PDA

View Full Version : any Linux LLR guys out there?



NVM
03-14-2006, 01:55 AM
my question is is Linux faster than Windows doing this(LLRNet) ? i've switched a couple of Suse boxen over and am not too impressed so far.is it than my Linux boxes are AMD(Bartons) machines(i'm not saying that to raise your hackles, guys).

but , its early still so any experience you can pass on would be great.

Beerknurd
03-14-2006, 03:34 AM
I took 2 Linux classes in college an vowed never to use it willingly again. I hate Linux with a passion. Damn you Linus Torvalds!!!!!!!!!!!! (I got a B in Linux Administration.... It messed up my GPA!!!!!!!!!!!)

http://www.cs.helsinki.fi/u/torvalds/linux_logo.gif:rightfighter5:



:lol:

NVM
03-14-2006, 03:40 AM
i'm leaning more and more that its the cpu. for instance, in 17 or Bust,which this project is similar to, a P4 will rock compared to a AMD.

Beerknurd
03-14-2006, 04:06 AM
P4's normally rule the math projects. But Linux is supposedly faster in eon. But since the University of Texas took over and the new client was released I don't know if that is accurate anymore. I don't know of any other projects where the OS makes a difference. AMD is better in Boinc projects. My AMD 3300+ killed my P4 3.4 HT machine.

Intel and AMD, Windows and Linux... They all have their advantages in some way. You just have to play with it and see what works best.

drezha
03-14-2006, 07:13 AM
Linux is faster in some projects because it uses less overhead than Windows therefore giving the DC project running more work.

Linux is bad on BOINC because the benchmark system favours windows :?
Though this is set to change with SETI beta because they benchmark differenlty or something. :)

AMDave
03-14-2006, 07:22 AM
Hi Mike.

Hardware-wise:

Per the project's forum, AMDs are better than Intels for Seiving
Intels and AMDs are both ok at LLR, but the Intel is marginally better.

Software-wise:

With window$ you are pretty much stuck with the pre-compiled version.
With Linux you get to roll your own :D so it is as fast as you can find compiler optimisations for it.