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: Jim McMannis who wrote (55770)	4/17/1999 1:14:00 AM
From: RDM	Read Replies (3) \| Respond to of 1579241

128.146.138.201 K7 vs Coppermine. Here's where the real fireworks begin! By the end of Q2 '99 the K7 will be out. Everybody knows it will start at 500 MHz and will have 200 MHz FSB. Hey, did I say 200 mhz FSB? It's very unlikely that current SDRAM DIMMs will operate at 200 mhz. So AMD will most likely use a trick…make a 128 bit CPU with 128bit L2-cache that accesses two parallel sets of 100 MHz FSB 64-bit memory banks. It is the same trick Intel used to make the Pentium work with the 32 bit memory of the 486 CPU. So I speculate that the K7 will be in fact a sort of double enhanced K6-3 with full speed L2-cache. It is heavily rumoured that the K7 will have 512 KB L2-cache on die, but in my humble opinion we will see the K7 with 256 KB L2 -cache, because 512 KB L2-cache would increase the die size too much. Don't forget that the die-size increases from the K6-2 to K6-3 with 135-81= +/- 54 mm² and an extra 256 L2-cache won't do much. So we won't see 512 L2-cache until AMD implements the 0.18 Microm process. Still the K7 will be the biggest threat Intel has ever seen: 128 bit L2-cache SMP support, so may be we can use 2x K7 under Windows NT 5.0! Enhanced K6 architecture Slot A, the K7 will not be limited by Super 7 as the K6-2 / K6-3. Last but not least : a Pipelined FPU, to compete with Intel in all areas. Intel has already indicated how it will react to this threat: a Katmai 533 with 133 mhz FSB, and then the Coppermine with 600 mhz. It still remains to be seen if the Katmai-600 CPU (because differences between Katmai and Coppermine are small) will dethrone the K7. First of all, going to a smaller fabrication process is a difficult task and some delays will more than likely prevent the Coppermine-600 from competing with the K7. Why will AMD lead with the K7? Simple, Intel's resources are divided. They are aiming at the high-end server with the Merced, the midrange servers with the Tanner , the workstations with the Katmai… So we can look forward to exciting times if AMD survives their cash problem and gets good chipset support for the K7… It is uncertain who will win the CPU-war in '99 but the days that Intel could charge $1000 for their high end CPU (remember the PII-300) are definitely over…