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: John Walliker who wrote (42439)	5/17/2000 10:08:00 PM
From: Bilow	Respond to of 93625

Hi John Walliker; Re shared masks between memory types... The thing to note is that the memory companies like to be able to redirect their prodcution at the last minute. Having all the earlier masks shared facilitates this. It also means that they can, if they have to, leave parts generic until they figure out what they want to do with them, though I don't think that they do this often. They would prefer to be able to produce DRAM in 2 weeks from start to finish, but that is not (yet) possible. So they try to keep their production as generic as possible. They do this with the x4, x8 and x16 SDRAM even though it undoubtedly increases the die size of the x4 part. The advantage is, in this case, more than enough to overcome the higher manufacturing costs. In addition, they save on having to make all the extra masks that 3 completely different dies would require, which can be a considerable savings. The ability to turn DDR chips back into SDR is the reason that the memory makers aren't too worried about ramping production up quickly. It is not a matter of exactly what die they are currently using, it is instead the principle that they won't get stuck with a big pile of unwanted DDR chips in the event of a Camino type DDR fiasco. Basically, this means that they can support DDR at less risk than otherwise. I was kind of surprised that I couldn't find more links to the fact that they are producing DDR on the same dies. Maybe I'm searching on the wrong key words... But then again, though I searched back to 1995, I couldn't find any references to the sharing of the x4, x8 and x16 masks. It may be that the memory makers aren't keen on letting too many details of their processes out to the public. Maybe some of the stuff I get told by the memory suppliers isn't completely public knowledge. But it should be pretty obvious that the memory houses do use common masks for the three widths. I know I've seen references in print, but I can't quickly find it. The only thing I've got is this statement from dramreview, which obliquely implies that the x4, x8, and x16 parts share a die: Small memory systems contain only a few DRAM devices, usually soldered down in close proximity to the memory controller. These systems demand high bandwidth out of only a small amount of memory, so memory granularity is very important. DDR devices that are used in these systems are 32 bits wide, which keeps granularity low but also does not share the same die, package, or test infrastructure as the narrower parts [i.e. x4, x8, and x16] that will be used in main memory once systems are available. dramreview.com Note that if the narrower parts didn't share a die, there wouldn't be much reason to distinguish them, as a group, from the x32 parts. The x32 parts, in addition to having a different die, also come in a different package. -- Carl