brisbane dual core 65nm

Printable View

Show 100 post(s) from this thread on one page

12-08-2006, 07:46 PM
Asazman

Quote:

Originally Posted by frankR

I'm all for seeing a Brisbane running 3.5+ Ghz w/stock volts on air.:)

If it was availible for 939, I'd be even happier.

me too, cause i could get one haha
12-08-2006, 09:09 PM
MAS

agree, it's a pure fake
just remixed metro.cl's SS
but i saw your reaction -)
12-08-2006, 09:16 PM
keiths

Quote:

Originally Posted by Turtle 1

Ed is uasually pretty right on here's what he has to say.

http://www.overclockers.com/tips01073/

Ed had a few other things to rant about.

http://www.overclockers.com/tips01074/

You must admit Ed has a few very good points.

You have Hector's "it's not about core anymore" direction he's taking AMD to thank for their current situation. The disconcerting part is he's still pushing his mantra and AMD's roadmap reflects it. Intel's cores are projected to be faster for the forseeable future but in Hector's opinion, it doesn't matter. Time will tell whether the market agrees with him.
12-09-2006, 07:09 AM
Turtle 1

Well I think so far Hector has done a good job. If however K8L doesn't match C2D in performance. I believe that placing AMD's future on fusion is questionable. But thats a long ways off.
12-09-2006, 07:39 AM
savantu

It is impossible for K8L not to equal C2D.Only if they screwed up ib a big way which doesn't look very likely with the conservative aproach they've taken.
12-09-2006, 07:46 AM
Shintai

Quote:

Originally Posted by savantu

It is impossible for K8L not to equal C2D.Only if they screwed up ib a big way which doesn't look very likely with the conservative aproach they've taken.

Just curious, how is it impossible not to equal C2Ds?
12-09-2006, 08:01 AM
MAS

Quote:

Originally Posted by savantu

It is impossible for K8L not to equal C2D.

AMD officially showed quadcore opteron perf. rating

opteron is (u can measure yourself) 14 percent faster than Woodcrest (imo both at the same frequency)
http://img118.imageshack.us/img118/9...uadperfxr5.gif
12-09-2006, 08:02 AM
couppi

Quote:

Originally Posted by Shintai

Just curious, how is it impossible not to equal C2Ds?

K8L is teh greatest evar. Didn't you get the memo?
12-09-2006, 08:44 AM
Silvermirage

Quote:

Originally Posted by brentpresley

That graph is complete marking BS. Some me some numbers.

Until then I take ANYTHING that AMD and Intel say as complete lies.

But mac tells the truth. :P
12-09-2006, 09:08 AM
Shintai

One also have to wonder how the 2.8Ghz Opterons can be just as fast as the 3Ghz C2D Xeons. Kinda abit rigged benchmarks, if benchmarks was even included.

So, a 2.67Ghz C2Q smacks 3Ghz FX74 with regular DDR2 around. Yet a 2.8Ghz Opteron with ECC memory is just as fast as the Xeon 5160, that got a 25% faster FSB, dual FSB, quadchannel memory and a 10% higher clock. FUNNY!

Reminds me of a quote from the nVidia marketing executive.
"In marketing, we cannot speak the truth. We have to make the truth more understandable."
12-09-2006, 09:25 AM
MAS

Quote:

Originally Posted by Shintai

One also have to wonder how the 2.8Ghz Opterons can be just as fast as the 3Ghz C2D Xeons.

maybe in certain server applications

nevertheless Xeon 5355 at 2,66GHz is 13-14percent slower than 2,5GHz K8L Opteron (2,5GHz is the max. clock for K8L Opti, see the table up) - AMD said
12-09-2006, 10:15 AM
Shintai

Quote:

Originally Posted by MAS

maybe in certain server applications

nevertheless Xeon 5355 at 2,66GHz is 13-14percent slower than 2,5GHz K8L Opteron (2,5GHz is the max. clock for K8L Opti, see the table up) - AMD said

Ye sure, and Apples new HW is 2-5x faster than their own PPC, that again was 2-3x faster than their X86..ohh wait.

Its the same benches/estimates that are already wrong. Nothing but fancy PR.
12-09-2006, 10:41 AM
Turtle 1

Quote:

Originally Posted by savantu

It is impossible for K8L not to equal C2D.Only if they screwed up ib a big way which doesn't look very likely with the conservative aproach they've taken.

I think that K8L has to do more than = c2d its has to best it. As Wolf/York will be a bit faster than conroe . I will be conservitive and say 5 % in clock to clock and in apps that use sse4 i would say 15% clock for clock.

