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

KCL on a SPARCstation 1



A big hello to everyone on the KCL mailing list from a new kcl user,
and special thanks to Ron Olphie for his welcome message.

I saw on the KCL mail archive some questions about availability of
KCL on SPARC machines. As a matter of fact, I have just installed
successfully AKCL (KCL + the austin modifications) on my SUN 
SPARCstation 1 running SunOS 4.0.3. I ftped the compressed tar files
kcl.tar.Z and akcl-1-222.tar.Z from rascal.ics.utexas.edu, untarred
them, and followed the instructions of the README file from akcl. 

Everything went very smoothly, and I was able to run the gabriel 
benchmarks (also ftped from rascal.ics.utexas.edu) a few hours later.


Here are the benchmark results on the SPARCstation 1 (fast indeed!) :


       For kcl Common Lisp
Fri Oct 13 15:32:20 MET 1989

BOYER        3.600
BROWSE       5.583
CTAK         1.667
DDERIV       1.967
DERIV        1.183
DESTRU-MOD   0.408
DESTRU       0.525
DIV2         2.700
FFT-MOD      1.400
FFT          192.450
FPRINT       0.362
FREAD        0.233
FRPOLY       44.483
PUZZLE-MOD   2.100
PUZZLE       2.283
STAK         0.988
TAK-MOD      0.817
TAK          1.183
TAKL         0.350
TAKR         0.142
TPRINT       0.496
TRAVERSE     8.600
TRIANG-MOD   18.617
TRIANG       23.100

However, I should warn you that KCL alone (without the austin mods) 
does not run on SPARC chips, as there are about 30 lines of assembly
code without a SPARC version. A quick look into akcl showed that at 
least a part of this assembly code has been replaced by C macros to
be expanded inline. Not only it is portable, but a comment says it 
is also faster!!

In my opinion, the austin mods are very worthwhile as they provide for
a faster, more portable and bug-free KCL. (I checked many of the bugs
signalled in kcl-mail-archive, and none remained in my akcl version). 

I want to thank all the nice souls out there (Yuasa, Hagiya ,Schelter)
for providing us such a nice and professionnal implementation of 
Common Lisp. Bravo!

I am still looking for an implementation of CLOS running under KCL. 
Please let me know if one is available for ftp somewhere.

(bye)  gregory