From caret@pyramid.com  Ukn Jul  4 00:27:39 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA22792; Sun, 4 Jul 93 00:27:37 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA07377; Sat, 3 Jul 93 21:26:59 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA17834; Sat, 3 Jul 93 21:26:48 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA17826; Sat, 3 Jul 93 21:26:46 -0700
From: caret@pyramid.com (Neil Russell)
Message-Id: <9307040426.AA17826@sword.eng.pyramid.com>
Subject: Video on the motherboard
To: riscy@pyramid.com
Date: Sat, 3 Jul 93 21:26:46 PDT
X-Mailer: ELM [version 2.3 PL11]
Sender: riscy-request@pyramid.com
Reply-To: riscy@pyramid.com
Status: RO
X-Status: 

The question is whether to put the video on the motherboard or not.

Many (most?) of the people interested in this board want to make a
nice fast workstation.  Any good workstation has fast, big colourfull
video.  The issue here is just how do you get that glossy video
and how much to pay for it.

Its true that the video will be the most expensive optional part or
the board, coming in at about $100.  I offer some alternatives.

First, video on the board:

	*  1 MB of Video RAM (supports 1024x1024 8-bits per pixel;
	   better resolutions if 1,2 or 4-bits per pixel are used)
		or
	   2 MB of Video RAM (supports 1280x1024 8-bits per pixel)
	*  TI 34076 RAMDAC (has pixel multiplexing, 256 out of 2^16 colors).
	*  NSC 1882 (generates Video sync, and DMA requests.?)

If the peripheral controller is a 3730, then special logic would be
provided to generate the shift register load cycles for the VRAM,
and, of course the main CPU does the video support.  If the peripheral
controller is the 3041, then the VRAM may be connected to either
processor.  If it was connected to the main CPU, the 3041 could still
access the VRAM anyhow, albeit a little slower.  The 3041 could
generate the special VRAM cycles.

Advantages of video on the motherboard:
	*  High bandwidth direct connection to CPU
	*  Linear memory organization (good for software)
Disadvantages:
	*  Adds significant cost to the end price of the motherboard.

For those that don't want the video section, the motherboard could
be sold with empty sockets for the whole video section.  A separate
video add-on package containing the chips to plug in could be put
together.


Video on a VESA or local bus connector:

(I understand that VESA and local bus are the same thing; correct me
 if I'm wrong.)

Advantages of VESA video:
	*  High bandwidth direct connection to CPU
	*  Linear memory organization available
	*  Cards available to do this already
	*  No work for us to design video
Disadvantages of VESA video:
	*  CPU is assumed to be an x86 (hard work to connect to a R3000)
	*  VESA is apparently going to die soon.
	*  VESA and its replacement, PCI, require licenses to use.
	*  Higher total cost than motherboard solution


Video on the ISA bus (assuming there is one):

Advantages of ISA video:
	*  Cheapest video solution
	*  ISA easy to implement
	*  Accelerator cards available that support reasonable speed
	*  Almost everyone already has a ISA video card
	*  Software already exists
	*  No work for us to design video
	*  Cards will be around for a while (ISA will not die soon)
Disadvantages of ISA video:
	*  Still not as fast as direct connect video, even with accelerators
	*  Higher total cost than motherboard solution (if using fast card)
	*  Very complicated interface (VGA)


No video at all:

Advantages:
	*  Cheapest solution of all
Disadvantages:
	*  Must have external terminal
	*  Not as cool!


My personal opinion is that the video should be on the motherboard as
an option.  That is, the video designed and tested, but motherboards
shipped with empty sockets.  I'm one of those that want the really cool
huge colour monitors.


Comments?
-- 
Neil Russell		(The wizard from OZ)
Pyramid Technology			Email:  caret@pyramid.com
3860 N. First Street			Voice:  (408) 428-7302
San Jose, CA 95134-1702			  FAX:  (408) 428-8845

 
From andy@piggy.waldorf-gmbh.de  Ukn Jul  4 12:33:46 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA01698; Sun, 4 Jul 93 12:33:44 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA05854; Sun, 4 Jul 93 09:33:14 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA01186; Sun, 4 Jul 93 09:32:57 -0700
Received: from goss.pyramid.com
	by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA01171; Sun, 4 Jul 93 09:32:54 -0700
Received: from mail.Germany.EU.net 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA05847; Sun, 4 Jul 93 09:32:59 -0700
Received: by mail.Germany.EU.net(EUnetD-2.3.0.e) via EUnet
	id XB28496; Sun, 4 Jul 1993 18:29:55 +0200
Received: from piggy.waldorf-gmbh.de 
	by wegy.waldorf-gmbh.de with UUCP (5.61/GEN-1.0.7)
	via EUnet for unido
	id AA24878; Sun, 4 Jul 93 18:11:16 +0200
From: Andreas Busse <andy@piggy.waldorf-gmbh.de>
Date: Sun, 4 Jul 93 18:50:05 GMT
Message-Id: <9307041850.AA19987@piggy.waldorf-gmbh.de>
Received: by piggy.waldorf-gmbh.de (5.65b/GEN-1.0.7)
	via EUnet for wegy
	id AA19987; Sun, 4 Jul 93 18:50:05 GMT
To: riscy@pyramid.com
Subject: Re: Video on the motherboard
Sender: riscy-request@pyramid.com
Reply-To: riscy@pyramid.com
Status: RO
X-Status: 


