AMD: Barcelona quad-core 50% faster than Intel’s quad-core Xeon

[MAS]

source - http://www.tgdaily.com/content/view/31729/135/

There are lots of question marks surrounding AMD’s first quad-core CPU, the Barcelona server/workstation processor. How fast will it be and - most importantly - will it be able to outpace Intel’s quad-core Xeon 5300-series (“Clovertown”) processors. It will, says AMD, and in fact upgraded its performance expectations.

So far, we the company had claimed that Barcelona will surpass the performance of Clovertown by about 40% at any given clock speed. Now the company says that it believes that Barcelona will have a 50% advantage over Clovertown in floating point applications and 20% in integer performance “over the competition’s highest-performing quad-core processor at the same frequency.”

[Shintai]

Please no more of those synthetic memory benchmarks, we want to see something REAL! Else I gonna get some Sandra SSE benchmarks....

Its the same old info we got for ages, nothing new.
And they are still too embarrased to add 2S dualcore K8 system that would trash the Barcelona.

[MAS]

the old info was 40 per cent advantage, now it's 50 per cent

[gallag]

“over the competition’s highest-performing quad-core processor at the same frequency.”

and this in a bench that k8 already kicks ass at, how does this look for gameing, media etc etc?

cant see a 2.5k10 besting a 3.33penyrn

[Movieman]

There are lots of question marks surrounding AMD’s first quad-core CPU, the Barcelona server/workstation processor. How fast will it be and - most importantly - will it be able to outpace Intel’s quad-core Xeon 5300-series (“Clovertown”) processors. It will, says AMD, and in fact upgraded its performance expectations.

So far, we the company had claimed that Barcelona will surpass the performance of Clovertown by about 40% at any given clock speed. Now the company says that it believes that Barcelona will have a 50% advantage over Clovertown in floating point applications and 20% in integer performance “over the competition’s highest-performing quad-core processor at the same frequency.”

That's the real key to this statement. What will be the highest speed that AMD will release this cpu at.
If they are going to make a 2400mgz as top end and are comparing to Intels clovertown at that same speed they may well be right but does it matter?
What I want to know is how their top quad compares to Intels top quad.
That's all that matters. The rest is smoke and mirrors.

[[XC] gomeler]

Enough already with the extremely vague benchmark results.. disgusts me that all we hear are fluffed up numbers from AMD. I really have my fingers crossed that K10 will bring a breath of fresh air into the market but I'm beginning to get the feeling that this thing is going to crash and burn Please be amazing..

[Shintai]

the old info was 40 per cent advantage, now it's 50 per cent

Old info was actually 46% (First 42%). And then some site made that into a nice round 50%.

And its no secret its still SpecFP_rate!

[Movieman]

Enough already with the extremely vague benchmark results.. disgusts me that all we hear are fluffed up numbers from AMD. I really have my fingers crossed that K10 will bring a breath of fresh air into the market but I'm beginning to get the feeling that this thing is going to crash and burn Please be amazing..

Agreed, and I want K10 to be the monster they say it is to keep prices down and competition between these 2 giants in place.
Trust me, the last thing you ever want to see happen is AMD go out of business.Intels prices would skyrocket and R&D would get a budget that wouldn't buy a Big Mac.

[red]

Wasn't old info 40% better than Intel's top bin? Now it's 50% at the same clock. On a bandwidth bench.

[MAS]

prices will go down - recollect Q6600 price in Q3 (266USD)

[Movieman]

If they can pull a 50% better math score than I've seen on a dual clover then good luck to them. They will have the market.

[MAS]

AMD has corrected numbers because of penryn tests

"if intel may speed up 45nm shrink up to 40 per cent - why may not I up to 50 for absolutely new architecture?" - AMD thought

[saaya]

50% better fpu and 20% better integer at the same clockspeed as a clover!

from what i heard so far K10 will come at 2.4Ghz at launch and intel is already 3ghz with their clovers. So that doesnt sound good at all...
it means intel will need a 3.6Ghz quadcore to beat K10 in fpu and only a 3ghz quadcore to beat integer performance. intel already has 3ghz quadcores today! and they should be able to reach 3.6Ghz with 45nm easily just a few months after K10 is launched. this sounds like bad news...

