[Wien] Which time is relevant in the benchmark list ?

[Peter Blaha]

"Previously", there was little difference between "cpu" and "wall"-times 
(with the wall time a bit longer due to system overheads), but with the 
modern dual core cpus, the "cpu-time is no longer so much relevant, but 
the walltime is what really counts.

So in essence:
The timing on single cpus is the cpu-time (but the wall clock should not 
be larger than 1 sec on well installed and empty systems);
on parallel machines it is the wall-time.

PS: In any case, if the timing routines are ok, the result of the unix 
"time" command and the timing in output1 should be quite similar).

 ( > > Hello,
> 
> I am testing Wien2k benchmark on different platforms, and I would like to
> know which time is given
> in the following list : http://www.wien2k.at/reg_user/benchmark/
> That is to say :  Is it "real time" given by the UNIX command time, or a
> time indicated at the end of the output file ?
> If it is a time of the output file, which one ? "Total CPU time" ? "Maximum
> CPU time" ? "Sum of Wallclock Time" ?
> 
> This information is important for me to compare with accuracy different cpu
> systems,
> Thank you for your help.
> 
> Raphaël.
> 
> _________________________________________________________________
> MSN Hotmail : créez votre adresse e-mail gratuite & à vie ! 
> http://www.msn.fr/newhotmail/Default.asp?Ath=f
> 
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> 


                                      P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------