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: Pravin Kamdar who wrote (11507)	10/5/2000 12:22:25 AM
From: Joe NYC	Read Replies (1) \| Respond to of 275872

Pravin, I am holding my breath and will breathe much easier when I see the first high speed Mustang shipping in volume. Ditto for the mobile Mustangs. Agreed This roadmap doesn't have the server based Mustang, but it is not very encouraging that neither Morgan nor Palomino will arrive in Q4. I was hoping for at least December introduction: watch.impress.co.jp On another note, is Intel's C0 stepping Coppermine-T? I think the roadmaps showed a version between regular Coppermine and Toulatin (sp?) - .13u part. According to the roadmap, the Coppermine-T was supposed to arrive in late Q4 and be a shrinked version. Well, the stepping C0 is a slight shrink, and it is Q4 (early, not late), so theoretically C0 can be Coppermine-T. If it is, then AMD has a lot of breathing room left for some time. Joe

To: Pravin Kamdar who wrote (11507)	10/5/2000 12:29:22 AM
From: Chung Lee	Read Replies (1) \| Respond to of 275872

.....AMD will have 1.5 GHz in Q1 It would be very dangerous to assume this. The changes that AMD will be implementing to get to 1.5 GHz will provide many opportunities for production glitches. 1. AMD has delivered what it promised in the last 3Qs at least, it has built trust the old fashion way, results! 2. Every promise is dangerous, but without promise/roadmap, investors will be even more in the dark, every 50/100 Mhz increment release is potential danger, but also potential oppertunity and sweet rewards. I am holding my breath and will breathe much easier when I see the first high speed Mustang shipping in volume. Ditto for the mobile Mustangs. If one waits to invest until one can see the whites in the eyes, one might miss a multiple bagger, with more misfire from the Intel camp (seems to deliver pretty consistently lately), and with Mustang shipping in volume, AMD will be in Niceguy heaven, I think Niceguy said $250. As for me, I am placing small bets every couple dollars AMD goes down, I have stow away quite a few shares which at this point not showing much :-(

To: Pravin Kamdar who wrote (11507)	10/5/2000 1:50:07 AM
From: Charles R	Respond to of 275872

Pravin, <It would be very dangerous to assume this. The changes that AMD will be implementing to get to 1.5 GHz will provide many opportunities for production glitches. I am holding my breath and will breathe much easier when I see the first high speed Mustang shipping in volume. Ditto for the mobile Mustangs.> All good points but the risk here, other than DDR platform risk, is pretty low. Chuck

To: Pravin Kamdar who wrote (11507)	10/5/2000 9:44:07 AM
From: pgerassi	Respond to of 275872

Dear Pravin: Given that Intel likes to tout P4 MHz advantage (ship 1.5 GHz in Nov 00), why doesn't AMD point to Kyrotech 1.6 GHz (Tbird?) in Nov 00? Since a Kyrotech solution adds about $300 to $500 and i850 plus 256 MB of RDRAM costs about as much over AMD 760 plus 256MB of DDR PC2100, the game would be equal at cost (assuming that the retail price of 1.5 GHz P4 is the same as 1.2 GHz Tbird) but, IMHO Tbird would just about destroy P4 in all but a very few benchmarks. BTW, I read some speculation on the non-moderated thread that P4 (internally named P68) was just an improved and longer pipelined double speed back end (xx8 (2nd try? or was P3, xx7?)) with a P3 decoder unit at normal speed (P6). If this is true, and looking at the diagrams for the pipeline stages for x86 decode and functions blocks in Intel IDF presentation it appears that it could be true, then any multitasking code or real time apps, where heavy context switching occurs, will cause the P4 to run the same as a P3 at half the clock rate of the P4. Thus, the benchmarks testing a 1.5 GHz P4 using heavy server type loads, would run like a P3 at 750 MHz (probably somewhat faster due to FSB bandwidth and short bursts in tight loops occur (the best code for P4)). Given the P4 benchmarks leaked to JC, it appears that the trace cache, doubled pipeline, and QDR FSB, gives a 40% improvment to overall throughput in normal real world office loads, a 90% improvement to loads with tight loops that work on massive data, and a 100% improvement on loads that can take advantage of SSE2 over x87 to the original speed P3. Thus, a QADG (quick and dirty guess), would put a 1.5 GHz P4 at a similar performance level to a 650 MHz Tbird on server loads, a 1 GHz Tbird on office loads, a 1.2 GHz Tbird on tight loop loads, and a 1.33 GHz Tbird on SSE2 loads. Without knowing the amount of performance improvement Mustang derived CPUs have over Tbird derived CPUs, it is harder to figure how P4 would compare to Mustang/Palomino/Morgan. I assume that they would outperform Tbird by at least 5% (if not more (WAG)). Thus the entire Palomino line could be higher in performance than the top P4 over all these type of loads. The only loads for which P4 would come on top, would be any load that does lots of prefetchable data movement where the FSB is the bottleneck. Pete