From those number K10 needs to run at least 2.4Ghz to regain the performance crown in some server segments.
For the desktop segment id say at least 2.6Ghz are necessary.
If K10 can even reach 3Ghz then intel will lose its performance crown completely in all segments.

fpu (+50%/clock)
2.4ghz K10 = 3.6Ghz clover/ 3.4Ghz york
2.6Ghz K10 = 3.9Ghz clover / 3.8Ghz york
2.8ghz K10 = 4.2Ghz clover / 4.0Ghz york
3.0ghz K10 = 4.5Ghz clover / 4.3Ghz york

int (+20%/clock)
2.4ghz K10 = 2.9Ghz clover/ 2.8Ghz york
2.6Ghz K10 = 3.1Ghz clover / 3.0Ghz york
2.8ghz K10 = 3.4Ghz clover / 3.3Ghz york
3.0ghz K10 = 3.6Ghz clover / 3.5Ghz york

[Movieman]

50% better fpu and 20% better integer at the same clockspeed as a clover!

from what i heard so far K10 will come at 2.4Ghz max and intel is already 3ghz with their clovers. So that doesnt sound good at all...

it means intel will need a 3.6Ghz quadcore to beat K10 in fpu and only a 3ghz quadcore to beat integer performance. intel already has 3ghz quadcores today! and they should be able to reach 3.6Ghz with 45nm easily just a few months after K10 is launched. this sounds like bad news...

EXACTLY the point I was trying to make. You said it better!
I'll be surprised if they can match what I'm seeing right now and thats at 3200 not 3600mhz.

[red]

Is Intel really threatened by SPECfp? Aren't AMD already ahead there? http://tweakers.net/reviews/661/2

movieman, is that a stock chip or no?

[K404]

AMD clockspeeds will ramp as well All these numbers are pointless, but AMD are releasing "easy" clocks on their chips so they dont need to worry about binning..once the initial rush is about dealt with they can think about binning and using the feedback they get from forums to see what the chips can REALLY do.

[Movieman]

Is Intel really threatened by SPECfp? Aren't AMD already ahead there? http://tweakers.net/reviews/661/2

movieman, is that a stock chip or no?

X5350 at 3271 for that screenshot.

Machine runs 24/7 at full load at 3200 on air.
Saaya can vouch for that or gomeler.

AMD clockspeeds will ramp as well All these numbers are pointless, but AMD are releasing "easy" clocks on their chips so they dont need to worry about binning..once the initial rush is about dealt with they can think about binning and using the feedback they get from forums to see what the chips can REALLY do.

I hope your right. Like I said, I want AMD to do VERY well with these chips.
MOST important to me and should be to all of us is the continued competition between these 2 companies.

[NightCrawler™]

I'm really getting fed up with these so called performances ratings.... Get the damn thing sampled and let some respected reviewers show the numbers..
Even beter get the damn thing in mass production so we get to play with it...

I don't know why AMD has to be this secretive about their upcoming line of products whether it's the R600 or the K10...
Intel had lots of C2D samples out in the open so we all knew what to expect...

Talk about bad marketing....
/waits for the next pricedrop from AMD to "compete" against Intel for the next 6-9 months...

[gOJDO]

The BS continues....No benchmarks, no system specifications, no numbers.....wrong asumptions("50% advantage over Clovertown in floating point applications" is not 50% faster)

The more bold claims AMD makes, the more I believe that K10 will be another failure, like 4x4. According to their K10 performance estimations based on hardcore marketing BS and their ridiculous statements:

The AMD Opteron™ processor continues to lead the industry in x86 dual-core performance.

http://www.amd.com/us-en/Processors/...l?redir=CPBM01
I don't believe that K10 will counter 65nm Core2.
From now on, I'll consider every word from AMD's mouth as FUD & BS.

Oh, and this part from the same link is so true:

Saying you’re the “World’s Best Processor” is one thing.
Actually being the “World’s Best Processor” is another.

[Brother Esau]

I clearly remember old quote was that it was 40% higher performer over Intels premier Quad Core.

[Shintai]

You have K10 sandra scores? Go ahead and get it, what are you waiting for?
AND BTW there is quite a difference between the spec benches and sandra, most of all the first one is and has been industry standard for years, and the other is a lame bench that has nothing to do with reality.

It is about time some real benches and results pop out...

Its the same PR mentality. As if I compared K8 SSE vs Conroe in Sandra.

Also lemme give you a hint. Goto www.spec.org and check specfp_rate2000 benchmarks between a clovertown and a 2S K8. Or a K8 vs a 51x0

Its bloated PR. And real world will be a real pain whenever this chip gets released.

[thunderstruck!]

Way to twist the words of Movieman. He hasn't said anything close to that. He is rooting for something good.

[erwinz]

I thought they will lunch the product today..

[Shaggy]

Seeing how those are both claiming their future cpu's will be a 40% increase intel will still control the market since AMD price hikes their cpu's a lot more like they did with the x2's.

[Shintai]

I thought they will lunch the product today..

More likely R600 only. Or a showcase of it. Not sure how much, if any K10 will be shown.