Re: faster representation of pow on some áreas of subsurface?
torvalds at linux-foundation.org
Sun Jan 19 12:00:13 UTC 2014
On Sun, Jan 19, 2014 at 11:29 AM, Tomaz Canabrava <tcanabrava at kde.org> wrote:
> I'v replaced the pow() call with the fastPow() that I found on the internet
> and it got *way* snappyer.
Yeah, no, that doesn't work. The error is too big, and it only works
on little-endian anyway. Ok, so we probably don't care about anything
else, but it's fragile.
That said, what the heck happened to my patch that caches the value of
"pow()" in add_segment()? That gets the math right *and* it sped
things up quite a bit.
I'll search for my patch and re-send.
More information about the subsurface