Sounds about the right ETA.......for the arrival of the E6 Diegos ;)Quote:
Originally Posted by largon
Printable View
Sounds about the right ETA.......for the arrival of the E6 Diegos ;)Quote:
Originally Posted by largon
so how are they likely to clock?
flytek,
I believe s7e9h3n is the only one yet who's had pleasure punishing E6 SDs. I myself might be able to tell something about it - in about 2-3 weeks. I already booked some model 146 samples...
Edit: Opterons i received were E4 after all. Oh well... ;)
Is anyone else thinking, "NOOOOOOOOOOOO!"
I am, my birthday is in 5 months, and I can only build a new computer then. I can only hope that this code change will lead to improvement, so I can still overclock the ownteron.
THe switch to rev6 isn't necessarily a "bad" thing, just consider it a "different" type of overclock :DQuote:
Originally Posted by manomanx2
Also, I've many times reffered to the new E6 1mb l2 cache cpu's as "E6 Diegos". Please, let me just say that I use the "Diego" name as a habit and the new 1mb cache cpus aren't in fact based on the E4 "San Diego" cores. So technically, the term "E6 Diego" is impossible :p:
stephen, politicians use words like "different". Any chance of cuttting out the politics and telling us what different means.
Paranoia reigns. AMD Conspiracy to take the well-overclocking chips away! :)
It seems Venice will also get a replacement. New OPN for the 512kB L2 core:
ADA3?00DKA4CG
I think it's a bit strange that they would end the big Opteron thing going on. They have made a big success from day one, AMD saw what was happening. All OC'ers completely loved them. So what did AMD do? They removed the 3000+, and changed the prices so the 144 and the 146 became cheaper (oct 31) than the 3200+ even though Opterons are more expensive to produce (1 MB cache).
In short, they made the opterons even more good looking after it became the overclockers dream, why would they suddently change again?
Mats,
They probably get better yields from E6 silicon than E4. Better yield -> more working cores -> more $$$ (-> price drop?)
AND
better silicon -> better OC
thank god.
Don't get too excited... These are pure speculations.Quote:
Originally Posted by manomanx2
More likely, it's a shift in production process. It's easier and more cost effective to produce one basic type of cpu and "lock/unlock" certain parts of the core to produce the different models ;)Quote:
Originally Posted by largon
The E4 Opterons will become cheaper since AMD will eventually stop producing that core, therefore they get thrown in the "clearance bin" :p:
Most of you don't realize that overclockers only make up 1-3% of overall sales. AMD doesn't care if their CPUs OC well or not, they will do whatever gets them the most profits.
If they did figure out a new process that improves OC and is cheaper or the same as the current process, then good, all new CPUs will OC better overall. But if the demand for these low end Opterons has exceeded supply (which it has), it's possible that the code change will mean lower OCs. They could of tweaked the process to produce higher yields, but lower OCs overall, for example.
Remember, Opterons are meant for SERVERS. The volume of sales for the server market is WAY lower than the desktop market. This is why Opterons are so rare compared to normal A64s, hence the price hikes.
Look back, and it's the same story when AMD introduced the Winchester, Venice, and even X2. All of the early samples were generally better OCers than the latter ones. Most 3700+ Diegos and X2 4400+ these days have a hard time reaching 2.6Ghz stable.
This is where all started :) .....http://www.vr-zone.com/?i=746&s=3
Imo, every cpu's week or Cxxx codes are always different in term of OCability...
That was a very interesting link, thanks for sharing :)
One thing worth noticing is that the multiplier is pre-programmed by the speed grading machine through the pins, not by the die or the package.
WTF? This means there's no laser cut or any other "hard" programming. I guess the guy who thought he could unlock A64 with the JTAG interface was right.
So, opteron 154 = fx57 :)
The OC'er market is not even an afterthought for AMD. However in the server market we have a chance to have a bigger impact than the general desktop market.Quote:
Originally Posted by HKPolice
How funny would it be that the 939 Opteron turns out to be a stunning sales success (based on non-server sales) and AMD goes on to claim they're winning a bigger share of the server market which in turn gives them more sales leverage with corporate sales. A brilliant plan. ;)Quote:
Originally Posted by HKPolice
Not entirely... Notice that ALL A64 cpus have "assembled in malaysia" written on them, yet the article you linked was from an AMD plant in Singapore. It seems pretty stupid to assemble the CPUs in malaysia then ship it to Sinapore to be speed graded, packaged and then shipped to the rest of the world. I know that the actual silicon is made in Dresden Germany, but couldn't it be assembled in Singapore as well as Malaysia?Quote:
Originally Posted by Dumo
Hence why there are different CXXX codes maybe? Each one representing a certain stepping from a certain assembly plant.
As you all know, certain countries make better OCing CPUs overall. IE Made in China P4s are the worse, while Made in Malaysia is usually the best.
Read it all :) ...Quote:
Originally Posted by HKPolice
http://img391.imageshack.us/img391/8...shot0106ua.jpg
There are probably varying calibrations in the test equipment. A plant in Malaysia may test a core to be 4000+ while the same chip could be tested as only 3700+ in another plant.Quote:
Originally Posted by Dumo
That would explain the varying OCs from different C codes
It could be also possible they drop the 144 and the 146. I hope that doesn't happen :( But it could also the fact they run out of E4 which seems to be more logical.
I ordered mine over an week ago in Germany, but they are out of stock...
Nearly impossible to get one :(