AMD to increase performance by 40%
According to Hans Deppe, vice president and general manager, AMD Saxony LLC & Co. KG, 65nm process at FAB36 is expected to provide a 40 percent improvement in transistor performance. Actually, AMD and IBM already announced this on Dec 6, 2005. Transistor performance is measured by switching speed. It seems that 4GHZ Opterons are not that far away. In December 2004, AMD and IBM announced 24% transistor performance increase with DSLSSOI, and Opteron speed went from 2.4GHZ to 2.8GHZ (3GHZ soon).
AMD has ramped FAB30 to 30,000 wspm, and is ramping FAB36 to 17,000wspm by the end of 2006, reaching 20,000wspm by mid 2007. Chartered FAB7 is on schedule to ship by mid 2006. It seems my previous estimate was a bit too conservative.
10 Comments:
Sharikou, today's report suggests that fab36 parts are only now hitting the market. How does that change your 1q-06 projection for capacity? Does the fab30's increased productivity offset this decrease?
Anyways, the 65nm transition if it happens to the high end parts by 2H06 will be a killer triumph for AMD.
Actually, today's report stated that FAB36 started shipping for revenue in March 2006. So that cut my estimate my FAB36 contribution by half. However, it is also reported that FAB30 increased by 20% to 30000wspm. So my 20% unit increase estimate holds.
According to today's report, AMD is ramping FAB36 to 17000wspm by the end of year, adding Chartered FAB7, we get over 20000wspm. This 1.5X FAB30 capacity even at 90nm.
The wafers from Chartered will be 90 nm not 65 as the Inq hinted. Which is OK and is more reasonable. It would not be reasonable for another company to be producing AMD's leading products. The thing to realize is that AMD's 90 nm products are eminently capable of competing with Intel's 65 nm products in performance. The only question is that of production capacity which has been addressed with the anouncements so far. Oh well, there goes the near term availability of the "Conroe Killer". Which is OK, it is not needed critically. Conroe will not be available when Intel says it will. Besides the Opteron is a proven product that has been around for three years now. Conroe will be an unproven new kid on the block. Companies will be hesitant to commit production systems to an unproven new product. The "Conroe Killer" will have to wait 'til mid 07 when the 2nd gen strained silicon on 65 nm comes out with the 40% speed advantage. Note I said "Killer" not "Competitor". Intel will be in sad shape at that point. I find the AMD enterprise videos interesting, Balmer and Mike Hurd pitching their products on the AMD web site. They are just pitching their respective companies products and the videos are not that interesting. What is important is that they are the top dogs. For them to be on the AMD website is simply amazing. I would guess that they realize that it is over for Intel, at least for the near term and probably for ever. For them to appear on the AMD website is simply amazing and indicates that nobody is afraid of the "Big Bad Wolf" anymore. My how things have changed.
humm this is completely not related to this thread.. but this is what i found in the forum of toms hardware guide.
Conroe's Super Pi benchmark
http://www.xtremesystems.org/forums/showthread.php?t=95021
the interest thing was that Conroe finished calculating Pi in 21 secs, which is a little impossible for a processor that has 99.9% similarity to Yonah.
My poor poor Opteron 146 @ 3ghz gets 27 seconds.
Super PI 1M calaculation fits nicely in the 4MB cache of the Conroe, so if Conroe can do it fast, I won't be surprised. Go up to 10M digits, it will be a whole different story.
I was being slightly sarcastic. I'll take my 3ghz Opteron 146 in Windows XP 64 over a conroe any day.
Super_Pi 1M does not fit into 4MB of cache. There is no 10M option, biggest is 32MB option.
Here a link to Conroe and then some overcclocked AMD K8's
http://vic.expreview.com/attachment/1144486148.jpg
http://www.xtremesystems.org/forums/showthread.php?t=70190
Comments ?
Whoever got that opteron what kind of cooling do you have
Would you really take a 3GHz Opteron 146 over Conroe in 64-bit Windows Xp any day?
I wouldn't, because I think a 2.67GHz Conroe will beat it.
Post a Comment
<< Home