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: Tony Viola who wrote (83286)	6/12/1999 4:09:00 PM
From: kash johal	Respond to of 186894

Tony, Re: RDRAM to cost $400 for 256Mb. >In the industry, Mb means megabit and MB means megabyte. You had me >sweating for a few seconds there. >also, article says: >said a 128-megabyte memory subsystem >stocked with Direct RDRAM will cost between $200 and $250 this year, >possiblydropping to less than $150 by the end of 2000. >So why are you extrapolating out to 256 megabytes? Because $400 has better shock value? No because hi-end CPUS nowadays come with 256Mbytes of RAM. Actually $500 has a higher SHOCK value - but I chose the lower number. And remember this is from a DRAM guy talking about price to OEMS. So in end user system a $300 higher RDRAM will translate to a $400-600 higher system cost. >I'm not saying much ado about nothing, but, still, I think you and >the article are overstating the importance of this. Starting at 600, >rather than 800 MHz is probably a damn good idea for getting a good >handle on the transmission line characteristics they're dealing >with, although it may be more a decision based on yield, or bin >split in CPU terms. Only 600 MHz, indeed! Man, that's practically >crawling! And, as was discussed here before, what's the overall >system performance degradation going from 800 to 600 MHz RDRAM? Is >it measurable? The Christmas thing, yes, I can see that, but, what >change does a mobo need to go through to support 800, after being >set up for 600 MHz. A crystal oscillator output, what else? BIOS? >OK, nothing is trivial. Well unfortunately due to the RDRAM architecture you have a narrow interface bus and it requires the high datarate. Due to the higher latency many folks have speculated that even 800Mhz RDRAM will have No speed advantage over PC 133 and so 600Mhz RDRAM could be slower than PC133. And AMD at the K7 presentation stated that they probably wouldn't even support RDRAM even next year and would probably go with 266Mhz DDRDRAM. It would seem logical that they would use RDRAM if it truly offered any significant system level advantages. And clearly the advent of 600,700 and 800Mhz RDRAM point to SIGNIFICANT yield problems at 800Mhz otherwise they would not have bothered with anything less than the 800Mhz speeds. Regards, Kash

To: Tony Viola who wrote (83286)	6/12/1999 4:18:00 PM
From: grok	Read Replies (2) \| Respond to of 186894

RE: <Only 600 MHz, indeed! Man, that's practically crawling! And, as was discussed here before, what's the overall system performance degradation going from 800 to 600 MHz RDRAM? Is it measurable?> Heck yes it's measureable! The problem is that latency increases by 33% which is huge since drdram is weak on latency to begin with compared to plain old sdram. And then to have to pay extra (in fact it may be much extra) to have the honor of helping Intel achieve their master plan will just cause each box maker to look harder at AMD. This "Pay More, Get Less" Rambus fiasco shifts the overall competitiveness between Intel based PCs and AMD based PCs in AMDs direction and at a time when AMD is getting more competitive with 0.25u yields up and K7 looking good.