Path: utzoo!utgpu!utstat!jarvis.csri.toronto.edu!mailrus!cornell!uw-beaver!tektronix!tekcrl!tekgvs!toma
From: toma@tekgvs.LABS.TEK.COM (Tom Almy)
Newsgroups: comp.sys.ibm.pc
Subject: Re: 16 bit cards vs 8, SHAFTED?
Message-ID: <4719@tekgvs.LABS.TEK.COM>
Date: 3 Mar 89 15:46:47 GMT
References: <37020@tut.cis.ohio-state.edu> <14313@bigtex.cactus.org> <3778@peora.ccur.com> <4710@tekgvs.LABS.TEK.COM> <14505@bigtex.cactus.org>
Reply-To: toma@tekgvs.LABS.TEK.COM (Tom Almy)
Organization: Tektronix, Inc., Beaverton,  OR.
Lines: 85

In article <14505@bigtex.cactus.org> james@bigtex.cactus.org (James Van Artsdalen) writes:
>In <4710@tekgvs.LABS.TEK.COM>, toma@tekgvs.LABS.TEK.COM (Tom Almy) wrote:
<< In article <3778@peora.ccur.com< joel@peora.UUCP writes:
<< <Another thing I'm wondering about is whether a 16 bit card card will
<< <actually faster than 8 bit cards for most cases. 
<
<< Well, if you are going through the BIOS, your won't see any improvement.
<< I have made the following observations:
<
<How so?  VGA cards have their own BIOS.  I assume those vendors code
<their BIOS to use their hardware properly.  The code in the
<motherboard BIOS isn't generally used.  The video card BIOS is certain
<to use 16 bit moves to scroll the screen.

Yes the BIOS is possibly faster, but the problem is that the BIOS routines
are constrained to operate like IBM's BIOS, which was not designed for
efficiency.  For instance, you cannot write a string of characters, but
must call the BIOS for each character, enduring the overhead on each one. 
(Yes, the VGA BIOS has a string output function, but this is not available
on earlier adapters)  If you want to draw a line, there is no line drawing
primitive, nor is there area fill or area copy (you can read or write only
single pixels).  The BIOSes also scroll the display by moving all characters
rather than reprogramming the display start address: this is much faster,
FANSI-CONSOLE does it, but is incompatible with some application programs.
Really fast programs don't use the BIOS except for setting display modes.

<
<< 3) Most text applications are not display speed limited, and thus won't show
<< much speed improvement by the clock, but they sure look lots faster since a
<< page of text snaps instantly to the screen.
<
<"snaps instantly to the screen"?  "won't show much speed improvement"?
<
<If you can see it, it's definitely an improvement.  If you can't, it
<usually isn't.  Whether or not something is an improvement from the
<CPU's perspective is entirely irrelevant: on the user's perspective
<matters.
<
Yes and no.  I think we are in agreement here -- people want the programs
to appear fast, and a 2x faster display, even in a 90% compute bound program
looks more than 5% faster!  But by the clock (which is what you see in
magazine reviews), it will only be 5% faster, so people might believe it
is not worth the additional cost.

<< 4) Speed improvement in graphics modes is less since you cannot do 16 bit
<< graphics manipulations.
<
<Oh?  I've been doing it for years.  If you store sixteen bits to the
<screen, you get that bit pattern displayed.  

Well, no.
The binary value 0000000110000000
displays as
.              .
which is not what I would expect!

<You just have to store the right sixteen bits.

Well, thats the problem, isn't it?  If you put "1" in AX, and write it to
the screen, shift left and write, etc., you see this:

       .		Time 1
      .			Time 2
     .			Time 3, etc.
    .
   .
  .
 .
.
               .
              .
             .
            .
           .
          .
         .
        .


EGA/VGA cards were designed for 68000 processors, being big endian in their
bit ordering.  

Tom Almy
toma@tekgvs.labs.tek.com
Standard Disclaimers Apply