Nick Warne wrote:
> On Friday 25 May 2007 03:59:55 Laurence Vanek wrote:
>
>>> ./configure CXXFLAGS=-O3 ...
>>>
>>> If you wish to include further flags, they need to enclosed in double
>>> quotes:
>>>
>>> ./configure CXXFLAGS="-march=athlon-tbird -O3"
>>>
>>> This is my full FG configure option:
On Friday 25 May 2007 03:59:55 Laurence Vanek wrote:
> > ./configure CXXFLAGS=-O3 ...
> >
> > If you wish to include further flags, they need to enclosed in double
> > quotes:
> >
> > ./configure CXXFLAGS="-march=athlon-tbird -O3"
> >
> > This is my full FG configure option:
> >
> > ./configure CX
Nick Warne wrote:
> Hi Georg,
>
> On Wednesday 23 May 2007 19:05:43 Georg Vollnhals wrote:
>
>> But where do you apply this "-O3" parameter? Configure or Make when
>> compiling the FlightGear sources? Thank you very much in advance for your
>> help.
>>
>> Regards
>> Georg EDDW
>
> Sorry, I read
Nick Warne schrieb:
> If you read the questions when using 'cmake -i .' one of them is what type of
> build to do:
>
> ...Debug | Release ...
>
> etc. Default is NONE. So you set that to 'Release'.
>
> Then later there is a question about what CXX flags the compiler should use
> when making a
Nick Warne schrieb:
> Sorry, I read this too quick. You can build SG/FG with optimisations like
> this:
>
> ./configure CXXFLAGS=-O3 ...
>
> If you wish to include further flags, they need to enclosed in double quotes:
>
> ./configure CXXFLAGS="-march=athlon-tbird -O3"
>
> This is my full FG conf
Hi Georg,
On Wednesday 23 May 2007 19:05:43 Georg Vollnhals wrote:
> But where do you apply this "-O3" parameter? Configure or Make when
> compiling the FlightGear sources? Thank you very much in advance for your
> help.
>
> Regards
> Georg EDDW
Sorry, I read this too quick. You can build SG/FG
Hi George,
On Wednesday 23 May 2007 19:05:43 Georg Vollnhals wrote:
> >> cmake -i .
>
> Thank you for the hint, this "cmake -i" configuration process works
> very nice for me but does not improve the framerates.
>
> > Oh my. After building OSG with 'Release' and then building FG with -O3,
> >
On 5/23/07, Georg Vollnhals <[EMAIL PROTECTED]> wrote:
>> Once you have a fresh checkout, in the trunk directory (top layer of
code)
>> to build the CMAKE files, issue:
>>
>> cmake -i .
Thank you for the hint, this "cmake -i" configuration process works
very nice for me but does not improv
Hi Nick,
>> I was a bit confused with this - I couldn't work out what 'the
>> configuration' screen was, nor meant.
>>
>> But I sussed out CMAKE now, so for anybody else with problems.
>>
>> Once you have a fresh checkout, in the trunk directory (top layer of code)
>> to build the CMAKE files, iss
On Tuesday 22 May 2007 17:48:02 Nick Warne wrote:
> On Monday 21 May 2007 23:46:34 gh.robin wrote:
> > > Configure osg with ist's new build system as a Release version. Or
> > > provide CMAKE_CXX_FLAGS and CMAKE_C_FLAGS on the command line of cmake.
> > > I run cmake with something like
> > >
> > >
On Monday 21 May 2007 23:46:34 gh.robin wrote:
> > Configure osg with ist's new build system as a Release version. Or
> > provide CMAKE_CXX_FLAGS and CMAKE_C_FLAGS on the command line of cmake.
> > I run cmake with something like
> >
> > cmake -D CMAKE_CXX_FLAGS="-O3 " \
> > -D CMAKE_C_FLA
Hi,
On Monday 21 May 2007, Tim Moore wrote:
> In case it's not obvious how to do that (I sure missed it!), in the
> ccmake configuration screen change CMAKE_BUILD_TYPE to "Release".
That is also possible ...
Greetings
Mathias
---
On Mon 21 May 2007 19:46, Mathias Fröhlich wrote:
> Hi,
>
> On Saturday 19 May 2007, gh.robin wrote:
> > I try to understand the big FPS differences i get when i build FG
> > first with an old OSG library 10 april 2007,
> > second with a recent OSG library 1.9.4 version
>
> I believe that you buil
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Mathias Fröhlich wrote:
> Hi,
>
> On Saturday 19 May 2007, gh.robin wrote:
>> I try to understand the big FPS differences i get when i build FG
>> first with an old OSG library 10 april 2007,
>> second with a recent OSG library 1.9.4 version
> I beli
Hi Tim,
Just tried this patch and had a couple observations (beyond noticing that it
generally works pretty well.) :-)
1. The application continuously scrolls the following message:
Viewer::checkWindowStatus() - 1
I couldn't immediately see where that was coming from.
2. The numeric keypad ke
Mathias Fröhlich schrieb:
> Hi,
>
> On Saturday 19 May 2007, gh.robin wrote:
>
>> I try to understand the big FPS differences i get when i build FG
>> first with an old OSG library 10 april 2007,
>> second with a recent OSG library 1.9.4 version
>>
> I believe that you built osg without op
Hi,
On Saturday 19 May 2007, gh.robin wrote:
> I try to understand the big FPS differences i get when i build FG
> first with an old OSG library 10 april 2007,
> second with a recent OSG library 1.9.4 version
I believe that you built osg without optimization.
When building without optimization y
Hello Tim, Mathias,
This is a digest of tests i have done , (and a repeat of previous separated
mails)
I try to understand the big FPS differences i get when i build FG
first with an old OSG library 10 april 2007,
second with a recent OSG library 1.9.4 version
FG was patched with Tim Moore pa
Hi,
On Friday 18 May 2007, Tim Moore wrote:
> renderer.cxx already contains a lot of OSG specific code; in fact it
> would be fair to say that is all OSG specific code. I did add some
> osgViewer specific code to renderer.cxx because the details of accessing
> the scene root, controlling the came
On Tuesday 15 May 2007, Tim Moore wrote:
> Howdy,
> This patch implements the option of using OpenSceneGraph's osgViewer
> instead of SDL or glut. The major user visible difference is the
> availability of OSG statistics, as seen in
> http://www.bricoworks.com/moore/osgstats.png, which show the tim
On Thu 17 May 2007 22:28, Tim Moore wrote:
SNIP
>
> > The "Cull" is basically very high compared to the other values but when
> > i fly over the sea (without tiles as i said in an other topic).
> > What is exactly the meaning of the "Cull" value ?
>
> It's hard to assign a meaning to these times i
On Thu 17 May 2007 22:28, Tim Moore wrote:
SNIP
> It's hard to assign a meaning to these times in isolation, except to
> note that 16 milliseconds total is the magic number that gives you a
> frame rate of 60hz. The large cull time indicates poor scene graph
> layout; possibly there's a problem wi
* Tim Moore -- 5/18/2007 9:20 AM:
> renderer.cxx already contains a lot of OSG specific code; in fact it
> would be fair to say that is all OSG specific code.
I agree. Main/fg_os*.cxx are there to handle the operating system dependent
interfaces for window management/keyboard/mouse: glut and later
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Curtis Olson wrote:
> Hi Tim,
>
> I took a peek at the diffs and had a couple questions. Originally the
> idea of the fg_os files was to have a single interface within the
> flightgear code so that the details could be hidden in fg_os.[ch]xx, but
> I
Hi Tim,
I took a peek at the diffs and had a couple questions. Originally the idea
of the fg_os files was to have a single interface within the flightgear code
so that the details could be hidden in fg_os.[ch]xx, but I see with your
changes that you have a lot of OSG specific code now that has f
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
gh.robin wrote:
> Hello Tim,
>
> I spite of a very low fps with osg 1.9.4 (as i said before) i have built FG
> with that patch, which works perfectly (but the cursor change, you are
> working on it).
>
I'll turn this into a Wiki page if it gets t
gh.robin wrote:
>Hello Tim,
>
>I spite of a very low fps with osg 1.9.4 (as i said before) i have built FG
>with that patch, which works perfectly (but the cursor change, you are
>working on it).
>
>How may we understand the values which are given
>Event
>Update
>Cull
>Draw
>Gpu
>
>Are they pe
On Tue 15 May 2007 00:32, Tim Moore wrote:
> Howdy,
> This patch implements the option of using OpenSceneGraph's osgViewer
> instead of SDL or glut. The major user visible difference is the
> availability of OSG statistics, as seen in
> http://www.bricoworks.com/moore/osgstats.png, which show the t
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Howdy,
This patch implements the option of using OpenSceneGraph's osgViewer
instead of SDL or glut. The major user visible difference is the
availability of OSG statistics, as seen in
http://www.bricoworks.com/moore/osgstats.png, which show the time sp
29 matches
Mail list logo
