Re: [AD] About Elias' bug

["Elias Pschernig" <eliaspschernig@xxxxxxxxxx>]

> I've attached a patch which does that. But I'm not very pleased with
it:
> - it adds a large amount of code,
> - some very weird things happen: in normal conversion mode, it runs
20%
> slower than the current 3.9.37 CVS code on my machine, although it
uses the
> same code with one less test per pixel (but supposedly crunched by the
loop
> optimizer).

I did some more timing with your patch, and now I understand the weird
timings i got yesterday: It seems 32->16 switches to dithering without
the KEEP flag.

And I compiled the patch with MSVC, there the results are even more
reasonable:

MSVC 24->15:
CVS:             350 blits/s
patched, normal: 340 blits/s
patched, keep:   300 blits/s

MING 24->15:
CVS:             450 blits/s a lot faster than MSVC :)
patched, normal: 260 blits/s this seems to be the GCC optimization
thing..
patched, keep:   220 blits/s

MING 32->16:
CVS:             450 blits/s
patched, normal: 150 blits/s that's what I noticed yesterday :)
patched, keep:   220 blits/s

Elias