Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : 3DFX -- Ignore unavailable to you. Want to Upgrade?

To: Patrick Grinsell who wrote (13138)	6/7/1999 2:27:00 AM
From: Waldeen	Read Replies (1) \| Respond to of 16960

Pat, yes, I remember this. But that was then and this is now. I view Intel's focus as changing. (Though you are right, they likely would still be difficult to work with.) At that time they thought graphics was the push for more processing power recent acquisition of DLGC and LEVL (now delayed), Shiva, Williams Comm. look like a push into Interent Telcommunications and home networking. Not to mention that they are rumored to be looking at acquiring or increasing investments in COMS and/or BRCM. the-adviser.com A good quote from here is "The Internet is a problem for Intel - Close to 50% of US households now have a computer in their home. Semiconductor companies such as Intel and AMD have tremendous capacity to make old but still acceptably fast computer chips. Current software requirements are increasingly being handled by these older technologies. In prior years, Intel counted on PC buyers who consistently upgraded old and slow processing PCs to handle new spreadsheets and word processing demands. This is no longer the case. PC's produced in 1997 are more than capable of handling today's Internets requirements. New software such as Microsoft 2000 will not drive upgrades. Neither will the Internet. The Internet is driving upgrades to modems and bandwidth related equipment." Therefore Intel wants to be the embedded processor in the network equipment, the more functionality required such as Voice over IP the better. The more processor upgrades needed. The questions I have are: has Intel laid out a path for future successor to the i740 and will they be able to compete with the likes of 3dfx and Nvidia? Have they announced any intentions for future graphics products that will drive upgrades of their products, especially the high end/high margin processor product? Isn't it likely that addition of geometry acceleration onto the graphics processor will likely have the opposite effect, or at least slow the need for increasing the processor power by removing it as the bottleneck in geometry set-up? This last point I know to be true from years of using hardware geometry acceleration in high-end Unix machines for 3D CAD related applications (mostly Finite Element Modelling). Ask someone who has used an SGI. Geometry acceleration means moving tasks off the CPU onto a 'graphics processor' specific for those applications. Certainly Intel realizes this. Frankly, I can't see Intel to be a fan of hardware accelerated DVD decoding or they wouldn't have demonstrated a processor only solution. But the graphics card manufacturers are now doing this. So how are they going to like geometry acceleration too? My opinion is Intel tried to push graphics processing as a way to drive processor upgrades, through investments in 3DLabs, acquisition of Chips and Technologies, AGP 2X, 4X, etc. It didn't work, and they've moved on to networking... set-tops and StrongARM being related technologies. If this assessment is correct, the Intel monkey is at a minimum off 3dfx's back, and at best they are now more in the drivers seat of graphics. But to work with Intel, or even get them interested, you need to push more work onto the processor i.e. demand more processing power. Can't see them being a fan of geometry acceleration unless you can tie it to specific needs for MMX or SSE related functionality to be increased over time. Maybe last time, 3dfx did not demonstrate this? What Intel would like to see is the Voodoo4 working only with the Pentium III or better. Voodoo5 requiring Merced, etc. Do that, they'll be interested. Waldeen