Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Display Grid Methodology for Reduced-Cost Windowing

IP.com Disclosure Number: IPCOM000118753D
Original Publication Date: 1997-Jun-01
Included in the Prior Art Database: 2005-Apr-01
Document File: 4 page(s) / 171K

Publishing Venue

IBM

Related People

Jaaskelainen, W: AUTHOR [+2]

Abstract

Disclosed is a mechanism for reducing the product cost associated with windowing on computer displays. This is accomplished by the replacement of the typical display buffer on a display adapter by a significantly smaller addressing buffer pointing to screen image areas in main memory. This technique is particularly well suited for network computers or similar products where product cost is deemed more important that raw speed. Software functions which exploit this hardware are also disclosed.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 32% of the total text.

Display Grid Methodology for Reduced-Cost Windowing

      Disclosed is a mechanism for reducing the product cost
associated with windowing on computer displays.  This is accomplished
by the replacement of the typical display buffer on a display adapter
by a significantly smaller addressing buffer pointing to screen image
areas in main memory.  This technique is particularly well suited for
network computers or similar products where product cost is deemed
more important that raw speed.  Software functions which exploit this
hardware are also disclosed.

      In a typical personal computer running multiple applications,
each application has an area of main memory which holds its current
screen image.  A part of the operating system called the display
manager is assigned the task of copying portions of each
application's screen image into appropriate areas (windows) of the
display buffer on the display adapter.  The display buffer thus
contains a duplicate copy of  information which is already held in
main memory.

      If, for example, the display space is 1024x1024 pixels and each
pixel is capable of holding one of 64K concurrent colors, the display
buffer size requirement is 2MB.  There is significant cost associated
with 2MB of duplicate memory.  This disclosure presents alternatives
for replacing the display buffer with an addressing buffer and
associated hardware which provides pixels directly from the
applications' main memory screen images via a Direct Memory Access
(DMA) or equivalent technique.  The addressing buffer contains
addresses in main memory of  the pixel(s) to be displayed, along with
associated parameters to be discussed below.  The addressing buffer
itself might reside either in a  small memory on the display adapter
or in main memory to be DMAed on demand for further cost saving if
performance allows.  Four refinements  of the basic idea are then
presented with progressively larger reductions  in total memory.  The
example given at the start of this paragraph will  continue to be
used to illustrate relative savings of the refinements.

      Assume that each pixel address in the addressing buffer points
to the pixel in an application's screen image.  Assuming a 32-bit
address space, this requires a 4-byte address per pixel or 4MB total
-- twice the original size.  Not a good solution but one on which the
four refinements will be based.

      Refinement #1 defines the granularity of window placement to be
on a grid of NxN-pixel boxes rather than on a pixel-by-pixel basis.
All pixels in each box belong to the same application.  The pixel
addresses in the addressing buffer in the preceding paragraph would
now be box addresses.  For N=2 there are only 512x512 boxes instead
of 1024x1024 pixels.  The addresses of the 4 pixels in the box can be
calculated, using simple adders on the display adapter, from a single
box address and  offsets.  The offset for horizontally adjacent
pixel...