At 06:27 AM 9/13/2001 +0300, Jarkko Hietaniemi wrote:
> > I think we should use int32_t instead of IV for all code related
> > data. The IV is 64-bit on 64-bit machine, which is significant waste.
>
>I always see this claim ("why would you use 64 bits unless you really
>need them big, they must be such a waste") being bandied around, without
>much hard numbers to support the claims. IV is usually the native
>integer datatype of the CPU which means that it will run fast.
>Unless you are thinking of huge and/or multidimensional arrays
>of tightly packed integers, I don't think you should care.
FWIW, test runs I did on Alpha systems clocked 32-bit ints as faster than
64-bit ones. It was a little surprising, but ultimately ended up being a
bandwidth issue--the time spent doing the extra twiddling to get only the
32-bit portion of the integers was less than the time to get the larger
values in from memory.
Granted, this was on a TurboLaser, which is seriously backplane bandwidth
limited these days, so it's not representative of modern Alpha hardware,
but 32 bits seemed sufficient for our purposes as inlined integer constants
and opcode parameters/numbers/whatever.
Dan
--------------------------------------"it's like this"-------------------
Dan Sugalski even samurai
[EMAIL PROTECTED] have teddy bears and even
teddy bears get drunk
