Optimised for Linux?
26 Jun 2006 11:25:19 UTC
Topic 191460
(moderation:
Just trying to find out if there is a way to run optimised s5 apps on a linux machine. All the ones ive seen posted are for windows crunchers.
I guess if worst come to worst i can see if i can get my old windows CD working again but id much prefer to keep at linux as im just getting to love it after some teething problems.
Any help would be much apreciated
Thanks
Edit :just noticed i have no credit yet in my sig. rest assured i have a 4 Wu's finished and 2 credited finally and plan to crunch more no matter which way i end up doing it optimised or not
Optimised for Linux?
)
I hope that Akos cooperate with Bernd on new wersion of official application and his new improvemets will be take in official app.
The 4.x boinc client will run
)
The 4.x boinc client will run under wine. There was an issue with boinc 5.x and wine, and I'm not sure if it was ever resolved. THe S4 windows science apps ran under wine as well, so I assume the s5 ones will as well.
So far, Akos's s5 improvements have been from outsmarting the compiler, not making architectural changes that can be ported into the C++ source. While they could be added to the x86 based apps for other platforms embedding asm would go against the projects apparent goal to keep the science apps source as similar as possible between platforms. S5's long enough that once akos's tweaks stabilize an optimized binary might be made for x86 *nix and mac's, but I wouldn't hold my breath.
RE: Just trying to find out
)
we have all been asked not to use Akos's patched windows versions for now.
Just a note I think you will find in Windows std app vs Linux std app that Bernd has done a good job with linux and it's app is faster anyway. At least on my computers my AMD 2400+ XP Linux system crunches long WU in ~28500 vs my AMD 64 X2 Windows XP system crunching in ~35700.
Am i wrong or does the linux
)
Am i wrong or does the linux app perform better at S5?
Same Host (Dual Boot)
Win: 1_0796.0_S5R1__685_S5R1a 80756s
Lin: l1_1206.0_S5R1__710_S5R1a 64203s
Both are long WUs but the linux one is about x1.26 faster
edit: CPU is a P3m 1GHz
Well it looks like the
)
Well it looks like the optimised apps have been pulled so ill just live with my current one.
The majority of my hosts are
)
The majority of my hosts are linux boxes (none of them too flash), yet since starting S5 my rac seems to be increasing, is this a result of linux now getting equal credits for equal wu's (compared to windows)?
The benchmark based credits
)
The benchmark based credits have never been fair to Linux users as the BOINC Linux benchmark was sluggish.
The FpOps counting method used by S5 has some overhead (program-wise) but it still reflects the real performance of a machine better.
I have noticed this also, my
)
I have noticed this also, my RAC was normally +/- 55, but with S5 units I get 170 credits a day, and this computer is a old one a celeron 700 Mhz (0.7 Ghz in modern terms).
Redhat 9 with 2.6.14.3 kernel
Before i started this project
)
Before i started this project my RAC was sitting at around 100 on a 2Ghz machine. Ive since added a second machine to my little farm and its only a cleron 700Mhz yet my RAC has almost doubled and is still climbing. I guess getting equal credit with windows boxes has a lot to do with this.
RE: Am i wrong or does the
)
they are both "long" but still different. you can ONLY really compare between WUs that give the same Credits at the end.
same at Stephen R. Its hard to compare, if you also post the credits you got for the WUs besides the time that would help...