Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!rutgers!ames!ucbcad!ucbvax!decvax!dartvax!uvm-gen!vmix!einstein
From: einstein@vmix.UUCP (Fredric J. Einstein)
Newsgroups: comp.graphics,sci.electronics
Subject: Re: Questions about building a graphics system.......
Message-ID: <191@vmix.UUCP>
Date: Tue, 9-Jun-87 15:47:40 EDT
Article-I.D.: vmix.191
Posted: Tue Jun  9 15:47:40 1987
Date-Received: Fri, 12-Jun-87 01:05:55 EDT
References: <4332@hc.DSPO.GOV>
Organization: Vermont Microsystems Inc. Winooski, Vt. 05404 (802) 655-2860
Lines: 26
Keywords: Am95C60, hardware, HELP!
Summary: Problems with the AMD QPDM
Xref: mnetor comp.graphics:744 sci.electronics:801

In article <4332@hc.DSPO.GOV>, naivar@hc.DSPO.GOV (Mark A. Naivar) writes:
> 
> 
> 	Hi there. I'm thinking about designing a graphics system around
> the "hot new" graphics processor from AMD: the Am95C60 Quad Pixel
> Dataflow Manager (QPDM) and I have a few questions for all you graphics
> "experts" out there. But first, let me explain the basic layout of the

	I am using a four plane board based on the AMD QPDM (I wrote an
MS-WINDOWS driver for it.  The current state of the QPDM involves some 
VERY serious software bugs that won't be fixed till AT LEAST November, 1987.
Among these bugs are the inability to write bitmap data from the host to
the graphics board and to read back bitmap data from the board due to bugs
in the chip's firmware.  Although the chip is incredibly fast and easy to
write software for (especially MS-WINDOWS and DRI GEM drivers -- I've written
them for INTEL '786, TI-34010, and the VMI PGL interfaces), these bugs really
throw a spanner into the works.  If you don't care about these SERIOUS bugs
having to do with bitmap data, the AMD is a great, fast and flexible (if not
cheap) graphics co-processor.  As far a speed is concerned, the biggest 
bottleneck in any PC-based graphics system is the board communication with 
the host.  Number of planes etc make little difference except for the 
number of bytes that must be transmitted between the host and the graphics
engine.  I don't know anything about the SUN, but on the IBM-AT bus, this
is what really slows everything down.

Fred Einstein