Wave Issue 9610 10/11/96 Article 2-01

***3DLabs Charts the Future
by John Latta

3D Labs has had a leadership position in 3D acceleration since its days
within DuPont in England. WAVE recently visited 3DLabs to seek its views.
Although the company has had its impact in the upper end of the market it
is planning on how it will do the same across the total market including
the consumer space. Its planning is being shaped by how rapidly moving
markets evolve and one of the best descriptions is like a Tornado. A good
resource on this analogy is Geoffrey Moore's book "Inside the Tornado."

In spite of over 2.5 million 3D chips having been shipped already in
1996, 3DLabs has not seen where 3D is consumer market hit. 3D is being
pushed into the market and there has yet to be a buyer pull. Voicing a
view not often heard in much of the 3D community they see where VRML may
be the first 3D software technology to get wide spread use and thus
create buyer demand. It could well be that the Web will accelerate the
adoption of 3D. When this happens the industry will be subject to "Web
Time," that is, product cycles of 3 - 4 months will push the industry
both in hardware and software. 3DLabs believes that in order for 3D to
reach the mainstream there must be another ISV software development cycle
for high quality 3D applications and content to surface. This implies
another 12 - 18 months before 3D is a mass market Tornado.

In their view achieving mass market effectiveness with 3D will not be
easy. For example, the lower the price of 3D, especially in add-on
products, the more the consumers will expect in terms of functionality,
flexibility and integration. This is just the opposite from the high end
of the market where 3D acceleration typically has a very focused use and
typically with only one application. In its planning for the future the
company sees a traditional market pyramid with the upper market at the
top and the consumer space at the bottom and the width of the pyramid
representing the size of the market. At the top are its GLiNT products
along with Intergraph, for example, and its workstations. In the mid-
band, which has not yet been well addressed, is its Permedia NT product.
The best example of a 2D product today is Matrox Millennium but it has
not addressed the 3D market.3DLabs sees this market as being largely
under served and it feels that this will emerge in 1stH 1997.

The consumer market today is being driven by S3 and ATI, yet, the Tornado
of this market has not begun. The key reason being the lack of consumer
demand, in part, based on immature software products. 3DLabs sees this
market emerging in 1998 and it intends to be a strong player with its
next generation products based on Permedia technology and its
relationship with TI. Yet, today the company has its focus on the high
end of the market which it shaped in 1995 with the 300SX product and now
is based on its 500TX products. Winning OEM designs is also critical in
this market and 3DLabs has landed Compaq, HP, DEC and Siemens. An area
which 3DLabs has also led the industry is in the role of set-up engines
and 3D geometry acceleration as a critical element in 3D performance.
Based on its own tests some acceleration is required to keep the
rasterizaton pipeline full, even with Pentium Pro 200MHz processors.
Today this need is being fulfilled by the Delta set up processor but
coming in Q1 1997 will be a full geometry engine called Gamma. (See the
article below on the Gemini benchmarking for more background on the
importance of geometry acceleration in hardware.)

www.3dlabs.com