Re: To RISC or not to RISC
- From: Phil Carmody <thefatphil_demunged@xxxxxxxxxxx>
- Date: 31 Oct 2006 10:44:03 +0200
"f0dder" <f0dder.nospam@xxxxxxxxxxxxxxxx> writes:
Some registers have special purpose. SI/DI/SP source/destination/stack are
easy to remember, but having to remember that, for instance, that these were
r0, r3, r5 (arbitrarily chosen here) would be harder.
On other and properly designed hardware with more registers and a more
coherent opcode map, r0-rxx naming would make sense. But on x86, it makes
about as much sense at a/b's weird "line noise" language.
On all register-rich orthogonal architectures that I've played with,
the vendor-supplied development tools' assembler has always used
macros for these special purpose registers (tools ABI-defined, not
necessarily architecturally imposed as they are in x86), even if they
are just r28-r30 in reality. It's a compromise, and it seems to work.
Different ABI's for different OSes in the same architecture could often
use different reservations for those special registers. Best to use
symbolic names with macros than hard coded numeric register values if
any portabilit is desired.
Phil
--
"Home taping is killing big business profits. We left this side blank
so you can help." -- Dead Kennedys, written upon the B-side of tapes of
/In God We Trust, Inc./.
.
- References:
- To RISC or not to RISC
- From: KJH
- Re: To RISC or not to RISC
- From: Herbert Kleebauer
- Re: To RISC or not to RISC
- From: Donkey
- Re: To RISC or not to RISC
- From: Herbert Kleebauer
- Re: To RISC or not to RISC
- From: f0dder
- To RISC or not to RISC
- Prev by Date: Re: A modern view of the halting problem
- Next by Date: Re: Improving one's algorithm
- Previous by thread: Re: To RISC or not to RISC
- Next by thread: Re: To RISC or not to RISC
- Index(es):
Relevant Pages
|
