POV-Ray : Newsgroups : povray.beta-test : Windows console anecdotal slowness : Re: Windows console anecdotal slowness Server Time
26 Sep 2021 12:54:24 EDT (-0400)
  Re: Windows console anecdotal slowness  
From: clipka
Date: 7 Feb 2017 10:37:44
Message: <5899e9c8$1@news.povray.org>
Am 07.02.2017 um 01:50 schrieb dick balaska:
> I upgraded my renderfarm from 3.7.1-alpha to 3.7.1-beta.3
                                ^^^^^^^^^^^    ^^^^^^^^^^^^
Please make sure to only build from tagged versions. Any other versions
may claim to be something different than they actually are (as is
currently the case for the newest code in "release/v3.7.1") or identify
with a hopelessly ambiguous version name (as is the case for many of the
revisions in "master").

Otherwise we are very limited in what support we can provide, as there's
no way we can reliably identify the versions you are referencing.

> and I'm seeing some odd rendertimes.
> ( http://www.buckosoft.com/bsac/meta/
> click on the "Servers" tab)
> Often, my 8-thread windows i7 runs almost twice as fast as my 4-core
> Linux i5.  Minimally, it is about 20% faster.  (The Total FPH column)
> But with this build, I'm seeing an extreme slowdown with the Windows
> console version.  It is running at half the speed of my i5!. (The Cur
> FPH column).

Was the Windows version you used earlier on the i7 also a console
version (which you presumably must have built yourself), or was it an
official GUI version?

In case of the latter, what version of Visual Studio did you use to
build the beta console version?

We are currently aware of speed differences between builds created with
VS2010 and builds created with VS2015 (the latter being faster); the
official alpha builds were typically created using VS2015. Also, the
project settings for the console version may not have the same
optimization settings enabled as the official GUI version, which might
result in the console version being generally slower than the GUI version.

> These numbers can only be viewed as relative to each other.
> Also interestingly, my Ubuntu 16 box is "faster" than its average while
> my other 3 Ubuntu 14 boxes are "slower" than their average.  I would
> expect them to all weight the same against their average.

I don't see what you mean. What I see is "Cur FPH" being lower than
"Total FPH" for all five machines.

Post a reply to this message

Copyright 2003-2021 Persistence of Vision Raytracer Pty. Ltd.