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: jim kelley who wrote (58942)	10/26/2000 5:54:21 PM
From: Tenchusatsu	Read Replies (1) \| Respond to of 93625

Jim, <I have heard that Intel hobbled the 820's performance elsewhere. I do not know if that is true but if so it seems have been a mistake.> This is true, but it's no mistake. Due to power constraints, Intel had to add a feature to the 820 chipset which keeps only a fraction of the RDRAM devices in the memory channel active. This does two things: 1) It negates the advantage of having a huge number of banks (an advantage Rambus loves to brag about) 2) It increases the average latency of the memory subsystem, since it takes a few extra cycles to bring an RDRAM device from standby state to active state. Since the devices are juggled between active and standby states, latency grows longer than it should be. (Perhaps this is what prompted some of the anti-Rambus people to crow about the latency issue. It's not the packetizing of data that creates the long latencies, as they claim. Rather, it's the juggling of power states.) I think this power management "feature" points to a fundamental difficulty with implementing RDRAM on low-cost, high-volume PC platforms like the 820 should have been. This could have also been a foreshadowing of other problems Intel ran into which were Rambus-related, such as the MTH fiasco. Tenchusatsu P.S. - The 840 chipset is able to keep a larger number of RDRAM devices in the active state, which is why the latency in that chipset is lower. However, because 840 is a higher-end (a.k.a. more expensive) platform, the power constraints are a lot looser than the 820 platform.