Content area
Full Text
Chip suppliers and OEMs are trumpeting 1996 as the dawn of a new age in desktop computing. Both are predicting that unheard-of numbers of personal computers with full multimedia capabilities-including workstation-level graphics and three-dimesional sound and video-will hit the home market like a hurricane.
While there does appear to be a trend in this direction, and Intel Corp. has been wildly successful selling Pentium-based machines to the home market this year, some of the predictions of stellar growth may be overstated. Recent figures from Dataquest Inc., San Jose, Calif., indicate that the growth rate of multimedia PCs in the U.S. home market is slowing. After two and a half years of 300% and 400% annual growth, the market is increasing at only 15% to 20% a year. Furthermore, demographic profiles of home PC buyers show virtually no change in the last three years, which could mean market saturation.
Recent research indicates that the home multimedia PC market will be driven primarily by add-ons and replacements rather than new systems. Nevertheless, OEMs and chip suppliers are gearing up to deliver higher-quality multimedia desktop systems for as low cost as possible. And the effort is producing both technological innovations and confusion over competing technologies and new architectures.
Like multimedia notebook manufacturers, OEMs designing desktop PCs have to cope with the Pentium processor's voracious appetite for power and memory. New desktop multimedia PCs also must cope with the similarly inclined Windows 95 operating system. As if that were not enough, a new hardware concept-Unified Memory Architecture (UMA), designed to save bill-of-materials costs by eliminating dedicated graphics memory from home PCs-was apparently arrived at independently of Windows 95, and can seriously degrade the performance of multimedia features when implemented without enough, or fast enough, system memory. Finally, there is the high bandwidth needed by multimedia applications, and the continuing question of when and where further integration of multimedia functions will occur in silicon.
The Memory Problem
UMA is an idea that can cut the dynamic RAM bill-of-materials cost to OEMs by $20 to $40, depending on the total amount of DRAM in a system. Consumer products are under such price pressure that saving an average of $30 to $40 or so is a big plus. The idea is...