| Re: [AD] Color convertors |
[ Thread Index |
Date Index
| More lists.liballeg.org/allegro-developers Archives
]
> Yes it's the way to go. Even Intel recommends it! xorl executes in 1 cycle
> and is fully pairable. movl executes in 1 cycle but can only be placed in
> the U pipe if it involves another register than %eax, since it's longer
> than 6 bytes. xorl is also shorter (2 or 3 bytes), which reduces the
> pressure on the instruction cache.
Hmm... I thought it was exactly the opposite: movl was fully pairable on
Pentium for any register whereas xorl can only be placed in the V pipe, and
'xorl reg,reg' vs 'movl $0,reg' was a well-known optimization for 80386
because of the instruction length but no longer valid. When I wrote the
non-MMX code, I extensively timed the routines on my P200 (disabling the
write instructions) and all the routines ran as if they were fully paired.
Weird.
> Allegro.txt?
Section 'Windows specifics', DirectX windowed driver.
> The 15->16 one works under Windows (for me at least)
They work but I'll disable them.
> So under Windows, if the window has the same depth as the screen, then the
> screen -is- the window? or does it still pass via a second buffer like the
> other depths?
Sort of a mix of both ;-) The backbuffer is in VRAM, the blits are
hardware-accelerated.
> Already on CVS :)
Yes, but as my local copy is crippled with various non-commited patches and
yours spans several files, it's not very easy.
---
Eric Botcazou
ebotcazou@xxxxxxxxxx