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 : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?

To: Joe NYC who wrote (111898)	9/30/2000 1:13:37 AM
From: kapkan4u	Read Replies (1) \| Respond to of 186894

<I am a little confused about the code names. I used to like P5, P6 core names, which would imply that there would be P7 (Merced) - a departure from Intel architecture. My guess is that P68 means that the chip is partially the old generation, partially a new one. Can you elaborate a little about the original intentions?> In 1994 when P68 was conceived, the P7 name was already taken by a different CPU project. The original P7 was a RISC design inspired by Alpha's simplicity and high clock rates. In the fall of 1994 Albert Yu conducted a project review and declared that P7 design -- get this -- was not aggressive enough and did not meet the goal of 40% performance increase at the same clock speed. Of course the design was great, but the goal was dead wrong. The P7 project was killed and Albert Yu signed a deal with HP and started Merced. As far as P68 being partially old and new, your guess was right on the money. The P6 team who designed P68 considered the x86 instruction decoding to be the toughest problem, but which was already solved in P6. They did not want to mess with it. So they left the decoder logic unchanged and introduced the "inner" execution loop running twice as fast as the decoding logic. The inner loop execution was fed with uops from the trace cache. The thinking was that the trace cache faults would be rare, so the chip would be running twice as fast as P6 without changes to the decode logic. Kap