[comp.arch] Multiple-Precision Operations

phipps@garth.UUCP (Clay Phipps) (12/31/88)

In article <1080@l.cc.purdue.edu>,
cik@l.cc.purdue.edu (Herman Rubin) writes:
>I know of no computer reasonably designed 
>for multiple precision operations.
>What is needed is fast unsigned integer arithmetic.  
>... With the CPU such a small price of the computer, 
>get those instructions in.

How fast ?

Both the Intergraph Clipper (32 bits) and the ELXSI 6400 series (64 bits)
seem to have the instructions that you need.
Although multiple-precision operations were considered when the ELXSI
was designed, it seems to be missing an unsigned division instruction.
The Clipper has that instruction, however.

Are there people out on the net who have tried to write code to perform
multiple-precision operations on these machines, and found 
the instruction sets unsatisfactory ?  Were some subtleties overlooked ?

-- 
[The foregoing may or may not represent the position, if any, of my employer]
 
Clay Phipps                       {ingr,pyramid,sri-unix!hplabs}!garth!phipps
Intergraph APD, 2400#4 Geng Road, Palo Alto, CA 93403            415/494-8800

phipps@garth.UUCP (Clay Phipps) (12/31/88)

In article <1080@l.cc.purdue.edu>,
cik@l.cc.purdue.edu (Herman Rubin) writes:
>
>There is need for the software operations 
>which those who understand the use of bit patterns
>find useful to be included in the languages.
>That these operations are not much used is no evidence
>[for their exclusion].

I deal regularly with "bit patterns" as a languages/compilers person, 
but I have never attempted (or wanted) to write code for things like 
multiple-precision arithmetic, FFTs, or data compression.
I assume that those are examples of the bit-pattern dirty work 
that Rubin has in mind.

Now that Rubin has provoked a reaction, here's your chance -- what, 
specifically, would you "find useful" for doing that work in an HLL ?
 
>What is needed [for multiple precision operations]
>is fast unsigned integer arithmetic. 

Don't you need access at the language source-code level to carry-bits 
and things like them, e.g., for multiple-precision operations ? 

[Note that the discussion has been split: into "comp.arch", where it 
originated, and into "comp.lang.misc", for language features only.]
-- 
[The foregoing may or may not represent the position, if any, of my employer]
 
Clay Phipps                       {ingr,pyramid,sri-unix!hplabs}!garth!phipps
Intergraph APD, 2400#4 Geng Road, Palo Alto, CA 93403            415/494-8800