[MPlayer-dev-eng] [PATCH] Re : Re : Direct3D vo performance increase

[Jim Hauxwell]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


> No, you failed to read till the end, where it says (yes, they use a
> ridiculously overcomplicated way to explain it):

> There should be at most one
> IDirect3DDevice9::BeginScene/IDirect3DDevice9::EndScene pair between any
> successive calls to present



You have read it wrong.  The key word here is 'should'.  This indicates
an optimal amount.  The only failure case with this call is if you call
begin, followed by another begin.  The D3D runtime would easily be able
to trap more begin/end pairs before a present, but it doesn't.

If you are familiar with opengl then an endscene is like a glFlush().
It pushes the currently collected state to the GPU pipe for rendering
(which is why msdn mentions about having the largest possible gap
between endscene and present (which is like swapbuffers).


Its kind of a moot point as it got taken out anyway.  I just wanted to
clarify.



Jim
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkkpJasACgkQhrNWoHjgI1A2kwCeLOasFnO4mJnSBC6Tjep0g+rP
X2UAoL/WCtG9IfRH73uoSdrdhQX/bQ9K
=lhFp
-----END PGP SIGNATURE-----