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 : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?

To: peter_luc who wrote (48183)	7/18/2001 7:30:45 AM
From: pgerassi	Respond to of 275872

Dear Peter Luc: I hope that AMD optimizes both IPC and clock. Given the numbers so far, Sledgehammer will be around 2.5 GHz for each core. Thus, Sledgehammer will be 5 times as fast as Tbird at 1 GHz. IPC will be up as SSE2, and there will be twice as many 64 bit GP registers as there are 32 bit GP registers now both for integers and the same (128 vs 80 bit) for floating point. This will improve IPC as well. Athlon MP is already as fast as a P4 running at 150% of the MP's clock rate. Thus, a 3 GHz P4 only runs as fast as a Athlon MP at 2 GHz. Since IPC will be no worse and probably better than either Athlon 4, Throughbred or Barton, a 2.5 GHz Clawhammer will be 33% faster than a 3 GHz P4. P4 will not get to 4 GHz by the end of 2002. The x86-64 will be a plus and start to do to IA-64 what the PC industry did to microchannel, push it aside (that is what in the process of happening to RDRAM). Pete

To: peter_luc who wrote (48183)	7/18/2001 8:31:39 AM
From: combjelly	Read Replies (1) \| Respond to of 275872

"I remember that earlier this year AMD officials said that Hammer will be introduced at three times the clock speed of the original Athlon when it was introduced back in 1999. " That is not what I remember. IIRC, the statement was a performance level 3 times the Athlon Classic, which topped at 1GHz. Since the Hammers will have a higher IPC than the ponies, a speed of 2.5GHz is likely. Look at it this way. The ponies on 0.18 micron can likely reach 1.7GHz or so. Say that the gains from going to a bulk 0.13 micron process are fairly small, make it 20%, much of that gain coming from using a low-k dielectric than from the smaller geometries. That means 2GHz is about it on the bulk process. Now according to IBM and others, SOI should give an additional 30-35% and that means a clockrate of 2.6-2.7GHz for a Barton. The Hammers are supposed to have additional pipeline stages and a greater IPC (I'll believe this when I see it). Say it adds two stages, since the current designs have a 10 stage pipe, we should see an addition 10-20% boost over that. Assuming that there aren't any bottlenecks introduced into the Hammers, AMD should see close to, if not more than, 3GHz with the Hammers. Could be more, could be less if there are speedpath limited.