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: Tenchusatsu who wrote (30713)	9/26/1999 5:13:00 PM
From: grok	Read Replies (1) \| Respond to of 93625

RE: <You talk about the "superiority" of "Random Accesses" over "streaming accesses" without knowing what the hell you are talking about. (Rambus is better in both than anything, except perhaps VC.)> Tench, you have no excuse for not knowing that CL2, PC133 has better latency than any Rdram. Non-tech people on this thread may not be able to see through baloney like that Samsung Whitepaper but you certainly should.

To: Tenchusatsu who wrote (30713)	9/26/1999 5:53:00 PM
From: Dan3	Read Replies (1) \| Respond to of 93625

Re: your arguments against Rambus... Re: You talk about the "superiority" of "Random Accesses" over "streaming accesses" No, I stated that PCs most need random access rather than streaming access. Neither is "superior" to the other, they are just different aspects of memory performance. Streaming access is important for applications like Video cards that block shift regions of memory around. The aspect that is important to PCs is random access. Rambus advocates keep reiterating that it has better bandwidth, get kind of mumbly when it comes to latency - but never seem to want to consider whether or not there is any need for high bandwidth serial accesses to memory in PC. They get starry eyed and talk about the importance of "vision". They have a hammer, and don't bother to consider that the target might be a screw rather than a nail. Re: You say that PC166 can work in PC100 motherboards, yet you don't realize that's a non-argument since you'll lose all the benefits of PC166.... Overclockers are getting plenty of benefit from it now, and in the process its faults are coming to light in time to correct them. PC OEMs know how to overclock as well. What do you think testing involves? Part of it is to push a product past its specifications and see when it fails. Re: you say that PC166 comes from the same binsplit/process as Rambus 800. We've all posted the links from Samsung's, Micron's, Infineon's, Hyundai's, etc. semi pages too many times for you to not know that the .18 generation parts that constitute Rambus 800 are from the same .18 generation parts that yield 166mhz SDRAM. But many Rambus advocates keep comparing Q1 '00 rambus parts with last year's SDRAM specifications. You've looked at the datasheets before, you know where they are, you can look at them again. Sorry if you are a big believer in rambus, but I have real doubts about the whole approach, even after this current problem is solved. We all have to keep doing our best to evaluate these things, and try to make the best investment decisions we can. Unfortunately, to a large degree, my conclusions are based upon a series of articles that I read about 10 years ago in Byte and some other magazines regarding benefits of cache for various types of applications. These issues were no longer discussed much once all CPU's began having at least some cache on chip. They predate posting things to the web and, though I've searched, I can't find them. The gist of the articles, as best as I can recall, is that even with a small cache like 16K, a system executes many instructions before it goes out to main memory for more. And when there is a cache miss, it's usually not just to the next address following the end of the most recently retrieved cache line - so most accesses to main memory are not serial, where Rambus has its big advantage. I appreciate your input to these discussions, I know this isn't a great weekend for anyone with an emotional, financial, or work related connection to Intel. It could be that Monday's news won't be so bad after all. Maybe they can just solder in the RIMMs and everything will be OK. Dan