A naive compiler will branch when generating (remainder255 >= (255 +
128)); that's a comparison, and comparisons generate branches.
A smarter compiler might have a clever way of avoiding the branch,
perhaps involving a subtract followed by a cntlzw (?). That's just off
the top of my head; I haven't tested it.
If "alpha" is a constant across the whole image operation, you could
multiply it by a cleverly-chosen constant so that "(alpha * multiplier)
>> something" gives you numbers scaled to any range you prefer. This is
a trick I've used in the past with good results. (The constant tends to
be something wacky; I remember getting 0x8102 or 0x10203 for various
operations in the past.) It only has one caveat; some PowerPCs can
multiply tiny numbers (i.e. 8-bit values) a little faster than big
numbers. I have no idea if this is ancient history or if G4s and G5s
still have this restriction. Still, I'd rather pay an extra two cycles
on the multiply than add extra instructions to the inner loop.
On Oct 21, 2004, at 6:20 PM, Keith Bauer wrote:
Wow, did I ever get hammered on that little "optimization". For
those watching at home, the optimization Chris is talking about is
probably this one:
tmp = alpha * red;
remainder256 = tmp & 0xFF; // get alpha * red (mod 256)
dividend = tmp >> 8; // compute alpha * red / 256
remainder255 = dividend + remainder256; // compute alpha * red (mod
255)
dividend += ((remainder255 >= (255 + 128)) & 1) + ((remainder255 >=
128) & 1); // add 0, 1, 2 depending on how large the remainder (mod
255) is
This is provably correct and even rounds correctly. And yes, it
would be much, much easier in AltiVec.
No, there are faster ways of doing it, without branches.
Chris
Er, that code doesn't have any branches in it that I can see...
And if you know such optimizations, why not contribute to the thread
and enlighten us by posting them rather than just sounding superior
about your knowledge?
Thanks,
Keith
_______________________________________________
Do not post admin requests to the list. They will be ignored.
PerfOptimization-dev mailing list
(email@hidden)
Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/perfoptimization-dev/
email@hidden
