is not being used. I don't really care bout the points, but why I am running all these boxes if the WU's are not even being used? I hope Stanford gets there act together soon. I am not talking about the beta SMP client but the console clients seem to be giving me more trouble than the SMP's!
</rant>
</end>
 
fenderltd said:
<start>
<rant>
I am about fed up with Stanford right now. 11 clients are not sending WU's and 8 have got work and NOT send results. It's great that all the hard work we do is not being used. I don't really care bout the points, but why I am running all these boxes if the WU's are not even being used? I hope Stanford gets there act together soon. I am not talking about the beta SMP client but the console clients seem to be giving me more trouble than the SMP's!
</rant>
</end>
Click to expand...

the trick to these things is not pay attention.

Pass the pickles now please.
 
You must be correct Mary, if you never see that it's not sending the WU and not getting work, you would never know there are problems. This new theory could be a break through!
 
Heh with a very few exceptions I never have any idea what any of my boxen are doing.. I just watch the points roll in and if they drop off sharply then I know there is a problem somewhere)
 
Patience is the key with Stanford and their beta clients. I suggest that if you are fed up with Stanford's SMP issues, just drop back to the regular console client and FOLD!

There is always UD as well. No sense in quitting over something like this.
 
Oldbenwa said:
I don't know how you guys are having problems with this client. I've had it on four machines since it came out and have never had a problem.
Click to expand...

I have had zero problems with the SMP client as well, other than the recent troubles with sending/receiving WUs, but that is not the SMP client's problem. That's trouble at Stanford, in which case being patient is all that is needed.
 
An extra couple hours, or even days to get the wu to them is not going to hurt. I want to see cancer cured and new discoveries found just as much as all of you, but we have been living with these diseases for thousands of years, so a couple hour delay is not going to ruin everything. They are still going to use the data your computer computed. F@H is having some growing issues at the moment, and I trust that everything will get fixed up soon.
 
AtomicMoose said:
Patience is the key with Stanford and their beta clients. I suggest that if you are fed up with Stanford's SMP issues, just drop back to the regular console client and FOLD!

There is always UD as well. No sense in quitting over something like this.
Click to expand...

Besides one thread that a guy claimed he was going to quit, who else is saying this? I am not quiting even if I need to fold my Ti-83
 
I had a wu that couldn't send this morning. It finished at 5am and sat idle for 3 hrs til I checked on it. Restarted the client and it got a new wu, but it took til 3:30 to get the old wu sent in. At least my box didn't sit idle twiddling it's thumbs all day (imagine doing that 2.9 billion times a second for 11 hrs) :D
 
I got everything straightened out with the SMP client now. My production is back up where it should be now:)
 
I installed the SMP client last night and noticed that my processor usage is nowhere near 100%. Is this normal? When I used to have 2 clients running, both my processors were fully loaded. Why would anyone run a client that has issues when 2 regular clients have always worked very well? What's the main advantage to running the SMP client?
 
APOLLO said:
I installed the SMP client last night and noticed that my processor usage is nowhere near 100%. Is this normal? When I used to have 2 clients running, both my processors were fully loaded. Why would anyone run a client that has issues when 2 regular clients have always worked very well? What's the main advantage to running the SMP client?
Click to expand...

Points is the only "advantage" to the folder themselves.. Stanford gets work back faster that they think is more important is the upside to the science. Just make sure that you have enough PC to actually do SMP and meet deadlines. If you aren't sure then just stick with the console clients.. Also stick with teh console if these machines are in production anywhere simply as the SMP client isn't stable enough (IMO) to not require some handholding pretty much every day
 
Tormond said:
Points is the only "advantage" to the folder themselves.. Stanford gets work back faster that they think is more important is the upside to the science. Just make sure that you have enough PC to actually do SMP and meet deadlines. If you aren't sure then just stick with the console clients.. Also stick with teh console if these machines are in production anywhere simply as the SMP client isn't stable enough (IMO) to not require some handholding pretty much every day
Click to expand...
I have a dual unicore Opteron system and I was seeing only one processor being utilized and not completely either. It might have been some kind of bug because the client was stuck at the "working on unit 01" line and didn't display anything beyond that for nearly half a day. I installed the regular client and now I see full CPU usage. The standard client reports all the lines including the number of steps. I guess the SMP client is indeed problematic.

So, what you're saying is that having a working SMP client will award the user more points than two fully operational standard clients will combined?
 
APOLLO said:
I have a dual unicore Opteron system and I was seeing only one processor being utilized and not completely either. It might have been some kind of bug because the client was stuck at the "working on unit 01" line and didn't display anything beyond that for nearly half a day. I installed the regular client and now I see full CPU usage.The client reports all the lines including the number of steps. I guess the SMP client is indeed problematic.

So, what you're saying is that having a working SMP client will award the user more points than two fully operational standard clients will combined?
Click to expand...

Well a C2D e6400 running the standard client x2 will net you about 250PPD. The same machine running the SMP client will net you about ~1200PPD so yes the bonus points on the SMP client make it point wise FAR superior to the standard client.. That being said I have about 100 cores folding right now and run 3 instances total of the SMP client (and 1 of the PS3 which is kind of a SMP client but will net you ~900PPD per PS3)
 
Tormond said:
Well a C2D e6400 running the standard client x2 will net you about 250PPD. The same machine running the SMP client will net you about ~1200PPD so yes the bonus points on the SMP client make it point wise FAR superior to the standard client.. That being said I have about 100 cores folding right now and run 3 instances total of the SMP client (and 1 of the PS3 which is kind of a SMP client but will net you ~900PPD per PS3)
Click to expand...
Wow, thanks for the info. I had no idea the difference between clients was an order of magnitude. Are there any special flags to add or options to enable/disable when installing the SMP client?
 
APOLLO said:
Wow, thanks for the info. I had no idea the difference between clients was an order of magnitude. Are there any special flags to add or options to enable/disable when installing the SMP client?
Click to expand...

Read the read me ....... :p

Summary of instructions:-
1. Install Microsoft .NET framework.
2. Install Folding@Home SMP client.
3. Run install.bat from the F@H directory.
4. Start the F@H client.

The deadlines are tight but as long as you can do less than 55 mins a frame, you should be ok.
My X2's @2.2 Ghz will pull in around ~250 PpD with 2 CPU console clients, ~700 with the SMP client & ~900 with a CPU & GPU client.
At the moment my C2D @2.75 Ghz is pulling in ~1,650 PpD running both an SMP & a GPU client together but it not a "set and forget" setup as you first need to set the affinities then need to constantly monitor it so that each client only runs on its own core.
Its why I'm pulling in 6,500 PpD off 6 boxen ....... :D

Luck .......... :D
 
Tigerbiten said:
Read the read me ....... :p

Summary of instructions:-
1. Install Microsoft .NET framework.
2. Install Folding@Home SMP client.
3. Run install.bat from the F@H directory.
4. Start the F@H client.

The deadlines are tight but as long as you can do less than 55 mins a frame, you should be ok.
My X2's @2.2 Ghz will pull in around ~250 PpD with 2 CPU console clients, ~700 with the SMP client & ~900 with a CPU & GPU client.
At the moment my C2D @2.75 Ghz is pulling in ~1,650 PpD running both an SMP & a GPU client together but it not a "set and forget" setup as you first need to set the affinities then need to constantly monitor it so that each client only runs on its own core.
Its why I'm pulling in 6,500 PpD off 6 boxen ....... :D

Luck .......... :D
Click to expand...

Actually, it's 47 mins / frame or less from a post in the folding forum at Standford ;)


 
You must log in or register to reply here.
Back
Top