|
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
I recently compiled the unix-version source on a Win95 platform using
the free GCC-compiler for windows cdk, available from www.cygnus.com.
The libjpeg and zlib compiled just fine, as did the povray-sources.
I tried rendering skyvase.pov at 640 480 with an aa setting of 0.2 (if
I remember correctly), The new version completed in 2 minutes and 6
seconds. The old pov-ray binary took 2 minutes and 20 seconds.
That's about 10% faster, or 10% "less slow" depending on how you look
at it.
Not a lot, but if you're doing a 5000-frame animation it might be
benificial.
--
At the top of the food chain.
Post a reply to this message
|
|
| |
| |
|
|
From: William D Hayden
Subject: Re: Improving the speed of POV-Ray 3.02 on Win95
Date: 15 Jun 1998 23:10:03
Message: <3585E20B.5C3C@computek.net>
|
|
|
| |
| |
|
|
>
> I recently compiled the unix-version source on a Win95 platform using
> the free GCC-compiler for windows cdk, available from www.cygnus.com.
> The libjpeg and zlib compiled just fine, as did the povray-sources.
>
> I tried rendering skyvase.pov at 640 480 with an aa setting of 0.2 (if
> I remember correctly), The new version completed in 2 minutes and 6
> seconds. The old pov-ray binary took 2 minutes and 20 seconds.
>
> That's about 10% faster, or 10% "less slow" depending on how you look
> at it.
>
> Not a lot, but if you're doing a 5000-frame animation it might be
> benificial.
And a commercial, native compiler is even faster. GCC is a known
performance dog, but then it never claimed to produce fast code. It has
the advantage of being free, and available on a wide number of
platforms. I ALWAYS compile POV-Ray when I get it for my Solaris box.
I noticed a 20-30% speed increase over their GCC compiled release.
Of course for Windows, I think they use Watcom, one of the better X86
compilers. If your comparing the Windows version versus a command line
version, merly get Windows out of the way will help tremendously. If
you want to speed it up even more, pick up a Borland or Watcom compiler
and recompile it again. A bare bones version of each is available for
under $100 US. Pricey, for some, cheap for the guy doing "a 5000-frame
animation".
--
William D. Hayden
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Hmmm... I have a W95 K6/233 32MB system.
skyvase.pov 640x400 AA 0.2 took it 6m 17s with the v3.02 watcom .exe file.
At "no AA" it took 3m 49s
> I tried rendering skyvase.pov at 640 480 with an aa setting of 0.2 (if
> I remember correctly), The new version completed in 2 minutes and 6
> seconds. The old pov-ray binary took 2 minutes and 20 seconds.
--
Alan
Spam Bait:
root@localhost
admin@localhost
current board of the Federal Communications Commission:
Chairman Reed Hundt: rhu### [at] fccgov
Commissioner James Quello: jqu### [at] fccgov
Commissioner Susan Ness: sne### [at] fccgov
Commissioner Rachelle Chong: rch### [at] fccgov
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
[yan### [at] pacbellnet]
| Hmmm... I have a W95 K6/233 32MB system.
| skyvase.pov 640x400 AA 0.2 took it 6m 17s with the v3.02 watcom .exe file.
| At "no AA" it took 3m 49s
Funny. I have a Pentium II 233, 64MB System. This is the part where I
lecture you about the K6's FPU performance, but I won't. :)
--
At the top of the food chain.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
[William D. Hayden <wdh### [at] computeknet>]
| And a commercial, native compiler is even faster. GCC is a known
| performance dog, but then it never claimed to produce fast code.
True. That's the price for being multiplatform I suppose.
| platforms. I ALWAYS compile POV-Ray when I get it for my Solaris box.
| I noticed a 20-30% speed increase over their GCC compiled release.
Impressive. SUN makes the best platform for POV-Ray use, at least
that's my experience. IRIX boxes are strangely slow and can easilly be
beat by an INTEL at the same MHz.
I'd love to see a POV-Ray version with K6-2 3DNOW support. It should
really fly.
| Of course for Windows, I think they use Watcom, one of the better X86
| compilers. If your comparing the Windows version versus a command line
| version, merly get Windows out of the way will help tremendously.
