Return to BSD News archive
Xref: sserve comp.unix.sysv386:24258 comp.windows.x:45686 comp.os.linux:11225 comp.unix.bsd:5803 comp.os.mach:2224 comp.sys.ibm.pc.misc:21970 comp.sys.ibm.pc.hardware:32318
Newsgroups: comp.unix.sysv386,comp.windows.x,comp.os.linux,comp.unix.bsd,comp.os.mach,comp.sys.ibm.pc.misc,comp.sys.ibm.pc.hardware
Path: sserve!manuel!munnari.oz.au!spool.mu.edu!think.com!ames!decwrl!csus.edu!netcom.com!hasty
From: hasty@netcom.com (Amancio Hasty Jr)
Subject: Re: Free software and the future of support for Diamond products
Message-ID: <4m+p5fd.hasty@netcom.com>
Date: Tue, 29 Sep 92 14:27:04 GMT
Organization: Netcom - Online Communication Services (408 241-9760 guest)
References: <7s0nqk-.hasty@netcom.com> <bill.717458450@chaos.cs.umn.edu> <t00nqa=.hasty@netcom.com>
Keywords: Diamond, free-software
Lines: 121
Hi,
First off my apologies for not getting to some of you earlier.
My primary reason was to acquire further information from
S3 before sending my e-mail response.
My end-goal is to acquire an affordable high end graphics svga
which we can program and that the board can be purchased and supported.
Ideally, the card should cost below $200 (US) mark.
In dealing with both Diamond and Orchid, we have run into
a stumbling block as far as the ability to set the dotted clock
values in the S3 boards due to the use of custom clock synthetizers.
Both Diamond and Orchid, so far are not releasing the information
required to set the dotted clock values. In order to use
the higher resolutions available with svga monitors and to
increase the performance of graphics operations, the functionality
to set the dotted clock values is required.
When I first approach S3 on friday, they inform me that they would
release the schematic diagrams for their 8C911. Today they
claimed that they only release the information for the their
DRAM version of the chipset. They promised they mail me the
schematics and the neccessary information on the algorithm
on how to initialize the S3 chipset, which also includes
information on to set the dotted clock values for boards
without clock synthetizers.
The S3 set is a combination of an 8514/A and SVGA chipset.
Basically, all the control functions needed to set the
S3 are provided in the form of VGA registers. All enhanced
graphics functions such as bitblt follow very closely the
IBM 8514/A register convention. Under DOS and after
initializing the board using INT10, I have been able to
do biblt operations and line drawings using the enhanced
register set using Orchid's Farenheit 1280 (S3 based svga card).
The test code was taken right out of the
"Graphics Programming for the 8514/A" by Jake Richter &
Bud Smith, with some minor modifications to set the DAC
registers and the ability to access the S3 enhanced chipsets.
In vga land, you have to unlock the register set in
order for read or write access. In terms of how may
colors, depending on the DAC used it can support 24
bit colors, or 16 million colors.
The new upcoming 928 is going to be about 2 to 3 times
faster than the 8C911, far faster than the 8514/A.
According to the customer-marketing person which I spoke,
they believe that at this point the easiest road to achieve
programmability and availability of an S3 card is to oem
the board through a company. Artix located in Sunnyvale,
California was recommended.
I was not able to contact Artix today. The talks with
Artix will begin tomorrow, and for sure I will buy a board
from them as a test run.
On the way of hardware engineer we have to wait for the
package from S3 corporation to see what the actual contents are.
I expect the package to be in around Wednesday.
The actual cost of building an S3 8C911 is about $130-$150 (US)
depending on value. The cost of including the AMI bios is about $25,0000.
So provided that we do go ahead with the hardware project,
it will help alot if someone could recommend, an alternative
to the AMI BIOS. Also, there are no ASICs besides the S3 chipset
that are required to build the S3 svga card.
As for availability of a programmable S3 svga card, more on this
tomorrow nite after I get to test the board. I will summarize
my findings and e-mail them on Wednesday along with the
information provided by S3.
Why the S3?
(1) The chipset is cheap, and vga boards are becoming available
below the $250 mark. Also, it is supposed to be easy to
implement an S3 vga card. To my knowledge there is no other
acceptable solution at this time. There maybe others
but they are either unproven technology or they will
only release information on how to program the chipset
under a non-disclosure agreement.
(2) The S3 was designed with local bus in mind, and there are
existing VESA local bus implementations as well as for other
local bus implemantations.
(3) Because the S3 command register set is based upon the
8514/a and svga register set, there is a wide choice
of available literature.
(4) S3 corportation provides a software development kit
(5) Although limited, S3 allows access to their engineers
and marketing.
(6) With the 928 there is a future performance growth path.
Even if I should obtained a reasonable S3 svga card, I feel
that it is worthwhile for netters to have the ability to
produce svga cards based on the S3 chipset. The reason
is that as the ease of creating high-end graphics cards
increases I strongly suspect that manufacturers will
resort to idiosyncransies as a way to differentiate themselves.
Thank you for support,
Amancio Hasty