*BSD News Article 15983


Return to BSD News archive

Xref: sserve comp.windows.x.i386unix:1519 comp.os.386bsd.questions:2383
Newsgroups: comp.windows.x.i386unix,comp.os.386bsd.questions
Path: sserve!newshost.anu.edu.au!munnari.oz.au!constellation!osuunx.ucc.okstate.edu!moe.ksu.ksu.edu!crcnis1.unl.edu!wupost!howland.reston.ans.net!agate!ames!pacbell.com!att-out!cbnewsj!dwex
From: dwex@mtgzfs3.att.com (David E. Wexelblat)
Subject: Re: XFree1-2 + 386BSD performance
Organization: AT&T Bell Laboratories
Date: Wed, 12 May 1993 16:01:08 GMT
Message-ID: <C6x8Hy.Fw@cbnewsj.cb.att.com>
References: <1993May12.025731.29769@latcs1.lat.oz.au>
Sender: news@cbnewsj.cb.att.com (NetNews Administrator)
Nntp-Posting-Host: mtgz058.gaz.att.com
Lines: 64

In article <1993May12.025731.29769@latcs1.lat.oz.au> wongm@ipc5.lat.oz.au (M.C. Wong) writes:
> Hi,
>   I have a question about the performanc of running XFree1.2 on
> 386BSD 0.1. In multi-user mode running 3 xterms, xeyes, xclock,
> xload and xbiff, when I do serious compiling in one of the window,
> I realize that the performance of the overall X activities are
> intolerable slow and it chokes very very much, and I can hardly
> see my mouse moving/appearing on the screen for 5-10 seconds.
> 
>   I wonder :
> 
>   1) does the system resource limit (set by limit from tcsh) has
>      great effect on the performance ?

No (I think).

>   2) I am running on 486/33 with 16M of RAM, is this considered
>      the expected performance ? Besides I am running on a local
>      bus board.

Unfortunately, yes.  With an SVGA board (non-accelerated), the CPU must
move all of the video memory around.  There's a LOT to move, and hence
this is a serious CPU hog.  Conversely, when the server gets knocked out
of the CPU by a CPU-bound task, the server essentially stops.

This, specifically, is what accelerated hardware will deal with.  Hardware
cursor, BitBlt, etc, will have the most visible impact on things like this.

>   3) has anyone tried fine-tuning the performance of 386BSD + XFree
>      such that it runs better ? If so, can u please post some
>      useful guidelines, as I think many of us will like to know.

'nice'ing the server up and your compiles down MAY help; I don't know.
Nothing except accelerated hardware (and servers upport for same) will
help much.

>   4) Is the choking a pure limitation of the 486 CPU power, or
>      is it the size of memory ? If I go on to > 16M, say 20M, 
>      will that gives me better performance ?

No, it's a function of the CPU-intensive nature of moving the data around.
Hence there's not much that can be done about it.

>   5) Does the size of the swap space affect that ? If so, how can
>      I increase the swap space size ?

Probably not.

> 
>   Many thanks in advance!
> 
> -- 
> - wongm@latcs1.lat.oz.au


--
David Wexelblat <dwex@mtgzfs3.att.com>  (908) 957-5871  Fax: (908) 957-5627
AT&T Bell Laboratories, 200 Laurel Ave - 3F-428, Middletown, NJ  07748

XFree86 requests should be addressed to <xfree86@physics.su.oz.au>

"How many times must good men die?  How many tears will the children cry,
 'til we suffer no more sadness?  Oh, stop the madness.  Stop all the madness."
        -- Molly Hatchet, Fall Of The Peacemakers.