well, they still are somehow often. It's just braindead now to make
*large* tables for *simple* calculation, you might have wanted to say... i
would still use lookups for low-precision integer sine and arctan destined
to animations, like I did in the 386 and 486 days. that still is faster.
The table is small, and pumping in the floating point is expensive. I must
admit that i use a Cyrix... but this can show for purely integer
computations. If you want to get the weight of a byte (sum of all its bit
values) there is no faster way than using a 256-byte lookup, or even a
256-nybble packed lookup, or a 16-byte lookup accessed twice, unless i've
missed some new opcode or trick. The two latter variants are provided if
you don't like the idea of filling the cache.
I'm quite on your side, though, as I won't be advocating for the use of a
65536-entry table of multiplication :-) (and i wasn't smart enough in 1986
to be using one on a 8088 or 6809 -- i was a BASIC guy)
matju
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu