More A1 Core WU's under notfreds

W

Wesleynator

Limp Gawd
Joined
Dec 26, 2008
Messages
182
For the last week or so, I have been assigned a rash of A1 core WU's. I have 7 VM clients running and have had as many as 5 of them running the A1 core WU's at the same time. Right now I have 4 of them. PPD is half of the A2 core WU's. Is everyone else seeing the same? This does take away some of the appeal of running the VM's but looking at the glass being half full at least some of the SMP's are making roughly double what a normal Win SMP would. Lastly, I assume it would be in poor taste to "dump" these slow ones to try and get a faster WU. Poor taste aside, it would be a lot more effort and I assume all science is important or Stanford wouldn't put these in the assignment queue if they didn't need the help. Comments?
 
I just take whatever I'm given and process them the same. I've noticed it comes in waves like that, and you'll get back to having A2 cores before you know it.
 
My LinSMP clients haven't gotten any a1 units lately, but sometimes they do send them out. Nothing much you can do aside from just folding them and hoping you get a better unit next time (or you can wipe out the queue.dat file and hope that you get an a2).
 
I only check my SMPs once a week and frankly haven't checked to see how many A1's I've gotten. I've gotten some here and there, but the vast majority on my 9 Linux SMPs are A2 units.

 
One of my clients is grinding away at an A1 right now. Haven't checked to see if my PS3 has gotten any.
 
Currently, 6 out of 7 of my vm clients are running the slower A1 clients. This really drags down the ppd however if this is the same everyone is seeing then so be it. I just wanted to make sure I didn't have some switch set in the client that was leading to this. Adv Methods is set to yes.
 
My other client finished it's A2 and downloaded an A1 so I have them on both now. Guess it's going to be slow-going for a while.
 
I got sent two or three A1 SMP WUs several days ago. It tends to happen whenever the servers are down for a while, even if it's the GPU servers. I think it's a method for Stanford to slow down the faster clients in order to help free up their infrastructure when there's issues, but aren't sure of this. I just notice the correlation. Personally, I wish I can revert to a complete A1 SMP regime if at all possible.
 
APOLLO said:
I got sent two or three A1 SMP WUs several days ago. It tends to happen whenever the servers are down for a while, even if it's the GPU servers. I think it's a method for Stanford to slow down the faster clients in order to help free up their infrastructure when there's issues, but aren't sure of this. I just notice the correlation. Personally, I wish I can revert to a complete A1 SMP regime if at all possible.
Click to expand...

HAHA .. nice idea but not the way it works ....

Your system gets a work units from Server #1
When it is complete, it send it back to Server #1
Then it asks Server #1 for more work.
If Server #1 is out of work or down, you might get send to Server #2 or #3 & so on.

Some SMP servers have more A1 work, some more A2.

It just depends on what server you get.

All the work units need to be completed and someone has to fold them.
Your PPD may go down for a day, but look how much you are doing for science ...
 
see what ya do is try to find out which server has the most A1 WU's.. then ip block the other servers that dont usually have A1 work with your firewall or router firewall.. problem solved, apollo :p..
 
EvilAlchemist said:
All the work units need to be completed and someone has to fold them.
Your PPD may go down for a day, but look how much you are doing for science ...
Click to expand...
I don't care that much about my PPD. I would prefer to switch to A1 units because the A2 unit uploads are ridiculously high.
 
At first I was spoiled by the lure of the A2 core units and roughly 2x the production vs WinSMP. This made putting up with the extra babysitting of the vm stuff worth it. Currently 2 of 7 vm clients have an A2 core unit and its been this way for roughly a month. I suppose this is no different really than the GPU units which vary on my boxen form 3500 to 5700 ppd depending on the luck of the draw. As Ax said, they all need folding and the times you do get an A2 it is a bonus probably just attractive enough for me to stay wit the VM stuff. WinSMP is easier for me to admin because you can put it in your start up folder and if you have to reboot it comes back up automatically. If anyone knows how to do this with a VM please post. Keep folding.
 
Last edited:
Wesleynator said:
WinSMP is easier for me to admin because you can put it in your start up folder and if you have to reboot it comes back up automatically. If anyone knows how to do this with a VM please post. Keep folding.
Click to expand...

Task Scheduler.

I use it to run any startup program that needs confirmation for UAC. I can run pretty much any program with highest privlages, on startup, with no prompts. For instance, Riva Tuner which is a notorious offender of needing UAC acknowledge on startup and load.

I'm was just looking how I have startup currently handled. For the stuff important to me folding.

VM0 - shortcut in startup folder shortcut starts program minimized, short cut to the folding.vmx file, I use TrayIt! to minimize all clients to tray.
VM1 - shortcut in startup folder starts program minimized
GPU0 - shortcut in startup folder starts program minimized
GPU1 - shortcut in startup folder starts program minimized
WinACF - scheduled task (to gain admin privileges)
Trayit! - scheduled task (to gain admin privileges)
Rivatuner - scheduled task (to gain admin privileges) but this is the exact same command as the one normally in HKLM\software\microsoft\windows\currentversion\run which I now have to delete from the registry every frigging time I change riva tuner)
 
Last edited:
You must log in or register to reply here.
Back
Top