Two processes on the same machine (2 cpus) running same unit ?

[Markfw]

I have 2 directories cpu0, and cpu1, and most of the time they are running different units, but once in a while, they will be evaluating the same unit, but at different places in the eval (at work now, can;t give the details)

Does this happen to anyone else ? And do I still get credit for doing it twice ?

 

[Confused]

Which project are you talking about? Different projects work differently for multiple CPU systems 🙂


Garry

 

[Assimilator1]

Originally posted by: Confused
Which project are you talking about? Different projects work differently for multiple CPU systems 🙂


Garry

😉

 

[Markfw]

Stupid me. F@H. I'm sorry, brain fart....

 

[GLeeM]

Markfw900
When you say "same unit" do you mean same protein project, p***

Each project has 100s or 1000s of different WUs, the difference in WUs of same project is in the "Run **, Clone **, Gen ** numbers. If these are different then they are not the same WU.

You have two different directories which is needed.

Have you also configured each client with a different MachineID? This MUST be done. You can have up to eight MachineIDs per computer.

You only get credit for one 🙁

If you need more directions or have any questions, please ask 🙂

Thanks

 

[Markfw]

Originally posted by: GLeeM
Markfw900
When you say "same unit" do you mean same protein project, p***

Each project has 100s or 1000s of different WUs, the difference in WUs of same project is in the "Run **, Clone **, Gen ** numbers. If these are different then they are not the same WU.

You have two different directories which is needed.

Have you also configured each client with a different MachineID? This MUST be done. You can have up to eight MachineIDs per computer.

You only get credit for one 🙁

If you need more directions or have any questions, please ask 🙂

Thanks

The machineid is possibly wrong, if so, I have been only getting half credit on that machine for 3 weeks now !!! How do you change that ? Is the machineid field in client.cfg ? and the valid values are 1-8 ?

 

[GLeeM]

Bummer! Sorry Mark

A while back I questioned what seemed like low pointage for you. I should have thought to ask about your MachineIDs 🙁

Any time you run two or more clients (only one can be GUI) you MUST set MachineID on the console (GUI is ID 1). If you have installed the GUI at any time, you must also use the -local flag when starting a console.

Yes, the MachineID is in the client.cfg file. Maybe you can change it there, I don't know for sure.

I know for sure that you can change it during config. Make a new shortcut and add flags: -local -config (-configonly quits when done configuring)

Answer "yes" when you get to "Change advanced options". 1-8 are the only numbers accepted. If you have GUI it is 1.

Now after all of that I remember something!!
If I am sending results from a no-net computer (I use a thumb or pen drive to move WUs) and I forget to look at what MachineID is running on my net rig and what MachineID is on the thumb and they are the same, the newly started console sees that its MachineID is already running and quits. It leaves a message in the log that it will run reguardless on the next start attempt. So if you start a console and it quits immediately, that is a sure sign that MachineIDs are the same!

Thanks 🙂

 

[Markfw]

Yup, I fixed it, I set the cpu1 to ID2, and I see it in the window (running console), so does that mean on the current work units, I won;t get credir twice ? How do I force it to get me a new unit ?

I now have two of " Project: 1301 (Run 19, Clone 2, Gen 0)"

 

[GLeeM]

I think you can just delete the contents of the work folder. But to be safe, do this to the console with the least work done, rename the work folder to worktest. Now restart and I think it will make a new work folder and get a new WU.

🙂 Now your ppd will go up just in time to pass me! 😉

EDIT: Just thought of another way, the right way

Start with -queueinfo flag, it writes the info to the log, find out what queue position number the WU is in. Restart with -delete # flag, where # is the queue position number

 

[Markfw]

Thanks, but it looks like it may be the morning before I pass you. Maybe my stats will go up ! Remember I go on vacation sunday AM, and will shutdown most of my workers to not take a chance of a fire while I am 3000 miles away ! (accidents do happen, and its not worth it for F@H, PSU's do explode, but most of mine are Antec)

 

[GLeeM]

Enjoy your vacation 🙂

Gravity will hold down the fort until you get back.

Look at jliechty's post half way down this thread. 😕
Edit: and the last post too.

Thanks 🙂

 

[Markfw]

I had to re-install/reformat my drive for win2k, and forgot about these. Damn Tyan drivers ! Thanks for reminding me ! I must have f-'ed up after the re-format.