> The question is whether to put the video on the motherboard or not.
>
> Many (most?) of the people interested in this board want to make a
> nice fast workstation.  Any good workstation has fast, big colourfull
> video.  The issue here is just how do you get that glossy video
> and how much to pay for it.

I agree. Although the bombs were very small (if not inexistent)
my idea to have no video on board haven't found friends...

> Its true that the video will be the most expensive optional part or
> the board, coming in at about $100.  I offer some alternatives.

Yes, but a mid-class ISA card costs the same.

> First, video on the board:
> 
> 	*  1 MB of Video RAM (supports 1024x1024 8-bits per pixel;
> 	   better resolutions if 1,2 or 4-bits per pixel are used)
> 		or
> 	   2 MB of Video RAM (supports 1280x1024 8-bits per pixel)
> 	*  TI 34076 RAMDAC (has pixel multiplexing, 256 out of 2^16 colors).
>	*  NSC 1882 (generates Video sync, and DMA requests.?)

1024x1024 is a bit unusual :-) Better 1024x768. That wastes ~25% of
the VRAM. 1280x1024 wastes ~33% of the 2 Meg required.
Why not 1152x900 ? That would utilize 99% of 1 Meg VRAM.

> If the peripheral controller is a 3730, then special logic would be
> provided to generate the shift register load cycles for the VRAM,
> and, of course the main CPU does the video support.  If the peripheral
> controller is the 3041, then the VRAM may be connected to either
> processor.  If it was connected to the main CPU, the 3041 could still
> access the VRAM anyhow, albeit a little slower.  The 3041 could
> generate the special VRAM cycles.

Generating the shift register load and clocking out that stuff
could be realized by some GALs.

> Advantages of video on the motherboard:
> 	*  High bandwidth direct connection to CPU
>	*  Linear memory organization (good for software)
> Disadvantages:
>	*  Adds significant cost to the end price of the motherboard.
>
> For those that don't want the video section, the motherboard could
> be sold with empty sockets for the whole video section.  A separate
> video add-on package containing the chips to plug in could be put
> together.

This is a good solution. I like it !

> Video on a VESA or local bus connector:
> 
> (I understand that VESA and local bus are the same thing; correct me
>  if I'm wrong.)

Correct me too, please :-)

> Advantages of VESA video:
> 	*  High bandwidth direct connection to CPU
> 	*  Linear memory organization available
>	*  Cards available to do this already
>	*  No work for us to design video
> Disadvantages of VESA video:
> 	*  CPU is assumed to be an x86 (hard work to connect to a R3000)
> 	*  VESA is apparently going to die soon.
> 	*  VESA and its replacement, PCI, require licenses to use.
> 	*  Higher total cost than motherboard solution

The fact that VESA is on the way to die should be enough
reason not do use VESA video cards.

> Video on the ISA bus (assuming there is one):
> 
> Advantages of ISA video:
> 	*  Cheapest video solution
> 	*  ISA easy to implement
> 	*  Accelerator cards available that support reasonable speed
> 	*  Almost everyone already has a ISA video card
> 	*  Software already exists
> 	*  No work for us to design video
> 	*  Cards will be around for a while (ISA will not die soon)
> Disadvantages of ISA video:
> 	*  Still not as fast as direct connect video, even with accelerators
> 	*  Higher total cost than motherboard solution (if using fast card)
> 	*  Very complicated interface (VGA)

The greatest disadvantage of ISA video cards is loooow speed for
nearly the same price we would pay for *fast* on-board video.

> No video at all:
> 
> Advantages:
> 	*  Cheapest solution of all
> Disadvantages:
> 	*  Must have external terminal
> 	*  Not as cool!

Hmm. No video at all can't be a solution...

> My personal opinion is that the video should be on the motherboard as
> an option.  That is, the video designed and tested, but motherboards
> shipped with empty sockets.  I'm one of those that want the really cool
> huge colour monitors.

