POV-Ray : Newsgroups : povray.unix : Unix ? Server Time
29 Jul 2024 00:37:39 EDT (-0400)
  Unix ? (Message 11 to 13 of 13)  
<<< Previous 10 Messages Goto Initial 10 Messages
From: Miguel Alvarez Rodriguez
Subject: Re: Unix ?
Date: 19 Jun 1999 19:03:05
Message: <376C226A.49686973@abonados.cplus.es>
"Jon A. Cruz" wrote:

> Chris C wrote:

>

> > Nieminen Mika <war### [at] cctutfi> wrote:

> >

> > >  I wonder how is this done. Unless unix overclocks the cpu without telling

> > >you... :)

> >

> > It's quite obvious, if you think about it. Any application on a modern OS must

> > share the CPU with - at the very least - the OS, if not other applications. In

> > addition, the application will typically depend on the OS for various services.

> >

> > The more efficient the OS is in keeping out of the application's way, and the

> > more efficiently coded the services it offers are, the more time the

> > application has to do its work.

> >

> > Windows 95/98 are woeful in that respect - they still have chunks of 16-bit

> > code in them, which causes a contect switch every time they're called from

> > 32-bit mode.

> >

> > Windows NT - while a true 32/64-bit OS - still has more overhead than modern

> > unixes.

> >

> > So, a good Unix such as FreeBSD or Linux will typically run (compiler

> > optimisations not considered) the same code on the same hardware faster than

> > Windows.

> >

> > -- Chris

>

> And another factor is how memory and drive space are handled. NTFS was originally

> hyped as not ever needing defragmentation. NT5 is going to include a third-party

> developed defragmenter. Hmmm. Convey anything about their file system design?

>

> Also, the swapping in NT is horrible. Well, OK maybe not that bad, but it can be a

> factor. Especially with a program like PovRay.

>

> And Win9X is another big problem. Any 16-bit process executing will block all

> 32-bit processes until it complets. In windows multimedia development, I've had

> simple test cases where a timer callback for MIDI playing could get a 2100ms

> latency! 2.1 seconds of non-callback just because I happened to be accessing

> ethernet at the time. Ouch!


Hi. I've compared the speeds. On NT and 9x are exactly the same. And with a well
configured kernel, Linux is about 40% faster than NT, and with a bad configured
kernel, it's about 25% faster.


Post a reply to this message

From: Ron Parker
Subject: Re: Unix ?
Date: 21 Jun 1999 09:08:52
Message: <376e3964@news.povray.org>
On Sun, 20 Jun 1999 01:06:18 +0200, Miguel Alvarez Rodriguez wrote:
>IkpvbiBBLiBDcnV6IiB3cm90ZToNCg0KPiBDaHJpcyBDIHdyb3RlOg0KPg0KPiA+IE5pZW1p
>bmVuIE1pa2EgPHdhcnBAY2MudHV0LmZpPiB3cm90ZToNCj4gPg0KPiA+ID4gIEkgd29uZGVy
>IGhvdyBpcyB0aGlzIGRvbmUuIFVubGVzcyB1bml4IG92ZXJjbG9ja3MgdGhlIGNwdSB3aXRo

(several dozen more lines of this junk follow)

Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: base64

Please fix your newsreader.  No real newsreader would be stupid  enough 
to post plain text in base64.


Post a reply to this message

From: Miguel Alvarez R 
Subject: Re: Unix ?
Date: 21 Jun 1999 20:16:37
Message: <376ED6AE.57108011@abonados.cplus.es>
Ron Parker wrote:
> 
> On Sun, 20 Jun 1999 01:06:18 +0200, Miguel Alvarez Rodriguez wrote:
> >IkpvbiBBLiBDcnV6IiB3cm90ZToNCg0KPiBDaHJpcyBDIHdyb3RlOg0KPg0KPiA+IE5pZW1p
> >bmVuIE1pa2EgPHdhcnBAY2MudHV0LmZpPiB3cm90ZToNCj4gPg0KPiA+ID4gIEkgd29uZGVy
> >IGhvdyBpcyB0aGlzIGRvbmUuIFVubGVzcyB1bml4IG92ZXJjbG9ja3MgdGhlIGNwdSB3aXRo
> 
> (several dozen more lines of this junk follow)
> 
> Content-Type: text/plain; charset=iso-8859-1
> Content-Transfer-Encoding: base64
> 
> Please fix your newsreader.  No real newsreader would be stupid  enough
> to post plain text in base64.

Sorry about that, Netscape could read it.

   I've compared the speeds. On NT and 9x are exactly the same. And with
a well
configured kernel, Linux is about 40% faster than NT, and with a bad
configured
kernel, it's about 25% faster.


Post a reply to this message

<<< Previous 10 Messages Goto Initial 10 Messages

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.