Reference: HW Instructions to Speed Up Lisp? 
Author Message
 Reference: HW Instructions to Speed Up Lisp?

I read a paper once that talked about a small (2?) number of hardware
instructions that, if added to current RISC processors, could be used
to greatly speed up Lisp processing. Something about a fast-trap in
some condition, I vaguely recall. A friend asked me for a copy to help
in a project he is working on, but I can't seem to find it.

I looked through all my old issues of _Lisp and Symbolic Computation_,
_Lisp Pointers_, old LUV proceedings, and a huge stack of Lisp
articles (including Fateman's "Fast Floating Point Processing in
CL"). No luck. Could some kind soul point me to the reference? Or did
I just imagine this?
                                                - Marty
(proclaim '(inline skates))



Sun, 01 Jun 1997 00:36:11 GMT  
 Reference: HW Instructions to Speed Up Lisp?

   I read a paper once that talked about a small (2?) number of hardware
   instructions that, if added to current RISC processors, could be used
   to greatly speed up Lisp processing. Something about a fast-trap in
   some condition, I vaguely recall. A friend asked me for a copy to help
   in a project he is working on, but I can't seem to find it.

        [....]

You might be thinking of a paper written by Doug Johnson of Texas
Instruments. Unfortunately, I can't remember the title, but I think
it's in the proceedings of one of the ASPLOS conferences
(Architectural Support for Programming Languages and Operating
Systems?), around 1990 or so. If I remember correctly, it described
software simulation results of adding a fast trap to a SPARC
processor, and maybe some others.

--
Phil Chu

WWW:      http://kytan.bbn.com/



Sun, 01 Jun 1997 05:39:08 GMT  
 Reference: HW Instructions to Speed Up Lisp?

      I read a paper once that talked about a small (2?) number of hardware
      instructions that, if added to current RISC processors, could be used
      to greatly speed up Lisp processing. Something about a fast-trap in
      some condition, I vaguely recall. ...
   You might be thinking of a paper written by Doug Johnson of Texas
   Instruments. Unfortunately, I can't remember the title, ...

{ Johnson:acm:lfp:1990
, author=   "Douglas Johnson"

, title=    "Trap Architectures for Lisp Systems"
, crossref= "acm:lfp:1990"
, pages=    "79--86"
, refs=     20
, source=   "Dept. Library"
, abstract= "Recent measurements of Lisp systems show a dramatic
skewing of operation frequency.  For example, small integer (fixnum)
arithmetic dominates most programs, but other number types can occur
on almost any operation.  Likewise, few memory references trigger
special handling for garbage collection, but nearly all memory
operations could trigger such special handling.  Systems like SPARC$^{tm}$
and SPUR have shown that small amounts of special hardware can
significantly reduce the need for inline software checks by trapping
when an unusual condition is detected.

A system's trapping architecture now becomes key to performance.  In
most systems, the trap architecture is intended to handle errors
(e.g., address faults) or conditions requiring large amounts of
processing (e.g., page faults).  The requirements for Lisp traps are
quite different.  In particular, the trap frequency is higher,
processing time per trap is shorter, and most need to be handled in
the user's address space and context.

This paper looks at these requirements, evaluates current trap
architectures, and proposes enhancements for meeting those
requirements.  These enhancements increase performance for Lisp
11\%-35\% at a cost of about 1.6\% more CPU logic.  They also aid
debugging in general and speed floating point exception handling."
, reffrom=  Layer:Richardson:cacm:1991

Quote:
}


{ acm:lfp:1990
, title=        "1990 ACM Conference on Lisp and Functional Programming"
, booktitle=    "1990 ACM Conference on Lisp and Functional Programming"
, organization= "ACM"
, publisher=    "ACM Press"
, month=        jun
, year=         1990
, sjb=          "The papers were not formally refereed but were
reviewed by the program committee."
Quote:
}



Tue, 03 Jun 1997 03:27:40 GMT  
 
 [ 3 post ] 

 Relevant Pages 

1. Instruction Speed and instruction availability

2. --- Instruction speed ---

3. --- Instruction speed ---

4. Instruction speeds

5. Instruction Speed

6. pentium instruction clock speeds

7. P4/Athlon instruction speed

8. instruction -> speed

9. Speed of WAM instructions

10. S/390 assembler instruction reference

11. Instruction Set Reference

12. mmx instruction reference

 

 
Powered by phpBB® Forum Software