Roy Schestowitz wrote:
> __/ [ The Ghost In The Machine ] on Tuesday 09 May 2006 01:00 \__
>
> > In comp.os.linux.advocacy, John A. Bailo
> > <jabailo@xxxxxxxxxx>
> > wrote
> > on Mon, 08 May 2006 13:23:13 -0700
> > <y5WdnX6bgZoONcLZ4p2dnA@xxxxxxxxxxxxx>:
> >>
> >> "Which is faster for stats, Linux or Apple's OS X?
> >>
> >> Linux, says Jasjeet Sekhon, associate professor, Travers Department of
> >> Political Science Survey Research Center at UC Berkeley."
> >>
> >> http://www.p2pnet.net/story/8723
> >>
> >
> > The webpage also includes WinXP. Benchmark 1 shows OSX Core Duo
> > to be the slowest, Linux Opteron to be the fastest, with XP in the middle.
> > Benchmark 2 shows Linux P4 to be the slowest and OSX Core Duo to be
> > the second slowest, XP in the middle again and Linux Core Duo to be the
> > fastest.
> >
> > Windows XP. When one wants to be squeezed in the middle. :-)
>
> To be fair, this depends on many factors: the programming language, the
> way in which the program was compiled (or, in some cases, *both* the
> application which interprets the code *and* the code), the number of
> background processes and even the nature of the computation.
>
> With bloat, I doubt that Mac OS X can ever win benchmarks as such. It is
> an operation system that in its core and concept is more about flash than
> substance.
The "flash" of OSX that you speak of exists mostly in the display and
user-interface elements of the OS. These benchmarks that were run
essentially do nothing with the UI. The reason for the slow performance
on OSX is due to the Mach micro-kernel design of the OS. This overhead
is inherently present at the system level regardless of the presense
(or absense) of eye-candy in the UI. You can't blame the flash in the
GUI for the performance. The micro-kernel design simply places
additional overhead into every system call. This decision isn't based
on "flash" - it's for robustness.
> Technical people will not deny the fact that Macs are far too
> expensive to be used as affordable computational servers (or an array
> thereof). Linux, on the other hand, has origins in datacentres where flash
> was (historically) unimportant and fast delivery of files, execution of
> files (e.g. PHP Hypertext Preprocessor), and efficient processing of data
> was a clear priority. Stability and freedom from intervention was another.
> On the third *cough* hand, Windows was intended to bring computers into
> the homes, at lease at the time (so did the Macs).
>
> I suppose the conclusion one can make is that Linux *was*, and still *is*,
> the platform for work to be done quickly. Performance-wise, Windows may
> indeed be stuck in the middle.
"Middle" can have two meanings here. When used to describe the "order"
of results Windows did finish in the middle; 2nd out of 3. If what
you're describing is the performance then it finished slightly behind
Linux. Example... it took OSX 22.78 seconds to complete the first
benchmark. Linux did this in 8.81 and Windows did it in 9.38.
"Performance wise" as you state, the middle would be 15.8 seconds.
> For this argument I leave aside the
> off-topic discussion where one would prove Macs and GNONE to be more
> user-friendly than Windows XP, which leaves Microsoft a winner in no
> department (other than marketing). But what *else* is new?
Another spin (I'm playing Devils advocate here) would be that it
combines the best of both worlds. People make decisions every day and
decisions invariably involve compromise. If your computing goal is to
run benchmarks on your computer all day long then picking the fastest
OS is the right decision. People for whom usability is of paramount
importance would chose a Mac.
|
|
