"A month of sundays ago nathan.zook@amd.com wrote:"
[Charset iso-8859-1 unsupported, filtering to ASCII...]
> The bitwise ors allow the equality tests to proceed in parallel, and without
> branch aborts from mispredictions. If the evaluations are short and hard to
> predict, the bitwise version will run faster.
Oh, I see. We are talking about different things. Yes, if all your
tests produce 1 or zero, parallel (bitwise) evaluation of all the
terms in a disjunction might well be faster than (on average)
lazily evaluating half of them in order before coming up trumps.
However, in my opinion that should be left to the optimizer to decide.
Or just possibly C is missing a boolean type.
Peter
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed Feb 23 2000 - 21:00:33 EST