Talk:Register–memory architecture

New, question on the edit

I'm new here, so let me know if adding stuff that is true but may not be in the article. The information is in RISC article but I'm not sure of the process. — Preceding unsigned comment added by Azurnwiki (talk • contribs) 00:20, 17 June 2016 (UTC)

Change to title

Currently the title is "Register memory architecture". Would the title be better as "Register-memory architecture", or "Register–memory architecture" (to match Load–store architecture)?

104.228.101.152 (talk) 03:54, 23 March 2021 (UTC)

Yes, it would - 1) that's how it's written in the article, 2) that's how Michael Flynn spelled it in the book used as a reference, and 3) it's a compound adjective. I've made the move. Guy Harris (talk) 08:58, 23 March 2021 (UTC)

Is the distinction here "or" and "and"?

"the Motorola 68000 series, which supports integer arithmetic with a memory source or destination, but not with a memory source and destination." WithGLEE (talk) 13:45, 15 April 2023 (UTC)

The 68000 series ADD instruction supports:

• ADD Dn, <ea>, which fetches an argument from a data register and adds it to either a register or memory destination;
• ADD <ea>, Dn, which fetches an argument from either a register or memory source and adds it to a data regiister,;

but does not support ADD <ea>, <ea>, i.e. it doesn't support a memory source and destination. The same applies to the SUB instruction; the MUL and DIV instructions only support a register destination with an arbitrary source.

See, for example, chapter 3 "Instruction set summary" of the MC68020 32-Bit Microprocessor User's Manual.

So the distinction is, indeed, between "or" - which, here, means an "exclusive or", i.e. not both a memory source and destination - and "and", which means both a memory source and destination. Guy Harris (talk) 19:38, 15 April 2023 (UTC)