Path: utzoo!mnetor!tmsoft!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!elroy.jpl.nasa.gov!ncar!noao!coyote!jmh
From: jmh@coyote.uucp (John Hughes)
Newsgroups: comp.sys.tandy
Subject: Re: tandy 16B and 16 bit compress
Message-ID: <1990Dec21.212133.7531@coyote.uucp>
Date: 21 Dec 90 21:21:33 GMT
References: <1990Dec12.003253.18391@csis.dit.csiro.au> <1990Dec17.020713.17712@bilver.uucp>
Reply-To: jmh@coyote.UUCP (John Hughes)
Organization: Datalog Consulting, Tucson, AZ
Lines: 41

In article <1990Dec17.020713.17712@bilver.uucp> bill@bilver.uucp (Bill Vermillion) writes:
>In article <1990Dec12.003253.18391@csis.dit.csiro.au> ken@csis.dit.csiro.au (Ken Yap) writes:
>>I was trying to get 16 bit compress working on my 16B (~= 6000) and I
>>found I couldn't have a large array in bss. Nor could I malloc or sbrk
>>any space larger than 245k or so. Now at boot up the machine claims to
>>have 892k (or so) user memory. Where's all that memory?  Do I have to
>>reset some limit somewhere?
>
>Yup.  If you are using 3.2 use the config (or is it conf) to set the max
>memory any process can use.
>
>If you are using one of the older version you need the configure kit, or a
>direct patch, or if you have a copy of filePro use their patch.
>
>I set mine t6k up with about 750 k per/prc.  Be warned though, 16bit
>compress is SLOW and will swap like mad.   I made sure that all the
>compress I did was 12 bit, but had compress set to be able to unpack 16bit.
>I got a 400k news feed one time that was accidentally shipped as 16bit
>compress.   It did about 100k per hour uncompressing.  YIKES!!!


In the process of upgrading and reconfiguring moondog I ran into a problem
similar to Ken's - namely that I couldn't get a compile of compress 4.0
to work in 16-bit mode, even with 2.25 Mb of RAM installed. Using the cfg
utility to up the maxmem size to 1Mb solved the problem, just as Ken
found out (I arrived at my solution at about the same time as the follow-
up with Ken's solution appeared... so I didn't bother to add my two cents
at that time..).

However, I've found that compress is reasonably fast. I routinely tar and
then compress whole directory structures, and I've never seen performance
as bad as 100K per hour (sheesh... now THAT's slow...). If compress is
really running that slow, then perhaps something else is in need of some
serious tweaking.


-- 
|     John M. Hughes      | "...unfolding in consciousness at the            |
| noao!jmh%moondog@coyote | deliberate speed of pondering."  - Daniel Dennet |
| jmh%coyote@noao.edu     |--------------------------------------------------|
| noao!coyote!jmh         | P.O. Box 43305  Tucson, AZ  85733                |