Posts by arkayn

Is it possible to download the new optimized apps somewhere without the installer, i.e. just a zip file with all needed files inside, like it was possible before? Some of us don't need installers, that do more than it's actually needed...

Sure... but only for GPU apps. CUDA on Jason's site, OpenCL on Lunatics. All as separate packs.

AFAIK they dont contain V7 aistubs yet, no ?

you right. we need to change that.
For separate packs it's quite trivial task.

- delete app_info.xml from the project directory
- reset the project

New app_info.xml not created

Will those of us with Kepler class NVIDIA cards (ie GTX-670) need to delete the environmental variable changes we made (CUDA_GRID_SIZE_COMPAT)to accommodate the GPU problems in 6.10 CUDA ?

In the previous thread, there were some comments about VLARs going to Nvidia GPU's now. I"m still on v6 for another 10 hours while draining my cache, and I notice that I have 3 VLARs and a non-VLAR now running on my 670 with x41zc, Cuda 5.00. I don't see any adverse effects except that run times will be longer than normal and there is a little lag in the system (responsible for any typos in this post!) :)

But the odd thing is that my gpu temperature is more than 10 degrees below normal. No downclock and gpu utilization is at a constant 99%. CPU usage is below normal for 6.10 tasks. Why would these run cooler? I expected the opposite.

I see the exact same thing. My guess is that since VLARs don't do well parallelizing, you can't keep as many cores busy in the GPU. However, Precision X reports high CPU usage, especially with another task running on that same GPU. Less cores in use mean less heat.

I can see this is going to be a problem since not only does the VLAR run much longer compared to a normal work unit, but it degrades the other jobs running on that same GPU. I don't know if there is a workaround to this except for only running a single tasks at a time on all my GPUs. That's doesn't seem to be a very efficient use of GPU resources. I would much rather the GPU stay on the CPUs which do pretty well with them. I wouldn't care if all my CPU tasks were VLARs.

I haven't converted over to V7 yet. Still running the same x41zc with Cuda 5.0. Had some computer problems so I'm concentrating on that issue first.

Just awhile ago, I received a batch (maybe a dozen or so) VLAR tasks as GPU tasks. I was amazed to see this to say the least. This problem was fixed ages ago. I Rescheded them to the CPU and all was well.

A little while later, I received another VLAR as a GPU task and aborted it.

Question, has anyone else seen this behavior? Wonder if this has anything to do with the V7 apps now available? I know that makes no sense, but it's just a suspicion.

This works just fine as I got a load of stock CUDA42 work, I did have to make a few changes so I could run them full out instead of normal speed.

Haha, now for multiple instances, I'm going to have to either modify newer Boinc with adaptive flops/aDCF-equivalent, or backport app_config.xml support into modded 6.10.58. *sigh*

Probabily an "stupid" question, what i need to do to start crunching V7, i´m running CPU (Intel)/GPU x41zc apps. Anyone have a simple receipt? Thanks.

My recommendation (Subject to change after I try it) would be:
- Set No New tasks
- Wait until tasks all complete, upload & report
- delete app_info.xml from the project directory
- reset the project
- Allow New tasks

Extra notes: Probably need to check web options too in there somewhere [OK :)]. I'd try let the server decide amongst the compatible versions over some time with one instance per GPU (underutilised due to memory intensive Autocorrelations) then worry about multi instances after things settle in. You'd likely want either newer Recommended Boinc for that, or wait for installer for older Boinc clients to run with app_info (anonymous platform)

PSU - very goog. Enough. Be sure. SeaSonic X-660 (Gold)
GPU temperature - normal. Always control it.
Reinstall driver. Waiting for GPU WU. If error still occur maybe rollback to previous version. All began when I install new version.

Tell me I need to install http://jgopt.org ??
I'm advanced in PC and know how manual install.
Thanks

I did it

From Jason's post, note that AVX has a very high minimum system requirement - a second generation Intel Core i5 or Core i7 manufactured no earlier than 2011, or an AMD Bulldozer of the same vintage. Plus, in addition, Windows 7 Service Pack 1.

Arkayn and I happen to share a birthday, although he is significantly younger.

Happy Birthday Arkayn - Hope your special day was filled with fun.

Angela & Eric

Panic? Don't see a reason... people running stock probably won't notice anything or at least not much, people running opt apps will get messages about missing application in their app_info untill they update... so no real issues.

Is the "next Thursday" just a rumor or was it at least somewhere mentioned by anyone from the staff?

Eric said so yesterday: http://setiweb.ssl.berkeley.edu/beta/forum_thread.php?id=2012&postid=46049

No panic for this slow old OSX cruncher, as it seems only for Windows and Linux.
My even older XP machine has been assigned to another project meanwhile, it wasn´t contributing in a very meaningful way to SETI.

With a pizza a good red Italian

I should be mostly good as I am only crunching on the GPU's currently and I am already using the app that will be the stock app.

So I will just have to make some changes to my app_info and fire BOINC back up.

So....

I'm running Windows XP with BOINC 7.0.64. What would I have to change in this App_Info come next week?

Just set NNT, remove the app_info, reset the project, restart Boinc, unset NNT, the apps the project will supply will be newer than what you're already running, then wait for the following Monday.

Claggy

v7 will work with Windows XP Pro sp3?

Now, where is Dune Finkleberry when you need him .....

OK I checked it out and the draft beer is called Ninkasi from oregon and of course I had another. :))

http://beeradvocate.com/beer/profile/14400/37966