[comp.sys.transputer] PS/2 transputer boards

K312240@AEARN.BITNET (Klaus Kusche) (08/17/88)

Dear Mailing List!

There was a question regarding transputer boards for IBM PS/2
(MicroChannel) systems on this list a few weeks ago. I met somebody of
the IBM lab Yorktown heights two weeks ago (I was asked not to give his
name) who said that a US company doing PC boards definitely will have a
PS/2 board *really soon now*.

Unfortunately, he said that he is not yet allowed to tell me the
companie's name. Sigh!
(CSA, Microway, Definicon, ...?)

Hope that helps and causes the corresponding company to speak up on this
list!

What's the most important aspect to me: Will the host interface on that
PS/2 board be faster (DMA!) than the damned software polling solution on
all the PC boards now???

Klaus Kusche (K312240@AEARN.BITNET).

chris@nth.UUCP (Chris Chauvin) (08/25/88)

In article <8808161709.AA20575@tcgould.TN.CORNELL.EDU>, K312240@AEARN.BITNET (Klaus Kusche) writes:
> 
> There was a question regarding transputer boards for IBM PS/2
> (MicroChannel) systems on this list a few weeks ago. I met somebody of
> the IBM lab Yorktown heights two weeks ago (I was asked not to give his
> name) who said that a US company doing PC boards definitely will have a
> PS/2 board *really soon now*.
> 
> Hope that helps and causes the corresponding company to speak up on this
> list!

Well, I don't think we're the mystery company, but the company I work 
for -- Nth Graphics, Ltd. -- is introducing a single-transputer (T414 
default) board in September.  The board is intended as a graphics board:

    *  1/2 Meg frame buffer (4 bits deep)
    *  Custom VLSI for fast lines and polygons
    *  Generates its own video output
    *  2 Meg program memory

In early 1989 we'll have a 2-transputer PS/2 graphics board:

   1 T800 modelling chip (for databases, transforms, modelling, etc.)
      -- with 2 Meg program memory

   1 T414 rendering chip
      -- 1 Meg frame buffer (8 bits deep)
      -- custom VLSI for fast lines & polys

Plus an extension board sometime this fall to expand program and
frame buffer memory.
 
> What's the most important aspect to me: Will the host interface on that
> PS/2 board be faster (DMA!) than the damned software polling solution on
> all the PC boards now???

We won't support DMA on the single-transputer board.  We probably will on
the 1Q '89 two-transputer board.

This is all unofficial info.  Contact Fred Hudson at +1 512 832 1944 for
the real stuff.

-- 
-- Chris Chauvin  
   UUCP:  {...,uunet}!cs.utexas.edu!nth!chris
   INTERNET:  chris%nth.uucp@cs.utexas.edu