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 : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?

To: jim kelley who wrote (68565)	11/15/1998 2:49:00 AM
From: Tenchusatsu	Read Replies (1) \| Respond to of 186894

<There does not appear to be a I/O problem currently with desktops or servers. The big problem is ethernet and internet bandwidth. It is going to take sometime for the bandwidth problem to be resolved. By the time it is resolved you will see many more solutions to the I/O bandwidth which will then be an emerging issue.> Bandwidth is definitely a problem for single-card solutions like a gigabit Ethernet adapter. There is a definite need for proposed standards like PCI-X or NGIO in servers. Obviously desktops and laptops aren't going to need PCI-X or NGIO. Desktops already have AGP to take care of graphics and PCI to take care of just about everything else. Tenchusatsu

To: jim kelley who wrote (68565)	11/15/1998 3:45:00 AM
From: rudedog	Read Replies (1) \| Respond to of 186894

Jim - don't beat up on poor john thinking it's me - I have been off the Intel thread for about a week, just too busy to read here, and John was quoting one of my CPQ posts. You can think what you like about the CPQ consumer product margins, I think we have beat that one to death. If CPQ is making less than 20% GM on those products I will be very surprised, but then I was surprised when they stuffed the channel in late '97 so I'm ready to be wrong again. But on your other point - There does not appear to be a I/O problem currently with desktops or servers that is way off for servers. A single gigabit ethernet card can saturate a 32/33 bus, and a single array controller can saturate a 32/66. A big server with 8 or 10 array cards and a couple of NICs saturates the bus often. Just getting to a good 64/66 model helps a whole lot. PCI-X helps a little more. But the reality is that current designs, let alone 8-way Xeons with big I/O, will be bottlenecked on even the PCI-X flavor. NGIO is too slow by a factor of at least 4 for today's needs. CPQ already has several proprietary high end bus designs. What I expect them to do is get one which will solve the problem in the volume server market in a way that maintains backward compatibility while allowing technically advanced new solutions to reach their full potential. Intel has worked hard on NGIO and there are many good features in the design. It's just too slow and still too dependent on system resources which will bottleneck.