SI
SI
discoversearch

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: h0db who wrote (81782)4/4/2002 3:08:39 PM
From: Bilow  Read Replies (2) | Respond to of 93625
 
Hi h0db; Re: "When/if you finally see *real* dual-channel DDR chipset from VIA, SIS, Intel, do you think they will be able to use more than 2 DIMMs?" Yes. In fact, Intel's dual DDR channel chipset already allows 4 DDR DIMMs per channel (for a total of 8).

Re: "Will they be able to use more than 2 DIMMs and run at any speed faster than PC2100 with CL2 settings?"

Dual DDR at PC2100 matches the 533MHz bandwidth of future Intel processors. There's no need, at this time, to run faster, so why would you expect anyone to do it. But when there is a need to run faster, perhaps because Intel increases the FSB again or maybe when Hammer comes out, there are no technical reasons why faster DDR won't be able to work. But I would expect that to stop at PC2700. To get to PC3200 I would expect mainstream industry to use DDR2.

Re: "For Intel, I think that dual-DDR makes more sense, but why not just jump straight to DDR-2?" Intel has already nearly destroyed their chipset business by jumping straight to RDRAM. It's better to wait until the memory makers have the stuff down. There's no rush. Companies make money by using technology when it's ready, not by jumping straight into bleeding edge technology. DDR2 will arrive in its time, and Intel has already announced support for it. But it's neither needed nor economic to push technology into places where it isn't need when it isn't ready.

Re: "I mean the spec is all done and everything right?" Getting something into production starts with getting a spec published. There's a lot more to it than that.

Re: "BTW, which spec do you think they'll use-- JEDECs, or AMI2?" The JEDEC spec will be mainstream. Anything AMI2 comes out with will be picked up by JEDEC or remain in niche status.

Re: "That is to say, Micron's, or Samsung's?" As far as I can tell, there is zero difference between Micron's and Samsung's specifications for 400MHz DDR2. Maybe you're getting the DDR1 400MHz (which I think will remain a niche part but this is not a big deal) confused with DDR2.

Re: "And will that be DDR-II, or DDR-IIa?" DDR2a will be a niche product.

All I'm doing here is giving my viewpoint on this. A design engineer wants to use the mainstream parts. The safe route is to use standard JEDEC DDR2. That's what will be supported by all the memory makers. I know that the other engineers are thinking the same thing, so that is what I will use. The end result is that JEDEC DDR2 will be the mainstream product (in good time), and will then be the cheapest memory.

We've been doing this for 2 decades now. If you're familiar with the industry it's pretty easy to figure out what's going to be mainstream. The only real glitch was when Intel got married to Rambus. That made it very hard to predict for a while. A lot of engineers concluded (wrongly) that RDRAM would become mainstream because of Intel's push. When the Camino failed it became obvious that RDRAM was never to make mainstream status. That was in late 1999. Since then it's been quite obvious where memory is going.

-- Carl