POV-Ray : Newsgroups : povray.unix : benchmark speed on duron 1.6: gcc vs. icc Server Time
22 Dec 2024 19:31:37 EST (-0500)
  benchmark speed on duron 1.6: gcc vs. icc (Message 1 to 6 of 6)  
From: jhu
Subject: benchmark speed on duron 1.6: gcc vs. icc
Date: 27 Sep 2004 16:10:00
Message: <web.41587349eaa754af89be9050@news.povray.org>
i have some povray benchmarks. they're much faster than my old celeron 800.
they even beat my dad's celeron 2ghz by 33%. what i find interesting is
that the gcc compiled binaries beat the icc binaries. it was vice versa on
my old celeron. in a nutshell, gcc (3.2.x) is better than icc (7.x) for
povray on an athlon-xp core cpu

gcc.386:
Time For Parse: 0 hours 0 minutes 2.0 seconds (2 seconds)
Time For Photon: 0 hours 1 minutes 1.0 seconds (61 seconds)
Time For Trace: 0 hours 40 minutes 49.0 seconds (2449 seconds)
Total Time: 0 hours 41 minutes 52.0 seconds (2512 seconds)

gcc.586:
Time For Parse: 0 hours 0 minutes 2.0 seconds (2 seconds)
Time For Photon: 0 hours 1 minutes 4.0 seconds (64 seconds)
Time For Trace: 0 hours 43 minutes 18.0 seconds (2598 seconds)
Total Time: 0 hours 44 minutes 24.0 seconds (2664 seconds)

gcc.686:
Time For Parse: 0 hours 0 minutes 2.0 seconds (2 seconds)
Time For Photon: 0 hours 1 minutes 1.0 seconds (61 seconds)
Time For Trace: 0 hours 40 minutes 11.0 seconds (2411 seconds)
Total Time: 0 hours 41 minutes 14.0 seconds (2474 seconds)

gcc.athlonxp:
Time For Parse: 0 hours 0 minutes 3.0 seconds (3 seconds)
Time For Photon: 0 hours 0 minutes 60.0 seconds (60 seconds)
Time For Trace: 0 hours 39 minutes 43.0 seconds (2383 seconds)
Total Time: 0 hours 40 minutes 46.0 seconds (2446 seconds)

gcc.athlonxp.nounroll:
Time For Parse: 0 hours 0 minutes 3.0 seconds (3 seconds)
Time For Photon: 0 hours 0 minutes 60.0 seconds (60 seconds)
Time For Trace: 0 hours 41 minutes 41.0 seconds (2501 seconds)
Total Time: 0 hours 42 minutes 44.0 seconds (2564 seconds)

gcc.athlonxp.sse"
Time For Parse: 0 hours 0 minutes 3.0 seconds (3 seconds)
Time For Photon: 0 hours 0 minutes 60.0 seconds (60 seconds)
Time For Trace: 0 hours 40 minutes 46.0 seconds (2446 seconds)
Total Time: 0 hours 41 minutes 49.0 seconds (2509 seconds)

gcc.athlonxp.sse.387:
Time For Parse: 0 hours 0 minutes 3.0 seconds (3 seconds)
Time For Photon: 0 hours 0 minutes 59.0 seconds (59 seconds)
Time For Trace: 0 hours 39 minutes 48.0 seconds (2388 seconds)
Total Time: 0 hours 40 minutes 50.0 seconds (2450 seconds)

icc.586:
Time For Parse: 0 hours 0 minutes 3.0 seconds (3 seconds)
Time For Photon: 0 hours 1 minutes 1.0 seconds (61 seconds)
Time For Trace: 0 hours 42 minutes 13.0 seconds (2533 seconds)
Total Time: 0 hours 43 minutes 17.0 seconds (2597 seconds)

icc.586.ip:
Time For Parse: 0 hours 0 minutes 2.0 seconds (2 seconds)
Time For Photon: 0 hours 1 minutes 1.0 seconds (61 seconds)
Time For Trace: 0 hours 42 minutes 50.0 seconds (2570 seconds)
Total Time: 0 hours 43 minutes 53.0 seconds (2633 seconds)

icc.586.ipo:
Time For Parse: 0 hours 0 minutes 2.0 seconds (2 seconds)
Time For Photon: 0 hours 0 minutes 60.0 seconds (60 seconds)
Time For Trace: 0 hours 41 minutes 42.0 seconds (2502 seconds)
Total Time: 0 hours 42 minutes 44.0 seconds (2564 seconds)

icc.686:
Time For Parse: 0 hours 0 minutes 3.0 seconds (3 seconds)
Time For Photon: 0 hours 1 minutes 7.0 seconds (67 seconds)
Time For Trace: 0 hours 43 minutes 36.0 seconds (2616 seconds)
Total Time: 0 hours 44 minutes 46.0 seconds (2686 seconds)

icc.686.ip:
Time For Parse: 0 hours 0 minutes 2.0 seconds (2 seconds)
Time For Photon: 0 hours 1 minutes 8.0 seconds (68 seconds)
Time For Trace: 0 hours 44 minutes 41.0 seconds (2681 seconds)
Total Time: 0 hours 45 minutes 51.0 seconds (2751 seconds)

icc.686.ipo:
Time For Parse: 0 hours 0 minutes 2.0 seconds (2 seconds)
Time For Photon: 0 hours 1 minutes 6.0 seconds (66 seconds)
Time For Trace: 0 hours 42 minutes 7.0 seconds (2527 seconds)
Total Time: 0 hours 43 minutes 15.0 seconds (2595 seconds)

icc.katmai:
Time For Parse: 0 hours 0 minutes 2.0 seconds (2 seconds)
Time For Photon: 0 hours 1 minutes 7.0 seconds (67 seconds)
Time For Trace: 0 hours 45 minutes 5.0 seconds (2705 seconds)
Total Time: 0 hours 46 minutes 14.0 seconds (2774 seconds)

icc.katmai.ip:
Time For Parse: 0 hours 0 minutes 3.0 seconds (3 seconds)
Time For Photon: 0 hours 1 minutes 5.0 seconds (65 seconds)
Time For Trace: 0 hours 44 minutes 15.0 seconds (2655 seconds)
Total Time: 0 hours 45 minutes 23.0 seconds (2723 seconds)

icc.katmai.ipo:
Time For Parse: 0 hours 0 minutes 3.0 seconds (3 seconds)
Time For Photon: 0 hours 1 minutes 4.0 seconds (64 seconds)
Time For Trace: 0 hours 43 minutes 38.0 seconds (2618 seconds)
Total Time: 0 hours 44 minutes 45.0 seconds (2685 seconds)


Post a reply to this message

From: Thierry CHARLES
Subject: Re: benchmark speed on duron 1.6: gcc vs. icc
Date: 28 Sep 2004 04:56:40
Message: <41592748@news.povray.org>


> i have some povray benchmarks. they're much faster than my old celeron 800.
> they even beat my dad's celeron 2ghz by 33%. what i find interesting is
> that the gcc compiled binaries beat the icc binaries. it was vice versa on
> my old celeron. in a nutshell, gcc (3.2.x) is better than icc (7.x) for
> povray on an athlon-xp core cpu

I think there is something wrong on your dad's pc : a celeron 4 @ 2GHz 
CANNOT be slower than a celeron 2 @ 800 MHz !!!!

On an other hand, you must not compare ICC and GCC on a non-INTEL 
processor : INTEL has no reason to create optimized compiler for 
concurrent processors !

