2.0GHZ Barcelona benchmarks released

[Apokalipse]

2.0GHZ Barcelona benchmarks have been released by Anandtech This is the end result: AnandTech: AMD's Quad-Core Barcelona: Defending New Territory

Opteron 2350 (2GHz) equates to Xeon clock speed of:

General applications:
WinRAR 3.62 - 2.7 GHz
Fritz Chess engine - 1.8 GHz

HPC applications
Intel optimized Linpack - 1.9 GHz
3D Applications 3DS Max 9 - 2 GHz
zVisuel 3D Kribi Engine - 2.33 - 2.4 GHz
zVisuel 3D Kribi Engine (AA) - 2.4 GHz

Server applications Specjbb - 2.4 GHz
MySQL - 2.33 GHz

Basically, Barcelona is better than Clovertown clock for clock in all but the Fritz Chess engine, and highly optimized Intel binaries, and 3DS Max 9 which it equalled

2.0GHZ is only the initial release of Barcelona. It will be ramped up very quickly in the coming months. And given the HyperTransport bus, which is even faster with the new Opterons, it should scale even better when clock speeds are increased.
However, Barcelona still does not utilise Hypertransport 3.0, and uses registered ECC memory, which does create a bit of a performance hit

Agena will include HyperTransport 3.0, and will use regular DDR2 RAM at DDR2-1066 speed. So I am expecting Agena to scale even better. Also, it will be interesting to see how the K10 performs when running games, which is one of the things AMD's CPU's tend to be better in.

 

[NosBoost300]

good news to read this!!!!! can't wait to see the phenom!

 

[gurusan]

Nice, eager to see how they perform in games...and how they clock (should be pretty good considering the power usage)

 

[maroon1]

I did same calculation, the average is only 2.25GHz

This means than Barcelona 2.0GHz is on par with Xeon 2.25GHz (according to anandtech benchmark)

This means that K10 is just slightly faster than Core 2 Duo

Based on those benchmarks, Q6600 would be faster and it just half the price of Barcelona 2.0GHz

 

[maroon1]

it will be interesting to see how the K10 performs when running games, which is one of the things AMD's CPU's tend to be better in.

I have to disagree with you

E6750 outperforms 6400+ by about 10% in gaming

In normal applications, E6750 outperforms 6400+ by about 5% on average

In synthetic benchmarks like 3dmark, E6750 performs worse

So, no AMD CPU's don't tend to be better in gaming

 

[gurusan]

I did same calculation, the average is only 2.25GHz

This means than Barcelona 2.0GHz is on par with Xeon 2.25GHz (according to anandtech benchmark)

This means that K10 is just slightly faster than Core 2 Duo

Based on those benchmarks, Q6600 would be faster and it just half the price of Barcelona 2.0GHz

Interesting...and isn't Penryn supposed to be around 10% more efficient than the current C2D processors...putting it in the same performance bracket as Barcelona/Phenom?

It will be interesting to watch this...I am rooting for AMD though as I like to see each company taking turns bouncing up on top in the enthusiast market.

 

[maroon1]

Interesting...and isn't Penryn supposed to be around 10% more efficient than the current C2D processors...

Penryn is around 5-10% faster than the current C2D processors per clock, but in SSE4 optimized application it outperforms the current C2D by up to to 100%

K10 might perform slightly faster than Penryn per clock, but Penryn have the ability to reach higher clocks easily because of 45nm die shrink. In SSE4 optimized applications, there is no doubt that Penryn is going to be better.

By the way, Performance per dollar (and performance per watt) is far more important than performance per clock. We need to wait and see how Intel and AMD is going to price their upcoming processor

 

[maroon1]

Here are some other benchmarks for Barcelona
AnandTech: AMD Phenom Preview: Barcelona Desktop Benchmarks

This benchmark is comparing Barcelona with AMD K8, and both are running at the same clock speed

It is faster but not much faster.

......................

It seems that AMD Barcelona is a big disappointment to AMD fanboys.

 

[NosBoost300]

ouch that hurts

 

[Apokalipse]

I have to disagree with you

E6750 outperforms 6400+ by about 10% in gaming

In normal applications, E6750 outperforms 6400+ by about 5% on average

In synthetic benchmarks like 3dmark, E6750 performs worse

So, no AMD CPU's don't tend to be better in gaming

It wasn't a comparision between K8 and Core 2. It was a comparision between K8 gaming benchmarks, verses K8 benchmarks in other areas.
K8 does better in gaming than a lot of other areas.

Penryn is around 5-10% faster than the current C2D processors per clock, but in SSE4 optimized application it outperforms the current C2D by up to to 100%

how many SSE4 applications can you count?
AMD are going to make SSE5.

K10 might perform slightly faster than Penryn per clock, but Penryn have the ability to reach higher clocks easily because of 45nm die shrink.

How do you know? The clock speeds of Barcelona are going to be ramped up in the coming months.

In SSE4 optimized applications, there is no doubt that Penryn is going to be better.

When there are applications that actually utilise it. By that time, the SSE5 processors from AMD should be out.

By the way, Performance per dollar (and performance per watt) is far more important than performance per clock. We need to wait and see how Intel and AMD is going to price their upcoming processor

and it seems the Barcelona processor has the advantage.

Here are some other benchmarks for Barcelona
AnandTech: AMD Phenom Preview: Barcelona Desktop Benchmarks

This benchmark is comparing Barcelona with AMD K8, and both are running at the same clock speed

It is faster but not much faster.

This is when paired with DDR2-667 (which is ECC BTW), which does cripple one of their big advantages: the onboard memory controller.
The fact that the K8 and K10 have such a fast memory controller is what allow them to scale well with clock speed. But they also are affected significantly by slower RAM.
Plus, you seem to be cherrypicking here.

It seems that AMD Barcelona is a big disappointment to AMD fanboys.

Now this is jumping the gun.
I am guessing you are saying that because you want it to be true.

Also, you are forgetting that this is Barcelona, and not Agena, which doesn't have HTT 3.0, and has to use ECC RAM.
Barcelona is designed more for multi-socket systems. And it scales much better when you start adding cores into the system than the Xeon's.

Agena will be oriented to the desktop: it will have HTT 3.0, and use DDR2-1066