I don't need huge colour monitors. I'm satisfied with a 17" monochrome
display. But I don't mind if it's colour :-)


Andy

-------------------------------------------------------------------------------
Andreas Busse             | e-mail: andy@piggy.waldorf-gmbh.de (home)
Eislebenstrasse 5         |         andy@resi.waldorf-gmbh.de (office)
D-5300 Bonn 1             | Phone:  +49 (0)228-252687
-------------------------------------------------------------------------------


 
From tim@ubitrex.mb.ca  Ukn Jul  6 13:11:01 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA06874; Tue, 6 Jul 93 13:10:56 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA24825; Tue, 6 Jul 93 10:10:26 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA14276; Tue, 6 Jul 93 10:10:01 -0700
Received: from goss.pyramid.com
	by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA14215; Tue, 6 Jul 93 10:09:58 -0700
Received: from ubiserver.ubitrex.mb.ca 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA24806; Tue, 6 Jul 93 10:10:02 -0700
Received: from ska.ubitrex.mb.ca ([192.75.16.23]) by ubitrex.mb.ca (4.1/SMI-4.1)
	id AA09602; Tue, 6 Jul 93 12:08:54 CDT
Date: Tue, 6 Jul 93 12:08:54 CDT
From: tim@ubitrex.mb.ca (Tim Braun)
Message-Id: <9307061708.AA09602@ubitrex.mb.ca>
To: riscy@pyramid.com
Subject: Re: Video on the motherboard
Sender: riscy-request@pyramid.com
Reply-To: riscy@pyramid.com
Status: RO
X-Status: 

> The question is whether to put the video on the motherboard or not.

> First, video on the board:

I'd like video on the board.  With the following options:
  1.  640x480 monochrome (VGA compatible) o/p.  Simplest
      anything-should-sync output.
  2.  1024x768 mono SVGA o/p.  Most SVGA monitors should
      sync.
  3.  1152x900 mono.  The mega-pixel display.  Efficient
      use of memory.
  4.  colour - well, I guess someone's going to want this. 
      8-bit colour versions of the above modes.  This requires
      more memory bandwidth (likely VRAM and interface) and
      and a RAMDAC.  Personally, I don't want to spend money
      on these, but I can see the advantages of colour, and
      appreciate that this is desirable.

The nice thing about mono is we can do the support with
the 3730 alone, and the bandwidth requirements are low
enough to do it in main memory, no VRAM required.  You can
still connect your existing (colour or mono) VGA monitor.

I suppose the fact that I'm typing this on an ELC (monochrome
Sun SPARC) "colours" my perception of what's acceptable for 
video.

> 	*  TI 34076 RAMDAC (has pixel multiplexing, 256 out of 2^16 colors).
Hmmm, I think the VGA RAMDAC's would get us where we want to go,
cheaper.  But I could easily be wrong.  Isn't this pixel multiplexing
just the "Hi-Colour" 32K colour technique?

> 	*  NSC 1882 (generates Video sync, and DMA requests.?)
I'm not familiar with the NSC 1882.  Is it like the DP8520/1/2 series?


> If the peripheral controller is a 3730, then special logic would be
> provided to generate the shift register load cycles for the VRAM,
> and, of course the main CPU does the video support.

We definitely want the main R3k to be doing the video pixel bashing (dumb
frame buffer).  The feature mix on the 3730 is really hard to pass up.
 
> For those that don't want the video section, the motherboard could
> be sold with empty sockets for the whole video section.  A separate
> video add-on package containing the chips to plug in could be put
> together.

I'm for this, in some form.

> Video on a VESA or local bus connector:
>
> (I understand that VESA and local bus are the same thing; correct me
>  if I'm wrong.)

Well, you're less wrong today than in January.  There were some proprietary
implementations of a local x86 bus for video used last year, but by 
now most things called "local bus" are VESA compliant.  For our purposes
the terms are equivalent.

VESA is probably not worth the effort for a non-x86 bus.  Getting the
spec's costs money.  

Of course, on the other hand, if we could do simple, cheap video
on the m/board for folks (like me) who can't afford big-time video,
and use VESA to get a 24/32 bit 2 MByte video option for those who
want that, I'd be for it.


> Video on the ISA bus (assuming there is one):

I don't think so.  Of course, if the ISA slot is there,
and someone writes a driver, power to them.  But I wouldn't
want to rely on it for video.

> No video at all:

I don't think we can enough people interested in this configuration.

> My personal opinion is that the video should be on the motherboard as
> an option.  That is, the video designed and tested, but motherboards
> shipped with empty sockets.  I'm one of those that want the really cool
> huge colour monitors.

