The preferred deadline is 4 days and the final is 6 so you do fall into the range to do them but you DO NOT get any time bonus so you are turning in these WUs for their regular credit, 20k. It breaks down to about 5k ppd for the machine, which is not bad but you might want to try just running a SMP2 client on there because it should get better numbres due to meeting preferred deadlines.
If you install HFM.NET on your machine and share the folder you can easily let it run for a few days and check on it or other stuff at work or simply install it locally on that machine. It will automatically due break downs for ETA and credit along with PPD per MHz and total completed/failed.
The preferred deadline is 4 days and the final is 6 so you do fall into the range to do them but you DO NOT get any time bonus so you are turning in these WUs for their regular credit, 20k. It breaks down to about 5k ppd for the machine, which is not bad but you might want to try just running a SMP2 client on there because it should get better numbres due to meeting preferred deadlines.
If you install HFM.NET on your machine and share the folder you can easily let it run for a few days and check on it or other stuff at work or simply install it locally on that machine. It will automatically due break downs for ETA and credit along with PPD per MHz and total completed/failed.
That's what I thought. It's been averaging around 20 or better steps per 24 hours which is how I equated my total. I'll set it to -oneunit and finish it and then go back to SMP2. I'm using 6.30 is that the SMP2 you are referring to?
Why what? Why doesn't HFM.NET work on here? I have no clue. I guess the server has it blocked or something. It works I just don't get all of the information fields filled in. I'll see if I can get a screenshot.
Just after I first posted I noticed that the wu shut down at 40% and sent results. Then it downloaded something else.
Look back at the log and see what was the result when it decided to shut down, it will always give a reason in upper case and "_" for spaces.
If you don't know what a project you downloaded is you can always check here. A3 cores are bigadv. It stopped running the WU you were for some reason and downloaded a new one, I am running a 6900 right now as well.
Look back at the log and see what was the result when it decided to shut down, it will always give a reason in upper case and "_" for spaces.
If you don't know what a project you downloaded is you can always check here. A3 cores are bigadv. It stopped running the WU you were for some reason and downloaded a new one, I am running a 6900 right now as well.
I deleted the log. I always delete the log. It's a bad habit I know.
I got into the habit when I was crunching on 25+ systems. It was quicker to hunt for issues when I used a fresh log. The habit never lifted unfortunately. I have a good feeling that something I did caused it to fail.
During the day when I am at my workstation I really tense and stress the system. Sometimes I'll have a few windows apps going, Lexco, ArcGIS, AutoCAD, Photoshop, and then get called on to open up something totally new. It's a wonder I don't bork up more wu than that one.
I just added this to the everything about folding thread but it seems rather interesting regarding the current discussion here. It is how I pulled that 5k pdd for you running a bigadv. If you run SMP2, I forgot this, and you have more than a 4 threads/cores you should add the -smp # flag, where number is the number of threads/cores.
I had -smp 8 but I was reading something about it and it said that if you are running the GPU client the you should reduce it to -smp7. I should have paid closer attention I guess. 7im had remarked on it at the community i think.
The only thing I have read about not using the correct # for the smp flag was relating to the hyperthreaded machines running bigadv and the fact that they were intended to be able to run those WUs. But it seemed all the machines could run them but some had stability issues and changing to an odd number seemed to help resolve this issue.
But if someone did say that, interesting. Link it up.
I looked around for it last night but I didn't see it again. I think I saw it when I was looking up minimum requirements to achieve -bigadv's. It was an instance where someone was having issues with it and was advised to drop from -smp 8 to -smp 7. If I see it again I'll e sure to post it.
Comments
If you install HFM.NET on your machine and share the folder you can easily let it run for a few days and check on it or other stuff at work or simply install it locally on that machine. It will automatically due break downs for ETA and credit along with PPD per MHz and total completed/failed.
That's what I thought. It's been averaging around 20 or better steps per 24 hours which is how I equated my total. I'll set it to -oneunit and finish it and then go back to SMP2. I'm using 6.30 is that the SMP2 you are referring to?
HFM.NET doesn't work on this workstation.
Why what? Why doesn't HFM.NET work on here? I have no clue. I guess the server has it blocked or something. It works I just don't get all of the information fields filled in. I'll see if I can get a screenshot.
Just after I first posted I noticed that the wu shut down at 40% and sent results. Then it downloaded something else.
If you don't know what a project you downloaded is you can always check here. A3 cores are bigadv. It stopped running the WU you were for some reason and downloaded a new one, I am running a 6900 right now as well.
I deleted the log. I always delete the log. It's a bad habit I know.
Deleting the log is a terrible habit.
During the day when I am at my workstation I really tense and stress the system. Sometimes I'll have a few windows apps going, Lexco, ArcGIS, AutoCAD, Photoshop, and then get called on to open up something totally new. It's a wonder I don't bork up more wu than that one.
But if someone did say that, interesting. Link it up.