To: Dan3 who wrote (27199 ) 8/16/1999 9:06:00 PM From: Tenchusatsu Read Replies (2) | Respond to of 93625
Dan, it's kind of obvious that I can't post any Rambus benchmarks yet. There are non-disclosure agreements for a reason, you know. I have seen a few foils revealing internal performance data for Rambus, and in my limited view the results are encouraging. That's why I'm telling you and others to wait for IDF later this month. As for cost, the wide gap in prices between PC100 SDRAM and DRDRAM is due to two factors. The first one is a glut of SDRAM, leading to the extremely low prices. And the second is the scarcity of DRDRAM, which is no surprise given that DRDRAM volumes haven't ramped up yet. It's a double-whammy leading to an artifically inflated price gap. <Or something, but just wait until... To me, that's the issue with rambus - it was conceived when it looked as though 100MHZ was it for DRAM - it had been a rough transition from 66 to 100. But that was a long time ago, and the SDRAM variations have proven to have had more potential than was expected.> Dan, you're not giving Rambus the benefit of the doubt. I'm personally a Rambus advocate not because of immediate benefits, but because of long-term. No other technology will come close to the bandwidth-per-pin of DRDRAM. No other technology allows for better, more flexible streaming of data from multiple sources. No other technology is already integrated into the designs of two upcoming CPU's: Timna from Intel and Alpha 21364 from Compaq/Digital, with more on the way. SDRAM is quickly becoming a legacy technology, something that will only be supported because of the lower (near-term) cost and the existing infrastructure. As for VC SDRAM, I don't know much about it, but my impression is that it's a modified form of ESDRAM, which itself is soon to enter oblivion. My impression is that the benefits of virtual channel technology can be (and should be?) implemented in chipset buffers, thereby reducing the need for VC SDRAM. <The EV6 bus is designed to eventually run at up to 400MHZ at double data rate, just like rambus - and it's 64 bits wide instead of 16 bits.> Dan, you're getting confused again. When AMD says that the EV6 bus can run at 400 MHz, it's actually a 200/400 speed, which is half the speed of DRDRAM 400/800. Speaking of which, I can't think of anything better to feed that 400 MHz EV6 bus than with two DRDRAM channels. Can you? <VC is already here, you can buy it now, and the PC motherboards that use it will ship just before, or just after, the rambus motherboards.> VC is already here?!? What?!? Dan, statements like these are one reason why I'm on the verge of giving up trying to convince you. Where can I buy VC SDRAM? After all, if VC is already here, I should be able to drive to my local Fry's Electronics, pick up 128 MB of VC SDRAM along with a motherboard that supports VC SDRAM. And of course, I should see the same performance benefits that I'd expect from DRDRAM. Alternatively, I should be able to see vendors on Pricewatch.com who are already selling VC SDRAM. I expect to see DRDRAM in stores this fall. And I can assure you that DRDRAM is more real than a couple of blurbs on a VIA chipset feature list, or a few foils put out by NEC. <The discussion about rambus's huge performance benefit that guarantees market acceptance at any price - that's what I don't think makes sense.> Fair enough, but then you should keep your arguments constrained to this area, instead of hyping up alternatives to DRDRAM without applying the same harsh standards of judgement. By the way, why accuse Rambus of assuming that the competition will stand still? From the tone of your posts, you seem to be assuming that Rambus and Intel themselves are standing still. Tenchusatsu