I can't afford a huge colour monitor.  The empty socket approach has
merits.  I would prefer "everyone gets monochrome, sockets for 8-bit
colour VRAM and RAMDAC on board".
________________________________________________________________
Tim Braun                          |
Ubitrex Corporation                | Voice: 204-942-2992 ext 228
1900-155 Carlton St                | FAX:   204-942-3001
Winnipeg, Manitoba, Canada R3C 3H8 | Email: tim@ubitrex.mb.ca

 
From tim@ubitrex.mb.ca  Ukn Jul  6 13:11:01 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA06874; Tue, 6 Jul 93 13:10:56 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA24825; Tue, 6 Jul 93 10:10:26 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA14276; Tue, 6 Jul 93 10:10:01 -0700
Received: from goss.pyramid.com
	by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA14215; Tue, 6 Jul 93 10:09:58 -0700
Received: from ubiserver.ubitrex.mb.ca 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA24806; Tue, 6 Jul 93 10:10:02 -0700
Received: from ska.ubitrex.mb.ca ([192.75.16.23]) by ubitrex.mb.ca (4.1/SMI-4.1)
	id AA09602; Tue, 6 Jul 93 12:08:54 CDT
Date: Tue, 6 Jul 93 12:08:54 CDT
From: tim@ubitrex.mb.ca (Tim Braun)
Message-Id: <9307061708.AA09602@ubitrex.mb.ca>
To: riscy@pyramid.com
Subject: Re: Video on the motherboard
Sender: riscy-request@pyramid.com
Reply-To: riscy@pyramid.com
Status: RO
X-Status: 

> The question is whether to put the video on the motherboard or not.

> First, video on the board:

I'd like video on the board.  With the following options:
  1.  640x480 monochrome (VGA compatible) o/p.  Simplest
      anything-should-sync output.
  2.  1024x768 mono SVGA o/p.  Most SVGA monitors should
      sync.
  3.  1152x900 mono.  The mega-pixel display.  Efficient
      use of memory.
  4.  colour - well, I guess someone's going to want this. 
      8-bit colour versions of the above modes.  This requires
      more memory bandwidth (likely VRAM and interface) and
      and a RAMDAC.  Personally, I don't want to spend money
      on these, but I can see the advantages of colour, and
      appreciate that this is desirable.

The nice thing about mono is we can do the support with
the 3730 alone, and the bandwidth requirements are low
enough to do it in main memory, no VRAM required.  You can
still connect your existing (colour or mono) VGA monitor.

I suppose the fact that I'm typing this on an ELC (monochrome
Sun SPARC) "colours" my perception of what's acceptable for 
video.

> 	*  TI 34076 RAMDAC (has pixel multiplexing, 256 out of 2^16 colors).
Hmmm, I think the VGA RAMDAC's would get us where we want to go,
cheaper.  But I could easily be wrong.  Isn't this pixel multiplexing
just the "Hi-Colour" 32K colour technique?

> 	*  NSC 1882 (generates Video sync, and DMA requests.?)
I'm not familiar with the NSC 1882.  Is it like the DP8520/1/2 series?


> If the peripheral controller is a 3730, then special logic would be
> provided to generate the shift register load cycles for the VRAM,
> and, of course the main CPU does the video support.

We definitely want the main R3k to be doing the video pixel bashing (dumb
frame buffer).  The feature mix on the 3730 is really hard to pass up.
 
> For those that don't want the video section, the motherboard could
> be sold with empty sockets for the whole video section.  A separate
> video add-on package containing the chips to plug in could be put
> together.

I'm for this, in some form.

> Video on a VESA or local bus connector:
>
> (I understand that VESA and local bus are the same thing; correct me
>  if I'm wrong.)

Well, you're less wrong today than in January.  There were some proprietary
implementations of a local x86 bus for video used last year, but by 
now most things called "local bus" are VESA compliant.  For our purposes
the terms are equivalent.

VESA is probably not worth the effort for a non-x86 bus.  Getting the
spec's costs money.  

Of course, on the other hand, if we could do simple, cheap video
on the m/board for folks (like me) who can't afford big-time video,
and use VESA to get a 24/32 bit 2 MByte video option for those who
want that, I'd be for it.


> Video on the ISA bus (assuming there is one):

I don't think so.  Of course, if the ISA slot is there,
and someone writes a driver, power to them.  But I wouldn't
want to rely on it for video.

> No video at all:

I don't think we can enough people interested in this configuration.

> My personal opinion is that the video should be on the motherboard as
> an option.  That is, the video designed and tested, but motherboards
> shipped with empty sockets.  I'm one of those that want the really cool
> huge colour monitors.

I can't afford a huge colour monitor.  The empty socket approach has
merits.  I would prefer "everyone gets monochrome, sockets for 8-bit
colour VRAM and RAMDAC on board".
________________________________________________________________
Tim Braun                          |
Ubitrex Corporation                | Voice: 204-942-2992 ext 228
1900-155 Carlton St                | FAX:   204-942-3001
Winnipeg, Manitoba, Canada R3C 3H8 | Email: tim@ubitrex.mb.ca

 
