Boinc

[danzigrules]

Back when I ran boinc there was all this jazz about optimized clients and whatnot. Is this still going on, or are they all the same now?

Do people agree that optimized clients are the way to go, or still fighting about it?


Thanks

danzigrules

 

[petrusbroder]

Well, the optimized clients = (applications, i.e. the program which crunches) are the way to go ... nowadays there is very little fighting about it.
The developers themselves are creating optimized apps: e.g. in Einstein@home, Seti@home and some other projects the applications are quite efficient. Probably because these projects employed the best guys ... and the devs realized that they get more science crunched that way.
For the Mac-platform there are some more and some less optimized applications (those who use the Altivec-compiled apps deliver results much faster).
Seti and Einstein have now only Altivec-optimized apps for the Macs.
applications - OTOH: these are often well developed and use quite often SSE2 or what not.

 

[SETIdude]

optimized BOINC client.. don't see the point anymore, the application determins credit granted these days, not the BOINC client program.

optimized applications are still being wildly used due to how much faster they are than the stock apps. I run optimized apps on all my machines.

This might be confusing, so to put a simple way, use the stock BOINC program, and change the science program to an optimized one.

 

[petrusbroder]

I agree with SETIdude - unless the optimized applications does not deliver correct results (which happens now and then - but not often at all).

 

[Smoke]

So where's the best Seti client for Windows XP?

Download link and instructions would probably help a lot of the people joining in for the King's Gauntlet.

 

[petrusbroder]

Originally posted by: Smoke
So where's the best Seti client for Windows XP?

Download link and instructions would probably help a lot of the people joining in for the King's Gauntlet.

Here for example ... Instruction should be included ...

Here too are a few ...

The instructions:

Installation of the optimized app is simple.
Copy all files in the "Files to install" directory to your BOINC\projects\setiathome.berkeley.edu folder.
There is a file called "app_info.xml" included that will make BOINC use this app automatically.
To use it, exit and re-open the BOINC Manager if you're running it in single-user mode, or re-start the service if you're using service mode.
That's it, you're done! You can open your task manager to see that it's really the new optimized app crunching your WUs.
BOINC will tell you that it found app_info.xml and is using an "anonymous platform".
You can find updates, instructions and source code here:
http://lunatics.at/

Happy crunching!

Simon Zadra
simon<squiggle>zadra.org

Edit: added instructions.

 

[bryanW1995]

I just started using the program that petrusbroder listed a few days ago. I'm seeing about an 8% increase in RAC on the rigs that are using it. My cpus seem to run a couple degrees C hotter but they're working harder

 

[Khyron320]

Where can i get 1 that is optimized for HyperThreading =P

 

[petrusbroder]

The applications detect hyperthreading themselves - if hyperthreading is activated in BIOS. Please note, that the sum of two threads on one processor is not double of a non-hyperthreading processor, it is a few % more ... but the heat is quite a lot more, so watch out if you do not have a very efficient HSF or watercooling.

 

[Assimilator1]

Thanks guys ,useful info

 

[bryanW1995]

I think that I've greatly underestimated the impact of the optimized clients. I was consistently running 3700-3850 RAC for the past 2 weeks with all my machines running full throttle. I have changed nothing other than a very tiny oc reduction on my e6750 (from 3504 to 3488 to reduce vcore/temp), and my total RAC is now up to 4151. It has been steadily rising, too. I wouldn't be surprised if I end up closer to 5000 with no other changes at all!

 

[petrusbroder]

Originally posted by: bryanW1995
I think that I've greatly underestimated the impact of the optimized clients. I was consistently running 3700-3850 RAC for the past 2 weeks with all my machines running full throttle. I have changed nothing other than a very tiny oc reduction on my e6750 (from 3504 to 3488 to reduce vcore/temp), and my total RAC is now up to 4151. It has been steadily rising, too. I wouldn't be surprised if I end up closer to 5000 with no other changes at all!

Thanks for the info - this gives food for thoughts and actions ...

 

[Assimilator1]

Originally posted by: bryanW1995
I think that I've greatly underestimated the impact of the optimized clients. I was consistently running 3700-3850 RAC for the past 2 weeks with all my machines running full throttle. I have changed nothing other than a very tiny oc reduction on my e6750 (from 3504 to 3488 to reduce vcore/temp), and my total RAC is now up to 4151. It has been steadily rising, too. I wouldn't be surprised if I end up closer to 5000 with no other changes at all!