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.

Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?

To: Maxwell who wrote (37066)	9/16/1998 2:09:00 AM
From: Elmer	Read Replies (1) \| Respond to of 1571039

Maxwell check out this. news.com Intel is showing 500mhz Katmai systems today. Not Q4 not Q1'99 but they have working Katmai now. Do you really think release is far away? EP

To: Maxwell who wrote (37066)	9/16/1998 5:20:00 AM
From: Tenchusatsu	Read Replies (1) \| Respond to of 1571039

Maxwell, Believe it or not, only the Voodoo2 can unleash the power of K6-2 and PII. Without the Voodoo2, you only get about 30-40fps in QuakeII. With Voodoo2 you get 60-80fps. Voodoo2 really enhances geometry setup which is lack in the CPU. Believe me, I know. I have a Voodoo2. But I don't think two is necessary for me yet, since I "only" have a PII 266. But I'm talking about those wackos who feel like getting two Voodoo2 cards in SLI right now, along with a Pentium II 450 MHz. If I'm willing to pay the extra cash to go to Coppermine and RDRAM, so will they. 1) In 3D processing 3DNOW is same as KNI Interesting assertion, since KNI introduces an additional 8 128-bit registers. 3D-Now can only do four single-precision FP ops at a time by concurrently executing two instructions which each take 32-bit operands, adding up to a total of 64 bits. I wouldn't be surprised if Intel allowed eight single-precision FP numbers to reside in a single KNI register. Wouldn't that be double the processing power of 3D-Now? (Disclaimer: No one really knows the real details of KNI yet. I'm only basing these KNI assumptions on a link I saw in the Intel thread.) A) KNI is more versatile than 3DNOW but will not be any significantly faster than 3DNOW in 3D games. Thus KNI will never replace 3DNOW. Not if KNI can process eight single-precision FP numbers at one time. That's two 4-way vectors in one instruction! B) AMD already stole the thunder out of KNI. KNI will just be a 3DNOW METOO from Intel to gamers. Apparently for some people this rings true. But remember how Intel was able to move people over from Socket 7 to Slot 1? Or how easily Intel was able to move people from regular Pentium to Pentium MMX? C) Intel is half committed to KNI since not all their products have KNI. In other words, KNI may not be supported in the future CPU. I'm sorry, but this is the second most idiotic statement I've ever heard. (The first would be that statement which includes the words "legally accurate.") For one thing, Intel is already starting work on processors which will be released in the 2002 timeframe. Do you think for a second that Intel will drop KNI in those future processors? Already you can expect Katmai, Tanner, Coppermine, and Cascades in 1999, all with KNI. There's no doubt in my mind that the Willamette (a.k.a. the P7) will also support KNI, if not more instructions. If Intel were to drop support for KNI in the future, that would be analogous to Craig Barrett personally disabling the safety features of his microwave oven, sticking his head inside, and turning up the power to "high." D) If Intel markets KNI well and KNI takes off then 3DNOW would become popular due to "piggy-backing" effect. CeleronA would then be back in the dog house. You're assuming that the Celeron line won't receive KNI in 1999. If KNI doesn't take off, and/or if AMD's 3D-Now is more popular than expected, you can bet that Intel will push KNI down to Celeron as fast as it can. If not, you can bet that Intel is enjoying the fruits (i.e. profit margins) of keeping KNI in the mid-to-high range of its processor line. All your assumptions seem to stem from a faulth impression that Intel is a stationary target. I think AMD wishes that the game were that easy, but it isn't. AMD would have better luck exploiting the holes that Intel leaves open instead of duking it out with Intel one-on-one (and losing). Tenchusatsu

To: Maxwell who wrote (37066)	9/16/1998 10:52:00 AM
From: Yousef	Read Replies (1) \| Respond to of 1571039

Maxwell, Re: "Here is an article ..." Well Maxwell, here is another article that you might find "interesting". Sounds like AMD will soon be more than 100mhz behind Intel. Oh BTW, Dr Barrett appears to be agreeing with Yousef ... Imagine that. <ggg> ebnews.com "Barrett used his talk as a platform to convince developers that the copper interconnect and Silicon-on-Insulator (SOI) processes used by Motorola, IBM and other semiconductor makers are not the way to increasing performance within a microprocessor. Instead, Barrett said that Intel will concentrate first upon increasing the speed of the transistors within the chip, moving to the faster copper interconnect process only in its 0.13-micron generation." Make It So, Yousef