nu.kanga.list.mud-dev

25912: Re: [MUD-Dev] Is database access a bottleneck?

On Sat, Dec 14, 2002 at 03:01:50PM -0500, Amanda Walker wrote:
> On 12/13/02 5:47 PM, bradley newton haug <brad@faithanddisease.com> wrote:
 
>> I even have hashtables with calculation/result lookups that I
>> load at startup, to save a few cycles in exchange for
>> memory/slower startup.
 
> Heh.  I'm old enough to remember when you had to do that for, say,
> multiplication ;-).

> Ah, those were the days... when a 8MHz Z80H gave people the same
> rush as a 3GHz P4 does now...  Love that Moore's Law.

These days it's quite frequently a better bet to simply do the
calculation, at least if the hashtable is big enough that it will
mean a cache miss most of the time, whereas the calculation doesn't
touch memory much.  Raw CPU speed has increased a /lot/ faster than
memory speed, and these days the root of most optimisation is
reducing accesses to RAM and improving locality of reference.

--

	Jo

_______________________________________________
MUD-Dev mailing list
MUD-Dev@kanga.nu
https://www.kanga.nu/lists/listinfo/mud-dev