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: tinkershaw who wrote (75260)	7/2/2001 1:24:36 AM
From: Scumbria	Read Replies (3) \| Respond to of 93625

Tinkershaw, Nvidia has shown off a two-ported DDR system already. The claims of technical impossibility seem moot. The real issue at this point is politics as much as engineering. Intel has chosen to make i845 a lower performance chipset than the i850. This will have the effect of making the P4 RDRAM systems high performance, and the P4 SDRAM systems low performance. Rambus marketing will now have a reference platform to make a claim of RDRAM's superiority (something they never got with PIII.) This is exactly what Rambus needs to make their big move into the PC marketplace. With the RDRAM/DDR price delta rapidly diminishing, I believe Intel has defined the winner in the PC marketplace to be RDRAM. Scumbria

To: tinkershaw who wrote (75260)	7/2/2001 2:58:45 AM
From: Dan3	Respond to of 93625

Re: you pit two tables, not equivalently measured, There is no exact equivalent. The actual numbers in the sheets are DDR at 7.5 ns and RDRAM at 45 ns. But there is more to it than that (DDR bus latency called out as CAS 2 is built in to the RDRAM spec), and I wanted to present RDRAM in the best possible light, and show the worst case for DDR. Would you prefer the raw device numbers, 7.5 ns for DDR and 45 ns for RDRAM? I gave RDRAM every break I could, and it still loses. It's slower.

To: tinkershaw who wrote (75260)	7/2/2001 1:32:58 PM
From: jim kelley	Respond to of 93625

Tinker, These guys were probably trained by Goebels himself.<G> They have been spewing this BS incessantly for the last two years. JMO

To: tinkershaw who wrote (75260)	7/7/2001 1:10:58 PM
From: Bilow	Read Replies (2) \| Respond to of 93625

Hi tinkershaw; Re your engineering booboos: You wrote: "Yes, but that is because dual channel DDR requires 132 line traces plus control and addressing traces to function. DDR2100 had trouble with just the 66 line traces. Far too much EMI. RDRAM can do dual channels because it only taks 66 total line traces, of which only 60 are high speed." Actually, RDRAM has far worse EMI problems. The reason is simple, they're at much higher speeds. That's why RDRAM traces have to be provided with ground traces between the signal traces, and you forgot to list the ground traces in your trace count. Total center to center distance between two RSL (i.e. high speed RDRAM) traces is 40 mils, while the center to center distance between two SDRAM or DDR data traces is only 12 mils. See #reply-13853059 for links. In other words, RDRAM takes data at 3x the speed of DDR, but using traces that are 3.33x as wide. Now do you get it? The basic, underlying fact that you cannot deny is that AMD still isn't bothering with an RDRAM chipset, but Intel, along with all the independent chipset makers, saw fit to provide the P4 with DDR. Are you really so convinced that you understand engineering better than all those engineers who chose DDR? BWAHAHAHAHA!!! You're a lawyer and you couldn't even figure out what was going to happen in the *legal* side of this fiasco, now you're making statements about the engineering? The engineering questions have, for better than 2 years, already been debated at length on this thread. You're not a heavy hitter in this area. You're not even a light hitter. Stick to the legal stuff, that's where Rambus still has some sort of a chance. -- Carl