Actually, I compared the DOS protected mode-version to the one I
compiled. An unfair advantage to GCC, I know, but the dos-version
actually outperforms the Win95 version from the POV-Team. I've never
cared for the editor that goes with it. I use Emacs for my
scene-editor, and now that I have the unix-command line version I can
work in the same way no matter what platform I'm on.
| If you want to speed it up even more, pick up a Borland or Watcom compiler
| and recompile it again.
I'm considering it.
--
At the top of the food chain.
Post a reply to this message
|
|
| |
| |
|
|
From: Mark Arrasmith
Subject: Re: Improving the speed of POV-Ray 3.02 on Win95
Date: 16 Jun 1998 14:13:52
Message: <6m6ckj$ee3$1@oz.aussie.org>
|
|
|
| |
| |
|
|
Actually I see the same thing. I run two compiles for AlphaNT. A port of
the Unix command-line version of 3.02 and my compile of POV-Win 3.01. The
command-line version runs in about 25% less time than POV-Win. On the same
system and both under Windows NT 4.0 SP3.
Mark Arrasmith
Post a reply to this message
|
|
| |
| |
|
|
From: Mark Arrasmith
Subject: Re: Improving the speed of POV-Ray 3.02 on Win95
Date: 16 Jun 1998 14:18:15
Message: <6m6csr$eep$1@oz.aussie.org>
|
|
|
| |
| |
|
|
Well before you start to argue about speed make sure you can even compare.
Copy SkyVase.pov to a temporary directory. Create a file called skyvase.ini
and put the following in it . . .
-i SkyVase.pov
+v1
-d
+ft
-x
+a0.300
+r3
-q9
-w640
-h480
-mv2.0
-b1000
This is the official speed test. Run it and then post your times.
Mark
>[yan### [at] pacbellnet]
>| Hmmm... I have a W95 K6/233 32MB system.
>| skyvase.pov 640x400 AA 0.2 took it 6m 17s with the v3.02 watcom .exe
file.
>| At "no AA" it took 3m 49s
>
>Funny. I have a Pentium II 233, 64MB System. This is the part where I
>lecture you about the K6's FPU performance, but I won't. :)
>
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Mark Arrasmith" <arr### [at] mathtwsuedu> wrote:
>Actually I see the same thing. I run two compiles for AlphaNT. A port of
>the Unix command-line version of 3.02 and my compile of POV-Win 3.01. The
>command-line version runs in about 25% less time than POV-Win. On the same
>system and both under Windows NT 4.0 SP3.
Did you set the render priority correctly ? And were any other apps running ?
Was display on ? All of these things affect speed under Windows. Check in the
help file.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On Mon, 15 Jun 1998 22:10:03 -0500, "William D. Hayden"
<wdh### [at] computeknet> wrote:
>Of course for Windows, I think they use Watcom, one of the better X86
>compilers. If your comparing the Windows version versus a command line
>version, merly get Windows out of the way will help tremendously. If
>you want to speed it up even more, pick up a Borland or Watcom compiler
>and recompile it again. A bare bones version of each is available for
>under $100 US. Pricey, for some, cheap for the guy doing "a 5000-frame
>animation".
Surprise! I compiled POVRay for Windows with MS VC++ 5.0 SP3 today,
from the unmodified 3.01 sources, and it beat my official 3.02 by 10%
on skyvase, in two different resolutions. The official benchmark
resolution took 193 seconds on the official build, and 174 on my
build, both running on a PII-233 with 64MB of memory and NT 5 Preview
Beta 1.
Post a reply to this message
|
|
| |
| |
|
|
From: Mark Arrasmith
Subject: Re: Improving the speed of POV-Ray 3.02 on Win95
Date: 17 Jun 1998 00:01:06
Message: <6m7f45$fuq$1@oz.aussie.org>
|
|
|
| |
| |
|
|
>Did you set the render priority correctly ?
Yes
>And were any other apps running ?
No
>Was display on ?
No
>All of these things affect speed under Windows.
I know.
The command line compile was from someone else (using VC++ Risc Edition).
The POV-Win compile was mine (again with VC++ Risc Edition). The problem
might be with my settings on the compiler more than anything else. I'll
tinker with that alittle.
Mark
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |