By: Paul A. Clayton (paaronclayton.delete@this.gmail.com), February 23, 2013 8:09 am
Room: Moderated Discussions
Linus Torvalds (torvalds.delete@this.linux-foundation.org) on February 17, 2013 4:08 pm wrote:
[snip]
> So a single bit gets you far, but there are advantages to multiple bits.
One possible benefit of having multiple condition code bits is for vector operations where one could encode all-match, none-match, one-match, half-or-more-match. Such could facilitate using different paths to handle such cases. (Using a bit vector, population count, and compare would be more flexible and using add-across with true values set to -1 would accomplish the same thing; but using a condition code has some attractions.)
[snip]
> So a single bit gets you far, but there are advantages to multiple bits.
One possible benefit of having multiple condition code bits is for vector operations where one could encode all-match, none-match, one-match, half-or-more-match. Such could facilitate using different paths to handle such cases. (Using a bit vector, population count, and compare would be more flexible and using add-across with true values set to -1 would accomplish the same thing; but using a condition code has some attractions.)