Finally, GCC and ICC have been largely improved since you got your old 
celereon


Thierry


Post a reply to this message

From: Warp
Subject: Re: benchmark speed on duron 1.6: gcc vs. icc
Date: 28 Sep 2004 05:18:33
Message: <41592c69@news.povray.org>
Thierry CHARLES <thi### [at] les-charlesnospamnet> wrote:
> On an other hand, you must not compare ICC and GCC on a non-INTEL 
> processor : INTEL has no reason to create optimized compiler for 
> concurrent processors !

  You are somewhat prejudiced.

  It is well known that optimizing for P4 gives speed advantages also
in AMD processors. The P4-optimized binary might not be the fastest
possible binary which could be compiled for the AMD, but it's clearly
faster than a P3-optimized binary.

  I believe it's not like Intel has no reason to optimize for AMD; it's
like Intel doesn't really care, but their optimizations just happen to
give benefits when run on an AMD as well.

  Thus it's quite fair to compare icc and gcc on an AMD system. The gcc
binaries might usually win, but I think often the icc binaries give a
quite good challenge.

-- 
#macro M(A,N,D,L)plane{-z,-9pigment{mandel L*9translate N color_map{[0rgb x]
[1rgb 9]}scale<D,D*3D>*1e3}rotate y*A*8}#end M(-3<1.206434.28623>70,7)M(
-1<.7438.1795>1,20)M(1<.77595.13699>30,20)M(3<.75923.07145>80,99)// - Warp -


Post a reply to this message

From: Nicolas Calimet
Subject: Re: benchmark speed on duron 1.6: gcc vs. icc
Date: 28 Sep 2004 11:07:11
Message: <41597e1f$1@news.povray.org>
>   Thus it's quite fair to compare icc and gcc on an AMD system. The gcc
> binaries might usually win, but I think often the icc binaries give a
> quite good challenge.

	Let's start spreading a little teaser:

"Hold on.  Keep watching.  Soon in p.benchmarking near to you."

	;-)

	- NC

PS: damn I'm quite late on my schedule.  Hopefully my kind producers
won't go bankrupt because of that...  But hey, I'm my _own_ sponsor
afterall...  [does anyone follows ?]


Post a reply to this message

From: Nicolas Calimet
Subject: Re: benchmark speed on duron 1.6: gcc vs. icc
Date: 28 Sep 2004 11:12:42
Message: <41597f6a$1@news.povray.org>
> i have some povray benchmarks.

	You have to be more specific than that so anybody can
really grasp the meaning of your numbers.

> gcc (3.2.x) is better than icc (7.x) for

	Both are outdated compilers and do not reflect the current
state of nowadays compiler performance.

	I'd suggest you first start by presenting your results
with graphics (if you can post that on some website) and wrap
them with more explanatory text.

	- NC


Post a reply to this message

From: jhu
Subject: Re: benchmark speed on duron 1.6: gcc vs. icc
Date: 28 Sep 2004 13:50:00
Message: <web.4159a42d7161e9a28f9cd9930@news.povray.org>
i was actually referring to my duron 1.6 being faster than both. i'd like to
do a gcc 3.3/3.4 vs. icc 8.x but i don't really have much time now. and by
the time i do get around to it, there will probably more compiler upgrades.

Thierry CHARLES <thi### [at] les-charlesnospamnet> wrote:

>
> > i have some povray benchmarks. they're much faster than my old celeron 800.
> > they even beat my dad's celeron 2ghz by 33%. what i find interesting is
> > that the gcc compiled binaries beat the icc binaries. it was vice versa on
> > my old celeron. in a nutshell, gcc (3.2.x) is better than icc (7.x) for
> > povray on an athlon-xp core cpu
>
> I think there is something wrong on your dad's pc : a celeron 4 @ 2GHz
> CANNOT be slower than a celeron 2 @ 800 MHz !!!!
>
> On an other hand, you must not compare ICC and GCC on a non-INTEL
> processor : INTEL has no reason to create optimized compiler for
> concurrent processors !
>
> Finally, GCC and ICC have been largely improved since you got your old
> celereon
>
>
> Thierry


Post a reply to this message

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