RE: [clisp-list] RedHat 7.1 and CLISP 2.27

[Justice, R. -CONT(DYN) <Ran...@cn...>]

I will try to make an RPM..  I have not made one, but I have read the
instructions a few times..

I'm am running, "compiled code" in Windows 2k, RH 7 and RH7.1.   One a
different note, 
I have seen great improvement using "Compile-file" 

"> My general problem is that CLISP is running very slow. My CLISP
> program is timed based - It runs an algorithm for certain periods of
> time then exits. The program measures runtime not clock time.  It
> takes about 3X times longer to run the program.

compared to what?"

Looking at the "total execution time" in Win2K - Task Manager CPU TIME, and
RH 7.0 - "top",  compared to the expected/reasonable amount of completion.
For example, a test case is scheduled for about 4 hours 
of "cpu/runtime.",  and the machine will take a little over 4 hours of
"total execution time" - the total elapsed time might be 5-6 hours depending
on how "busy" the system. On RH7.1 the "total execution time" is not in
align with expected/reasonable amount of completion.  It's looks like 
the CLISP is constantly being preempted by something, except the "top"
command don't show anything
else being executed--except for the "normal" system stuff. 

I will look at "CLOCC/src/tools/monitor/metering.lisp".

Any Ideals?

Randy
 


-----Original Message-----
From: Sam Steingold [mailto:sd...@gn...]
Sent: Thursday, August 30, 2001 8:34 AM
To: Justice, Randy -CONT(DYN)
Cc: cli...@li...
Subject: Re: [clisp-list] RedHat 7.1 and CLISP 2.27


> * In message
<B4C...@pe...>
> * On the subject of "[clisp-list] RedHat 7.1 and CLISP 2.27"
> * Sent on Thu, 30 Aug 2001 07:29:02 -0500
> * Honorable "Justice, Randy -CONT\(DYN\)" <Ran...@cn...>
writes:
>
> I installed RedHat 7.1, full installation, and CLISP 2.27.  Both
> installations seem to went well without any problems.

can you make an RPM?

> My general problem is that CLISP is running very slow. My CLISP
> program is timed based - It runs an algorithm for certain periods of
> time then exits. The program measures runtime not clock time.  It
> takes about 3X times longer to run the program.

compared to what?

> I changed the "priority" of the lisp executable from "0" to "-10".
this is not likely to help.

> Can anyone help?  Could it be a bug in CLISP 2.27?  Could it be a bug
> in RedHat?  Could it be a RedHat Admin problem?  What demons don't I
> need?  I'm clueless on a solution.  Should I down grade to RedHat 7.0?
> Should I down grade CLISP?  FYI the RedHat system is non GUI.

> I know my CLISP program works well in Windows 2000, and RedHat 7.0.     

1. please make sure that you are running compiled code on both win2k,
   rh7 and rh7.1

2. if you still observe the performance hit, get
      CLOCC/src/tools/monitor/metering.lisp
   and find out which part of the program is hit.



-- 
Sam Steingold (http://www.podval.org/~sds)
Support Israel's right to defend herself!
<http://www.i-charity.com/go/israel>
Read what the Arab leaders say to their people on <http://www.memri.org/>
Software is like sex: it's better when it's free.