Re: [Sbcl-help] Re: READ-LINE Performance Question

[Marco A. <ma...@cs...>]

As an aside,  on the issue of READ-LINE I would really like to have a

	READ-BUFFERED-LINE line-buffer &optional stream eof-error-p eof-value 
recursive-p

that behaved like READ-LINE, but that would fill LINE-BUFFER (this 
being an adjustable character array with a fill pointer - the fill 
pointer being reset to 0 on each call)

Using READ-SEQUENCE for the intended use does not quite cut it IMHO.

Of course the performance issue should be dealt with at the 
implementation level.

Cheers

Marco





On Jan 25, 2005, at 9:00 PM, Pete Kazmier wrote:

> Peter Denno <pet...@ni...> writes:
>>
>> I'm not an expert in the internals of SBCL, so I can't comment on
>> most of this, but I see that you are running under SLIME. I have a
>> LispWorks application that runs twice as fast standalone as it does
>> under SLIME.
>
> Thanks for the suggestion, although in this case, it did not have a
> significant impact on the times.  Since my original post, I have
> tested reading the same log file in both Allegro and LW (per Edi's
> suggestion in a thread on the cl-ppcre list) to see how their
> READ-LINE speeds compare to the free CL implementations [1].  Both
> were faster of course, but LW was by far the fastest and on par with
> the equivalent python program.  Of course, life would be much better
> if SBCLs READ-LINE was much faster.  Any developers have any input on
> my previous post?
>
> Thanks,
> Pete
>
> [1] Speeds from LW and Allegro
>
> LispWorks:
> CL-USER 17 > (time (tester "/tmp/pgw-logs/trapd.log.ovnyc00p"))
> Timing the evaluation of (TESTER "/tmp/pgw-logs/trapd.log.ovnyc00p")
>
> user time    =      0.880
> system time  =      0.180
> Elapsed time =   0:00:01
> Allocation   = 36822608 bytes standard / 4070 bytes conses
> 0 Page faults
> Calls to %EVAL    34
> T
>
> Allegro:
> CL-USER(7): (time (tester "/tmp/pgw-logs/trapd.log.ovnyc00p"))
> ; cpu time (non-gc) 2,730 msec user, 90 msec system
> ; cpu time (gc)     80 msec user, 0 msec system
> ; cpu time (total)  2,810 msec user, 90 msec system
> ; real time  3,493 msec
> ; space allocation:
> ;  67 cons cells, 67,825,256 other bytes, 0 static bytes
> T
>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
> Tool for open source databases. Create drag-&-drop reports. Save time
> by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
> Download a FREE copy at http://www.intelliview.com/go/osdn_nl
> _______________________________________________
> Sbcl-help mailing list
> Sbc...@li...
> https://lists.sourceforge.net/lists/listinfo/sbcl-help
>
--
Marco Antoniotti					http://bioinformatics.nyu.edu
NYU Courant Bioinformatics Group		tel. +1 - 212 - 998 3488
715 Broadway 10th FL				fax. +1 - 212 - 998 3484
New York, NY, 10003, U.S.A.