POV-Ray : Newsgroups : povray.unix : GCC 3.1.1 Athlon Optimized Build : Re: GCC 3.1.1 Athlon Optimized Build Server Time
28 Jul 2024 10:27:10 EDT (-0400)
  Re: GCC 3.1.1 Athlon Optimized Build  
From: Ken Cecka
Date: 4 Aug 2002 12:18:25
Message: <3d4d53d0@news.povray.org>
From what I understand, SSE support was not added to the Athlons until the 
MP and XP variants.  I'm pretty sure it's not available in the Thunderbird.  
I've posted my build at http://home.attbi.com/~cecka2/pov35 .  Let me know 
how it works for you.

Ken

Fabien wrote:

> Your flags do not mention SSE instructions. My Duron does not either.
> I sometimes get an error with Micha'compile (which includes SSE
> instructions).
> I'd like to get my hand on your version to give it a try.
> 
> 
> Thanks
> 
> Fabien HENON
> 
> 
> 
> 
> Ken Cecka wrote:
>> Coming in on the heels of everyone else's compiling woes and optimized
>> builds, I just got around to doing my own, and thought I'd post my
>> results. I don't have the official linux build or the windows build
>> installed to do comparisons with, but I did a default build and an
>> optimized build, and have results for each on my system:
>> 
>> povray.unopt +a0.3 +w320 +h240 +v +D +ibenchmark.pov
>> Time For Parse:    0 hours  0 minutes   6.0 seconds (6 seconds)
>> Time For Photon:   0 hours  3 minutes   2.0 seconds (182 seconds)
>> Time For Trace:    1 hours  4 minutes  41.0 seconds (3881 seconds)
>>     Total Time:    1 hours  7 minutes  49.0 seconds (4069 seconds)
>> 
>> povray +a0.3 +w320 +h240 +v +D +ibenchmark.pov
>> Time For Parse:    0 hours  0 minutes   4.0 seconds (4 seconds)
>> Time For Photon:   0 hours  1 minutes  30.0 seconds (90 seconds)
>> Time For Trace:    0 hours 34 minutes  54.0 seconds (2094 seconds)
>>     Total Time:    0 hours 36 minutes  28.0 seconds (2188 seconds)
>> 
>> Quite a significant improvement!  OK, so this sucks compared to you
>> Athlon XP 1900 people, but I'm quite happy with it on my Athlon tbird 1.1
>> GHz. Some notes on the build:
>> - Hadn't read the thread on the min/max problem, so I ended up changing
>> all calls to min/max to fmin/fmax, and calls to min3/max3 to
>> fmin(fmin(),)/fmax(fmax(),).
>> - Configured as follows based on random suggestions scattered across
>> usenet: CXX='/usr/local/gcc-3.1.1/bin/g++ -O3 -march=athlon-tbird
>> -mcpu=athlon-tbird -fomit-frame-pointer -falign-functions=6
>> -funroll-loops -ffast-math -malign-double -m3dnow -mmmx'
>> LDFLAGS='-static' ./configure
>> 
>> Note that I linked this as a static binary so if anyone else wants to
>> give it a try, I can put it up for download - just let me know.
>> 
>> Ken
>> 
>> Oh, for reference:
>> 
>> $ cat /proc/cpuinfo
>> processor       : 0
>> vendor_id       : AuthenticAMD
>> cpu family      : 6
>> model           : 4
>> model name      : AMD Athlon(tm) Processor
>> stepping        : 2
>> cpu MHz         : 1109.921
>> cache size      : 256 KB
>> fdiv_bug        : no
>> hlt_bug         : no
>> f00f_bug        : no
>> coma_bug        : no
>> fpu             : yes
>> fpu_exception   : yes
>> cpuid level     : 1
>> wp              : yes
>> flags           : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca
>> cmov pat pse36 mmx fxsr syscall mmxext 3dnowext 3dnow
>> bogomips        : 2215.11
>> 
>> $ cat /proc/meminfo
>>         total:    used:    free:  shared: buffers:  cached:
>> Mem:  261648384 247451648 14196736  1597440 33746944 153714688
>> Swap: 388575232 92811264 295763968
>> MemTotal:       255516 kB
>> MemFree:         13864 kB
>> MemShared:        1560 kB
>> Buffers:         32956 kB
>> Cached:         132836 kB
>> SwapCached:      17276 kB
>> Active:         110336 kB
>> Inact_dirty:     70904 kB
>> Inact_clean:      3388 kB
>> Inact_target:        4 kB
>> HighTotal:           0 kB
>> HighFree:            0 kB
>> LowTotal:       255516 kB
>> LowFree:         13864 kB
>> SwapTotal:      379468 kB
>> SwapFree:       288832 kB
>> NrSwapPages:     72208 pages
>>


Post a reply to this message

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