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: Dan3 who wrote (29268)	9/11/1999 1:34:00 PM
From: Alan Bell	Read Replies (1) \| Respond to of 93625

If one is running on a desktop where power is not an issue, leaving all the chips in standby makes sense. But on a laptop where power consumption is an issue, one would only keep active the chips that are actually being used. In a system, there are always multiple applications running at a time. Each application only uses a part of the memory and the memory that a given application is using will often fit in one bank. While the banks that are active will be controlled dynamically, generally these changes will be correlated more with process switches than with individual memory access as you suggest. Taking the 90ns hit on a process switch is not a big deal. Secondly, Laptop usage is bursty. The CPU will be used intensely for a short while and then sit idle. For example, if one is typing to Word, when the character is typed, Word has to do a lot of work to properly update the display. It then sits in the idle loop. When one fetches a web page, one wants that page to display quickly, then one reads it and the CPU is in the idle loop. With DDR ram, all banks must be kept active during these short idle periods. With RDRam, only one bank needs to be kept active while in the idle loop. This saves a lot of power, John is actually being very conservative in assuming that 4 chips will be active at any given time. It will usually be much less, often just one. This makes the RDRam power savings more like a 3 to 1 ratio. -- Alan

To: Dan3 who wrote (29268)	9/12/1999 8:10:00 AM
From: John Walliker	Read Replies (1) \| Respond to of 93625

Dan3, On page 46 it shows an NAP exit delay adding an additional 50+40=90ns to the latency delay. OUCH! Going to NAP at every opportunity will make your Coppermine 733 equal in performance to about a Pentium 266. But you will be using a little less power than PC100 that way. 90 ns is not a lot if it is invoked only occasionally. Much better to have more memory than is needed most of the time and minimise swapping to disc. Then the excess memory can be switched in as required and be in power-saving mode the rest. It costs a little more, but gives a much more responsive machine. John