Than you have to consider the clock rate Wolf@ 4ghz with 6md shared cache. and York @ 3.7 ghz. with 12mb shared catch between all four cores as thats what has been reported. I believe Scali2 layed this out for you guys already and he was quit good at it.
12-09-2006, 11:03 AM
Turtle 1

Quote:

Originally Posted by PallMall

I don't think we have to doubt about the performance of amd's nextgen products,

it's kinda like Nvidia vs ATI,

ATI is always later, but they do end up with a better product right?;)
AMD is quite the same to me, i have a nice E6600 now,but i'm sure the equivalent will prolly beat it with 1 leg cutted of :)

But hey when the new stuff arrives, i already went quad lol

Well I won't argue that point with you. I hope you are correct as it will benefit all of us if true.

But would you mind laying out for us just exactly how AMD will make up its present clock for clock deficit. You must remember and use logic . That dothan at the same clock was = or > than AMD64. Now c2d has all the same improvements that K8l has plus it is 4issue with more pipes and wider pipes.

I just don't see were AMD's IMC can make up this deficit. Now use logic and not your heart in your presentation of the facts.
12-09-2006, 12:05 PM
informal

Quote:

Originally Posted by PallMall

The logic would say,

--look at the K8L specs
--look at the launch clocks(lower then conroe ;) )
--look at the launch date

al these facts, just drives me to say that it will be better no?:)

Right on target ;)
12-09-2006, 01:20 PM
Turtle 1

Ya if you say so. I really thought maybe you could point out the architectural advantages.
What in the specs is better than what the wider pipe 4 issue C2D core.

Lower clock rates only tells me that AMD can't scale there cpu as high.

Whats the launch date have to do with Core Architecture?
12-09-2006, 01:28 PM
Turtle 1

Ok . Your right. I was just hoping you could tell me something concrete.
12-10-2006, 02:22 AM
Lightman

I just found NEMO!!

http://www.scan.co.uk/Products/Produ...oductID=445618

Quote:

This is the New 65nm Energy Efficient Version

Do more in less time with true multi-tasking
Increase your performance by up to 80% with the AMD Athlon™ 64 X2 Dual-Core processor.
Work or play with multiple programs without any stalling or waiting. Dual-core technology is like having two processors, and two working together is better and faster than one working alone.

But on the specification page they say 90nm:slapass:
Which one is truth?? They are open tomorrow so I will try to find out.
Anyone willing to try this baby??
12-10-2006, 02:32 AM
MAS

"This is the New 65nm Energy Efficient Version" must be "This is the New 65Wt Energy Efficient Version", thougth it's no so new

there are no retail Brisbanes today, only oems
12-10-2006, 02:53 AM
zir_blazer

Did you notice than it is an ADO4200CUBOX? Retail 65nm A64X2 are ADOxxxxDDBOX.
12-10-2006, 03:01 AM
Lightman

Quote:

Originally Posted by zir_blazer

Did you notice than it is an ADO4200CUBOX? Retail 65nm A64X2 are ADOxxxxDDBOX.

Yes I noticed, but every other EE CPU on their web site is described as 90nm Windsor Core. This is only one with '65nm' in overview! We know that this data are putting in system people without knowledge, so best and simplest approach is to copy as much as possible from older product overview.

Besides this product was added recently. I was looking one week ago on froogle with same key words without any result.

I have shadow of hope ;) .
12-10-2006, 03:13 AM
Lightman

Never mind!!

AMD announced only 4000+/4400+/4800+/5000+ 65nm versions.
12-10-2006, 03:24 AM
Shintai

Quote:

Originally Posted by PallMall

The logic would say,

--look at the K8L specs
--look at the launch clocks(lower then conroe ;) )
--look at the launch date

al these facts, just drives me to say that it will be better no?:)

Not to burst your bubble.

We looked at K8L specs, it contains only a small handful of Yonah->Conroe improvements. That gave 0-20%, and the major improvements was speculative cache, 4 issue wide, from 1 to 3 SSE ports, double cycle to single cycle SSE, from 2 to 3 ALUs, bigger buffers, macro fusion and so forth. Specially the SSE is funny, its one of the big hopes since K8 will get twice potential SSE! However, Conroe got 6 times the SSE potential that Yonah got. And Yonah performed around the same as K8.

Look at the lower clocks, yes, did you also notice the regular K8 quadcore lower clocks? And since AMD seems to be limited for 65W at 2.6Ghz for 65nm. Its a no brainer that its a heat limit.

Look at the launch date, that makes no sense. Its like saying presler/smithfield etc would beat X2 just because it was launched later. It was miles away from that.

Now get back to reality!
12-10-2006, 03:37 AM
mzs_biteme

Quote:

Originally Posted by MAS

"This is the New 65nm Energy Efficient Version" must be "This is the New 65Wt Energy Efficient Version", thougth it's no so new

there are no retail Brisbanes today, only oems

"Show me" a retail or OEM or ES version of Brisbane, and I'll give you a cookie.... What...... You can't.... No cookie for you!!!!! :slapass:
12-10-2006, 03:46 AM
MAS

ask Dell for Brisbane - it has it
Dell needn't cookie ))

Show 100 post(s) from this thread on one page