PanoTools NG:
Re: Re: Processor Speedtest
rogerhoward 2007-Aug-03 05:48:24
On Aug 2, 2007, at 3:34 PM, JD Smith wrote:
> On Thu, 02 Aug 2007 08:33:10 +0000, Hans Nyberg wrote:
>
> >
> > 3 more speedtests on Mac and Windows XP on Mac
> >
> > 1.
> > Model Name: MacBook Pro 17"
> > Model Identifier: MacBookPro1,2
> > Processor Name: Intel Core Duo
> > Processor Speed: 2.16 GHz
> > Speedtest results PTGui 7.2
> > 1st run 1.08
>
> vs.
>
> > Win XP run by Parallels on the same machine
> > Memory set to 1064 mb
> > PTGui 6.03
> >
> > 1st run 2.30
>
> This test seems to counter that claim. Of course, it could be the
> memory
> limits you've imposed, or the older version of PTGui on the Win side.
> Does anyone have bootcamp with the same PTGui version under both
> OS's to
> test?
This doesn't counter my claim at all - you're comparing OSX running
on bare metal, to XP running in a virtual machine on top of OSX. In
that instance I wouldn't expect XP to outperform OSX. The overhead of
virtualizing the disk interface alone would make me avoid such a
comparison.
Of course, I also tend to be of the opinion that the only benchmarks
that matter are those that replicate how someone really works - so
for those running XP in Parallels, that's a very useful benchmark.
I run XP on the exact same hardware... some people say "in" Bootcamp,
but Bootcamp is really just a bootloader. Same version of PTGUI Pro
7. Same exact hardware. Same memory, same CPU, same disks. Of course,
different versions of PanoTools library, that I can't avoid.
Waiting for the sample files so I can run the test. Quite curious
myself! In the meantime, I did a quick test of my own. I tried to
minimize the differences by not using a third-party blender (which I
normally do) - I'll test enblend 3 performance differences in a later
round. I also may well have very different output specs than many - I
almost always output to blended+layers in PSB. I always use 16bit in/
out.
Existing project, created/optimized on another workstation:
6 frames, Sigma 8mm/Canon 20D
16 bit uncompressed TIFF
Output to 16bit 6400x3200 PSB
Blended with Layers, no layer color correction
PTGUI blender
Both instance, followed these steps:
launched PTGUI
Launched batcher from
Added test project to batch
Timed run of batch from second I clicked start, to second where batch
completed end of "Save" step
Tiger 10.4.10
Intel Core Duo 1.83
2GB RAM
Run after restart; logged in with no startup items
OSX: 11:08
XP SP2, all current patches
same machine as above
Run after restart
OSX: 9:48
There are so many variables one can change in a test like this. I
decided to only test the first pass since I only run each project
once anyway - I didn't want OS-level file caching, for instance, to
sway the results. I didn't test optimization (yet). I felt this was a
simple, straightforward test of how I actually work. The difference
isn't HUGE, but it's substantial enough not to be a rounding error.
Am going to try on a Mac Pro later. Will also run the same project on
my G5 (8GB RAM, fast disks) and other systems just for kicks.
--
<*> Wiki: http://wiki.panotools.org/
<*> Nabble (web) http://www.nabble.com/PanoToolsNG-f15658.html
<*> NG Member Map http://www.panomaps.com/ng/
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/PanoToolsNG/
<*> Your email settings:
Individual Email | Traditional
<*> To change settings online go to:
http://groups.yahoo.com/group/PanoToolsNG/join
(Yahoo! ID required)
<*> To change settings via email:
mailto:#removed#
mailto:#removed#
<*> To unsubscribe from this group, send an email to:
#removed#
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/