Rumour: Bulldozer 50% Faster than Core i7 and Phenom II.

[IntelUser2000]

Congrats on making Elite Member! :thumbsup:

You were long over-due for the recognition IMO, and well deserve the distinction.

Achievement unlocked: Anandtech Elite

Just kidding. Thanks guys.

 

[Martimus]

Those are awful names.

I actually like the names. Simple, and easy to differentiate between the model numbers.

 

[piesquared]

yep, great naming scheme actually.

 

[gdansk]

yep, great naming scheme actually.

"It is just plain common sensical." It was about time they dropped the name like Phenom II X4 955. The 955 part made sense, the II and X4 monikers are a little excessive.

 

[HW2050Plus]

"It is just plain common sensical." It was about time they dropped the name like Phenom II X4 955. The 955 part made sense, the II and X4 monikers are a little excessive.

The full names are a little longer though:
AMD Vision Black FX8000
old namings:
AMD Phenom II X4 955 Black Edition

But you are right the "II" "X4" "955" is too much.
Now you have one name ("Vision Black"), one prefix ("FX") and one number ("8000").

 

[dmens]

If the gate delay metric quoted was FO4, then its a process independent metric. Prescott was supposedly at 12.5 FO4 delays at 90nm. Penryn was 18 or something.

The "double-pumped" ALU's were clocked twice as fast as the nominal core frequency, which made that logic ~7 FO4/cycle. :awe:

OK so it wasn't static CMOS, but damn it was fast.

 

[Triskain]

Thus a 4GHz Zambezi at 95W is not unlikely.

That is my expectation as well.

nvo from the Semiaccurate Forums, who supposedly has access to insider info concerning BD says:

Well, from what I've heard, target frequency for the default P-state at 95W TDP is 3.0GHz to 3.3GHz, and 3.7GHz to 4.0GHz using TurboCore 2.0 with all cores active.

Source

OK so it wasn't static CMOS, but damn it was fast.

It also had more leakage than a punctured high pressure line.

 

[Skurge]

I actually like the names. Simple, and easy to differentiate between the model numbers.

They are simple, but to many numbers. Shouldve kept it like the old FX's just 2 digits. Like FX55.

 

[IntelUser2000]

Who knows how marketing dept will mess up the naming with 20 different SKUs per power band?

 

[dmens]

It also had more leakage than a punctured high pressure line.

Heh, it wasn't leakage that made prescott so power hungry.

 

[sawtx]

They are simple, but to many numbers. Shouldve kept it like the old FX's just 2 digits. Like FX55.

I like how they are using the first number to designate core count, hopefully the second number would be the generation, and last two for a specific chip designation.

 

[bryanW1995]

Heh, it wasn't leakage that made prescott so power hungry.

no need to defend prescott, especially when you guys are on such a roll these days. you don't see amd defending 2900xt do you?

 

[Phynaz]

no need to defend prescott, especially when you guys are on such a roll these days. you don't see amd defending 2900xt do you?

He's not denying it ran hot, he's stating it wasn't transistor power leakage that made it run hot.

Of course it would have been helpful if he completed the statement and say what did make Prescott run hot.

 

[Skurge]

I like how they are using the first number to designate core count, hopefully the second number would be the generation, and last two for a specific chip designation.

Like FX6150?

I just guessed.

 

[Castiel]

They should have named the 6 core BD FX2600 just to piss off intel

 

[podspi]

They should have named the 6 core BD FX2600 just to piss off intel

Better yet, FX2601 ()

 

[drizek]

They should come out with an FX 5800 and make it run slower than Conroe, just to piss off nvidia.

 

[athloln]

They should come out with an FX 5800 and make it run slower than Conroe, just to piss off nvidia.

Mandatory

 

[Soleron]

That is my expectation as well.

I would not be surprised if AMD ultimately reaches a retail SKU clocked at 4.5GHz at some point on 32nm, especially if they keep the 140W TDP SKU going at 32nm.

Is the "architecture is capable of 3.5GHz" statement made during ISSCC lowballing it then?

 

[jvroig]

Is the "architecture is capable of 3.5GHz" statement made during ISSCC lowballing it then?

Considering that they already exceed this with their current 45nm parts, and that this new arch is designed for high clock speeds, then yes, this would be a reasonable expectation.

 

[itsmydamnation]

Is the "architecture is capable of 3.5GHz" statement made during ISSCC lowballing it then?

not really they probably have speed/perf/power target.

 

[Idontcare]

Is the "architecture is capable of 3.5GHz" statement made during ISSCC lowballing it then?

"Capable of" - definitely.

"Will it" - that is highly dependent on GloFo's 32nm process tech...

 

[IntelUser2000]

They said 3.5GHz+ at ISSCC. So if you want to say something ambiguous, but confident this would be the way to go. You can't figure out what the exact clock speeds will be from this.

 

[Ares1214]

Is the "architecture is capable of 3.5GHz" statement made during ISSCC lowballing it then?

Of course its capable. Even 45nm is. The important part is if it stocks at 3.5 GHz, will it:

1. Scale as well as SB when OC'ing

2. OC as high as SB

3. Maintain reasonable temperatures at these clock speeds

Considering Intel is in it's 2nd 32nm gen, odds are they are getting better batches than AMD will be. Seeing the fantastic results SB is getting with the 3 categories above, im fairly skeptical SB will hit 5 GHz on air. Scaling depends on the arch, and so does temps, but im going to guess 5 GHz will be rather elusive for AMD.

 

[Arkadrel]

They should have named the 6 core BD FX2600 just to piss off intel

Better yet, FX2601 ()

had me lauffin. +1 karma points for that Castiel.

I vote they go with FX2600 too
(mostly because I dont like the FX2000,4000,8000,16,000? theme)

sounds like to much to have a 16,000 tag.

Wonder if they couldnt just do a FX-2k, FX-4k ect.
Think it sounds better, more memorable too that way.

then they could spell their cores as "Kore" 's.