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: Paul Engel who wrote (35305)	7/30/1998 11:38:00 PM
From: Elmer	Respond to of 1573215

Re: "But the K6-3 was just a K6-2 with extra L2 SRAM and a new L2 cache controller." This is what has me puzzled. AMD has already shown they can do large sram on a processor die with their LI, so what's the big deal with adding L2? Can the L2 controller be a problem? I don't think so. I wonder why they didn't just extend the L1? That way there would be no need for an L2 controller. There would be major layout changes for sure, but adding the backside interface to the L2 and the L2 controller is no small tweak in and of itself. It'll be interesting to see how this turns out. EP

To: Paul Engel who wrote (35305)	7/30/1998 11:47:00 PM
From: Maxwell	Read Replies (2) \| Respond to of 1573215

Dr. Engel: <<It is competition for the K6-3 because Mendocino may outperform - or at least match - a K6-2.>> A PII with cache at half speed can't even match the K6-2 clock for clock when playing games with 3DNow or even chess. bootnet.com GLQuake II....333MHz-K6-2.....333MHz Pentium II 640x480.......71.3fps.........68.8fps 800x600.......71.3fps.........62.0fps 1024x768......62.2fps.........44.7fps Another similar findings hardware.pairnet.com "....While 1024x768 was slower than 640x480, it is still almost 80 frames per second! Personally, I am more than happy to give up 3 frames per second to enjoy 1024x768. From what I have been told, the Pentium II consistently takes a near 20 fps loss by going to 1024x768. 3DNow! is giving the K6-2 a dominating lead over the Pentium II when running 1024x768 with a pair of Voodoo 2 cards...." This is due to the fact that the bottleneck of the PII system is the CPU. The PII couldn't keep up with 3D calculations. A Xeon is a PII with L2 cache running at full speed is only 1-2% faster than PII www2.tomshardware.com In games like QuakeII or many of the 3D games the speed of L2 cache is not as important as in 3D calculations. Thus Xeon WILL NOT OUTPERFORM K6-2 clock for clock. A Mendocino is just a PII with integrated 128K L2 cache. Mendocino is NOT FASTER than Xeon clock for clock. Thus there is no way Medocino will outperform K6-2 running 3DNow applications. You put too much hope on the Mendocino and Celeron. With the new price cut AMD didn't react to Intel's Celeron pricing. The K6-2 ASP is still at healthy above $100 level. The bottom line is that Intel doesn't have a CPU that can do 4 FLOPS per clock cycle. Dream on. Maxwell

To: Paul Engel who wrote (35305)	7/31/1998 12:59:00 AM
From: Jim McMannis	Respond to of 1573215

Paul, "McMannis - re: " What do I think of the delay? Assuming there is one..." The K6-3 isn't working right. AMD must know it will take 6 or more months to understand the extent of the problems, make corrections and turn new silicon and then re-verify the silicon before they are ready to ship. Clearly, this is comparable to Intel's Merced slip - which was, as you recall, only 6 months - from mid/late 1999 to early 2000. But the K6-3 was just a K6-2 with extra L2 SRAM and a new L2 cache controller.l," I can't believe you are comparing the delay in the Merced to a possible delay in the delay K6-3. The Merced is a totally new chip with a dual OS capability. The K6-3 is a K6-2 with a L2 onboard and a few other goodies... The K6-3 was supposed to be out in the 4th quarter. Maybe a delay of a month. It seems like the Merced will be delayed for 6 months. Jim