wineport: Use a lookup table to make our slow ffs() a bit faster.
amartinson at codeweavers.com
Fri Mar 11 12:16:11 CST 2011
On 03/11/2011 07:19 AM, Alexandre Julliard wrote:
> Henri Verbeet<hverbeet at gmail.com> writes:
>> Something like the following may be faster:
>> static const int bit_pos =
>> 0x00, 0x01, 0x0e, 0x02, 0x0f, 0x1b, 0x12, 0x03,
>> 0x1e, 0x10, 0x1c, 0x16, 0x18, 0x13, 0x0a, 0x04,
>> 0x1f, 0x0d, 0x1a, 0x11, 0x1d, 0x15, 0x17, 0x09,
>> 0x0c, 0x19, 0x14, 0x08, 0x0b, 0x07, 0x06, 0x05,
>> const unsigned int c = 0x7dc4d65;
>> return bit_pos[((x& -x) * c)>> 27];
>> I'd expect that mostly due to eliminating the branches and the
>> implicit shift& mask operations in something like "y.c".
> Maybe, but there's no point. This is just a fallback implementation for
> broken platforms, any platform that we care about should never use this.
For systems with fast multiplication that way is probably faster. This
was just something I noticed that took me 5 mins and was clearly better
than what we had.
More information about the wine-devel