4 Nov
2011
4 Nov
'11
4:46 p.m.
On 11/4/2011 9:23 AM, Cameron Kaiser wrote:
Oh, I'm not one to defend the x86 instruction set (I was merely standing
up for a 386 machine that's old enough to drink). It's certainly a
compromise, but then again that historical compatibility has proved to
be an advantage, just like all good viruses :). So in a certain sense I
can't complain about it too much. (And I don't spend time dealing with
x86 at the assembly level other than occasionally deciphering
disassemblies in the debugger, so honestly the impact it has on my day
to day activities is basically nil.) Would I rather the PC were based
on a more sane instruction set? Sure. Am I losing sleep over it? Nah.
- Josh
Yeah. The
80386 isn't a *terrible* processor on its own (ok, nowhere
near what I'd consider "good", but it's usable) but when hamstrung by
the amateurish, short-sighted, hackish PC systems-level architecture
things get much worse.
I think the real reason I despise the x86 ISA (sorry, Josh)
is because of
all the baggage it carries. It makes CISC look really bad because of the
weight of historical compatibility.
Intel seems to know this based on all the times they've tried to kill x86
off (iAPX 432, i960, etc., even Itanic^WItanium to some degree).