baum@Apple.COM (Allen J. Baum) (02/20/91)
[]
Is the new swordfish compatible with anything, like the old 320xx, or
is it a completely new instruction set architecture?
--
baum@apple.com (408)974-3385
{decwrl,hplabs}!amdahl!apple!baumgideon@taux01.nsc.com ( Gideon Intrater) (02/21/91)
In article <49343@apple.Apple.COM> baum@apple.UUCP (Allen Baum) writes: >[] >Is the new swordfish compatible with anything, like the old 320xx, or >is it a completely new instruction set architecture? > >-- > baum@apple.com (408)974-3385 >{decwrl,hplabs}!amdahl!apple!baum The Swordfish is assembly level compatible with other members of the 32K family. It is not binary compatible. It uses a superset of the 32K programming model (32 general purpose registers vs. 8 in the 32K, 3-Operand instructions vs. 2-Operands etc.). The Swordfish assembler can generate Swordfish code from a 32K assembly source file by 1-to-1 or 1-to-many translation. We target the Swordfish for the embedded control market, where most of the users has the source files of their applications. This is of course not the case in the general work station market. Binary competability is not required by most of our customers. Making an instruction issue logic for a superscalar CISC instruction set is very complex. We decided to reduce this complexity, and go for a pure RISC architecture which still preserves the 32K programming model. -- Gideon Intrater gideon@nsc.nsc.com National Semiconductor, P.O.Box 3007, Herzlia B 46104, Israel Phone: +972-52-522255, Fax: +972-52-558322
wolfe@vw.ece.cmu.edu (Andrew Wolfe) (02/21/91)
The article I have in front of me (Electronic Buyers News - don't ask) describes Swordfish: "the 0.8 micron CMOS part is a powerful 64-bit superscalar RISC extension to National's NS32000 family." Your guess as to what this means is as good as mine.