[FFmpeg-devel] [PATCH] Optimization of original IFF codec

[Sebastian Vater]

Hi Michael!

Michael Niedermayer a ?crit :
> On Mon, Apr 26, 2010 at 12:19:28AM +0200, Sebastian Vater wrote:
>   
>> When I understand you right, I have to create a lookup table the
>> following way:
>> For each of the 4-pair read bits:
>> {0000 = 0, 0001 = 1 << plane, 0010 = 0x100 << plane, 0011 = (1 << plane)
>> | (0x100 << plane), 0100 = (0x10000 << plane), ...}
>>
>> Is that correct?
>>     
>
> yes, with endianness correct
>
>   
I noticed this yesterday, because the image got garbarged with big
endianness lut ;)

Have you an idea how I can do this with 24bps? I'm thinking of using 3
luts...one for 0 <= bps < 8, one for 8 <= bps < 16 and one for 16 <= bps
< 24.

But I'm afraid that this will cause some of the data/code to be accessed
in L2 cache only instead of L1 which it does now (which would also
explain the huuuuuge improvement with inline keyword).

-- 

Best regards,
                   :-) Basty/CDGS (-: