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 : Rambus (RMBS) - Eagle or Penguin -- Ignore unavailable to you. Want to Upgrade?

To: Dave B who wrote (19479)	4/29/1999 4:16:00 AM
From: Tenchusatsu	Read Replies (3) \| Respond to of 93625

Dave and RMBS investors, I'm starting to understand why there's such negativity toward Rambus and Direct RDRAM. I think most of the criticism amounts to "Will it really make a difference?" I have a few anecdotes which should shed more light on the current situation with RDRAM. Before PCI was invented, the rage was "local bus" graphics. Older non-local-bus graphics cards were hampered by the ultra-slow 8 MHz ISA bus. The first local-bus cards were made for a local-bus standard called VESA local-bus, and they were much faster than the ISA graphics cards. Suddenly, everyone wanted one, and "local bus" became a marketing buzzword. However, VESA wasn't really the proper way to implement local-bus, so Intel created the PCI bus spec and made it an open standard. A brief VESA-vs-PCI war started, and PCI soon won out because the new Pentium systems were optimized for PCI. Then, there was a real need for PCI. Just in the past year or two, Intel predicted that even the good old PCI bus wasn't fast enough to support the new 3D graphics cards being released. Plus, Intel thought that it would be way too costly to put large amounts of memory directly onto a graphics card, so Intel created a new graphics-port spec called AGP. AGP allowed the graphics card to use system DRAM memory as its own, thereby giving the graphics controller access to a huge pool of memory. There were several people who were against AGP, one of which was Tom Pabst of Tom's Hardware Guide. (Paul Engel and I on the INTC thread like to call him Herr Uberclockermeister.) Tom felt that AGP was unnecessary, and he brings up two very valid points. One is that one of the fastest 3D solutions out there (until recently) was Voodoo2 SLI, which is PCI-based. The other point is that memory is super-cheap now, and graphics controller designers are able to put like 16-32 MB of graphics DRAM onto the card without adding much more to the cost. Though the need for AGP may not have been as dire as Intel once thought, the truth is that just about every new 3D graphics card on the market is AGP. Plus, AGP helps eliminate some barriers that games and graphics application writers have been working around for some time. The truth is that these days, hardware innovation drives software innovation. For example, without AGP, game writers wouldn't feel free to throw in huge amounts of textures into their 3D graphics. They'd have to resort to tricks like texture compression which affect image quality. Even Tom Pabst, once an AGP critic, is now accepting the fact that AGP is becoming useful. Now we could be seeing the same things happen with other aspects of the PC, like: 1) Processor (your basic sub-$1000 PC has enough computing power for the majority of users out there) 2) AGP-4x (currently, the bandwidth AGP-2x rarely ever gets saturated) 3) RDRAM However, Intel is hoping that if it drives the hardware, the software will eventually follow and utilize the new technologies. It might take a few months, or it could even take several years. For example, Intel introduced 32-bit mode with the release of the 386 processor back in 1990 (I think it was 1990). But it took Microsoft a very long time before Windows 95 opened up the floodgates of robust 32-bit software. So the next time someone shows up and says that RDRAM isn't necessary, ask them if they ever said that PCI wasn't necessary. Or AGP. Or UltraDMA. Or a Pentium processor. Or a Pentium II. Tenchusatsu