[tremor] [PATCH] 12% global performance gain on a StrongARM

[Nicolas Pitre]

On Thu, 19 Sep 2002, timmy brolin wrote:

> >The attached patch provides a 12% performance gain on a StrongARM SA1110
> 
> > - Split the lookup table into two tables since half of it (one value every 
> >   two) is only used in separate section of the code and only with large 
> >   block sizes.  Therefore the table size used for the common case is reduced
> >   by 2 accounting for yet better cache usage.
> 
> >Nicolas
> 
> It's the worst case that we want to optimize, not the common case. Doesn't
> matter how much you optimize for the most common cases, the CPU will still
> have to be dimensioned to handle the worst case.

Well, I disagree.  Of course the CPU must be able to handle the worst case
and my patch does not perform any worse than the current code in that case.  
However, if you can save memory bandwith and CPU cycles on the common case
then you'll save on power consumption as well which is as important for
portable devices.

> Still, looks good.
> What's the performance difference on the ARM7?

Don't know.  Don't have any ARM7 hardware to test on.

<p>Nicolas

<p>--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'tremor-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.