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: Dan3 who wrote (35965)	12/19/1999 5:46:00 PM
From: Don Green	Read Replies (1) \| Respond to of 93625

O.T. The Pentium III Tablet computer. You'll never see computers the same way again qbenet.com

To: Dan3 who wrote (35965)	12/20/1999 2:32:00 AM
From: Tenchusatsu	Read Replies (1) \| Respond to of 93625

Dan, <The advantage of Rambus is that it can support nearly unlimited burst length, which is the "efficiency" being touted in that table. What is absent from that table is latency - where DDR has a large advantage over Rambus.> I really wish you'd stop using the term "latency" like a buzzword, because in reality the latency question is a lot more complicated than you make it sound. No, DDR does NOT have a "large" advantage in core latency over Rambus. If anything, the difference is about 10 nsec due to the packetizing of data on RDRAM. What really counts is the average latency, in other words the latency that is seen in normal system operation. A large part of the average latency is determined by the efficient use of bandwidth. The less efficient the bandwidth is used, the worse the average latency will be in a heavily-loaded system. On the other hand, in lightly-loaded systems, average latency is very close to core latency. But improving the performance of lightly-loaded systems isn't the aim in high-end systems. Anyway, DDR attempts to brute-force its way past its inefficient use of bandwidth by boasting of a higher max bandwidth than RDRAM. A higher max bandwidth can help reduce your average latency depending on the situation. But this brute-force method also means bandwidth is utilized even less efficiently than before. That doesn't matter at long as overall performance is increased, but it does bring about diminishing returns. (Interestingly enough, Rambus can also pull off a similar trick and brute-force max bandwidth as well. The dual-RDRAM 840 chipset is a prime example.) If anything, all other things being equal, I'd expect single-channel DDR systems to be virtually neck-n-neck in performance with single-channel RDRAM systems. That's competitive enough for me. Tenchusatsu