[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: inline array refs?



In article <poeck-130594113955@wi6a65.informatik.uni-wuerzburg.de> poeck@informatik.uni-wuerzburg.de (Karsten Poeck) writes:

   Path: info-server.bbn.com!noc.near.net!paperboy.wellfleet.com!news-feed-1.peachnet.edu!news.duke.edu!convex!cs.utexas.edu!howland.reston.ans.net!EU.net!Germany.EU.net!netmbx.de!zrz.TU-Berlin.DE!zib-berlin.de!fauern!winx03!wi6a65.informatik.uni-wuerzburg.de!user
   From: poeck@informatik.uni-wuerzburg.de (Karsten Poeck)
   Newsgroups: comp.lang.lisp.mcl
   Followup-To: comp.lang.lisp.mcl
   Date: 13 May 1994 10:14:21 GMT
   Organization: University of Wuerzburg
   Lines: 56
   NNTP-Posting-Host: wi6a65

   I translated some code to test iterative repair alogirithms from c to LISP.
   The code mainly does fixnum operations and simple-array accesses. By adding
   appropriate (the fixnum ...) and (declare (type (simple-array fixnum (*))
   array) 
   I managed to get about half the speed of the corresponding c program
   compiled with MPW on the same mac.

I don't know about MCL yet, but in Lucid and Allegro, you should be able to
get the speed to be the same as C, perhaps better.  I am looking for pairs
of comparable real programs in Lisp and C, so i'd be glad to help you with
performance issues if your code could be come one of my benchmarks.

   The disassembly of the code shows calls to subroutines
   (jsr_subprim $SP-GETLONG) and
   (jsr_subprim $SP-MKLONG)

   I suspect that these calls slow down the programm. Does anybody know what
   these routines do and how I could avoid them? I always thought that adding
   enough declarations to a MCL program would give me the same speed as a
   corresponding c program.

   Anyhow I tried to run the same program under LIspWorks 3.2 on a Dec-Alpha
   and to my surprise a quadra 650 was more than two times faster than the
   alpha, even after declaring fixnum-safety = 0. From looking at the
   disassembly I suspect that the arefs are not inlined in LispWorks. Does
   anybody know how I could get LispWorks to inline the arefs too?

   ;test code for the array refs

   (defun test (array index)
     (declare (optimize (speed 3)(safety 0))
	      (type (simple-array fixnum (*)) array))
     (aref array index))

Unless you need a full 32 bits i recommend using simple-vector's instead.
Although (simple-array fixnum (*)) seems to be what corresponds to int
A[100]; in C, I think it should only be used when full 32 bit integers are
required.

   ->
   0 (SPUSH VSP)
   2 (VPUSH D1)
   4 (VPUSH D0)
   6 (MOVE.L (VSP 4) ATEMP0)
   10 (MOVE.L @VSP D1)
   12 (ASR.L 1 D1)
   14 (MOVE.L (ATEMP0 D1.L 7) D0)
   18 (SPOP VSP)
   20 (JMP_SUBPRIM $SP-MKLONG)

This looks like a jump to a primative.  I don't know if MCL can do any
better.

k
--
Ken Anderson 
Internet: kanderson@bbn.com
BBN STC              Work Phone: 617-873-3160
10 Moulton St.       Home Phone: 617-643-0157
Mail Stop 6/4c              FAX: 617-873-3776
Cambridge MA 02138
USA