i was wondering if any of you might know how i can find a t-bird 1200 mfg. in Dresden, germany rather than one made in austin tx.??
It probably doesnt matter but i want one with copper interconnects and a 266 fsb.
I spoke to someone at AMD today and he said,
" Copper interconnects only allow us to hit higher speeds later on.
This will become more important on our 64 bit processors when we need to hit very high speeds".
and other than better conductivity doesnt offer any performance increases.
well thanx if you know where I can get one.
It's not incorrect, the Al Athlons can only clock so high and thus all after a certain clock all are Cu. Actually I think after 800 or 850 all Athlons are Cu. That's kinda old info so it might have changed some by now. Anyone own a Al 900mhz (non-OC) Athlon? Correct me if I'm wrong, but when Tom did that vapochill OC test, weren't the Al cores only able to go up to 1150?
I decided to go to the source on this one and so i spoke to a couple of engineers at AMD heres what they said:
The 266 FSB parts are being manufactured and sold now. However, they may be
hard to find until they filter out into the retail channel. As for
different versions of the part? I would have to say that this is incorrect.
We do have 2 manufacturing facilities (Austin Tx, and Dresden Ger.). The
Dresden facility is equipped to manufacture processors using copper
interconnects. The interconnects are the things that internally connect the
transistors together. While the copper interconnects do offer better
conductivity than the aluminum, they do not change the performance of the
processor. Copper interconnects only allow us to hit higher speeds later
on. This will become more important on our 64 bit processors when we need
to hit very high speeds. In any case, there are not different versions of
the T-Bird and there are no price differences on our side. If resellers try
to market and price these parts different, I cannot say. But from our
perspective, a 1.2Ghz T-bird is a 1.2Ghz T-bird. The only difference will
be the FSB speed 200 vs. 266.
Hope this helps.
I hope this helps. If you have any other questions, please feel free to
contact me. If I have misunderstood the intent of your email, please
explain further and I will try to answer your questions.
Supervisor / Apps. Engineer
Well i guess i was only assuming that the 1000 mhz and larger chips are the only ones incorporating the copper. and as the response indicated the copper is there for future speeds and does not offer any performance increases at all.
i suspect that if you had a choice between a copper or aluminum and were planning to overclock it would make sense to get a copper one if you could.
Which is exactly what im planning to do, so i have been in correspondance with AMD again today and buy monday i will have some definitive answers regarding the manufacturing / assembly pts.
my 800 says it was assembled in malaysia.
i was told ( not by amd) that all athlons/t-birds under 1000 mhz were using aluminum.
I will let ya know monday afternoon.
IBM initially found that you get an extra ~30% clock by using Cu instead of Al. The conductivity of Cu isabout 30-40% greater than microprocessor grade aluminum (which is actually an alloy with some Cu). Clearly the conductivity offers a pretty linear increase in processor clock. I'm pretty certain the current upperend TBs are beyond what they could do with Al. I'm also pretty certain I remember reading from Tom that the Al processors were only Durons and slower TBs. If you look at the max clocks Tom was able to get in the "Processors On The Rocks VOL. 2" article it excludes the possibility of Al @ 1.2ghz. Here's why, the max clock achievable with a Cu was 1485mhz and 1066 with the Al. Take the max Al clock, multiply by 1.4 and you get the Cu clock. Current Cu athlons are close to maxed out and that's why you have a new process comming along. I'm guessing that the 40% is achieved because the implementing Cu is more mature than when IBM first developed it. I suppose also the capacitive effects are even more prevelant than when IBM originally implemented it in ~300-400mhz processors.
I'd be VERY interested if you can find anyone with a Al AMD processor clocked over 1ghz. (non-OC)
All ATHLONS ARE MADE IN DRESDEN GERMANY. COPPER IS THE ONLY PROCESS DRESDEN USES. THEY DON'T DO COPPER AND ALUMINUM IN THE SAME FAB. AUSTIN USES ALUMINUM. NO ATHLONS ARE BEING MADE IN AUSTIN. DURONS AND THE LAST K6'S ARE COMING FROM AUSTIN. IF YOU BUY A ATHLON YOU GET COPPER INTERCONNECTS, WHETHER IT HAS A 200MHZ FSB, 266, OEM OR RETAIL, IT'S COPPER. END OF STORY!
Dude , are you using caps cause your shouting?
its alright we wont discuss it anymore. shhhh, i think hes gonna blow...lol
just kidding, looked like you were pissed.
anyway I told you , monday Ill talk to AMD again, matter of fact if theres any specific questions that you would like me to forward let me know.
I dont care who thinks what is made of what if there isnt any logical supporting evidence, or source of information.
Im sure this information is available thru data spec sheets and such and im sure some poor bastard already read, decifered, translated , and posted this but nonetheless i want to hear it from the horses mouth.
Some of the lower clock Athlons have been Al, I'm just not sure exactly up to what clock. Eg. I *had* a Al TB 750. At these lower clocks about 50% of them were/are Al. Austin is primarly for the purposes you suggest and you're certainly right about the process at each local. I also really doubt that there's any difference between OEM and retail, that just doesn't make sense. AMD would not be purposefully providing chips that are more overclockable for a certain market segment.
You're right that initially T-Birds were fabbed in both fabs. Once they hit a certain speed grade, which I believe was 900MHz, T-Bird production shifted entirely to Dresden. It was also right about the same time that Duron production started. If you buy an Athlon that was manufactured in Austin it's going to be < 1GHz and it won't be one that was manufactured recently.