Superpi score proofs Phenom X4 is a bit faster than Athlon64 X2...

[nemrod]

Good points(and redbull's too).I can't remember in recent past,the time that AMD "pre-demoed" it's new uarch. and that it performed at the same level as it would when it hit the retail.We remember the early K7 "disastrous" tests with low FPU scores,than it was early K8's that could clock like *** and was later reviewed at Xbitlabs as 2800+ and barely beat K7 at the same clock ,then it was early Anand's AM2 preview that showed only decrease of perf. over s939(at the end ,on average with solid DDR2 RAM, AM2 chips were 3-5% faster than s939 with good DDR1,and rarely slower )

So yes,let's wait for the 19th and judge then.

In recent past, I've not seen a revolution with barcelona, is it?

[informal]

In recent past, I've not seen a revolution with barcelona, is it?

Have you seen one with K7->K8 transition??Was it like 50% faster?No it wasn't.It was typical 15-20%,just as Phenom will be over K8.Although,since the doubled SSE resources,Phenom will be a lot faster than K8 (than those average 20%) in SSE intensive apps.

[Zucker2k]

Why are all you AMD fanboys pretending superpi is the only benchmark Intel's CPUs are dominating AMDs newest phenoms? Show me a single popular benchmark where phenom or any other AMD cpu dominates an Intel CPU. Fact is all you have now is AMDs past glories to hang on to. Funny thing is, I could be there to but I choose to go Intel now because I SAVE TIME whenever I have to run a benchmark or encode files. As a matter of fact, I am encoding a 27GB HD DVD file to x264 format at 54fps, and still have an overhead to multitask. The best thing is, this is only a mild overclock of 3.5 Ghz on a Q6600. Goodluck trying to duplicate that on your AMD system.

[informal]

Why are all you AMD fanboys pretending superpi is the only benchmark Intel's CPUs are dominating AMDs newest phenoms? Show me a single popular benchmark where phenom or any other AMD cpu dominates an Intel CPU. Fact is all you have now is AMDs past glories to hang on to. Funny thing is, I could be there to but I choose to go Intel now because I SAVE TIME whenever I have to run a benchmark or encode files. As a matter of fact, I am encoding a 27GB HD DVD file to x264 format at 54fps, and still have an overhead to multitask. The best thing is, this is only a mild overclock of 3.5 Ghz on a Q6600. Goodluck trying to duplicate that on your AMD system.

Funny thing is that there are no benchmarks of Phenom available(apart from one Crysis which is fishy as hell since it's not threaded at all and the system had some low memory subsystem values;yeah one SPi bench was posted too-apart from the fact that SPi has no merit at all)

[Tulatin]

Problem is, so much as AMD may be a "platform company" What would be the point of buying a phenom + 790 mobo if you could go intel quad + P35 for the same money (the only fallancy here is the CPU), and then OC your way well past graceland? The GPU part of that equation is NOTHING special. You can run CF on anything, you know...

As much as I like Damit, they have to pull their heads out of their asses, and stop this "We're a platform/midrange company. We don't need performance parts" . If K10 is just ANOTHER K8 iteration just like AM2, then FFS, why don't we call it NetBurst?

Edit: and though the K7 -> K8 jump may have been 20% at the gate, K8 had a BIG difference: Headroom. What's the point of putting out a new line of chips if they're the fastest of that breed you'll ever see.

[John600rr]

Funny thing is that there are no benchmarks of Phenom available(apart from one Crysis which is fishy as hell since it's not threaded at all and the system had some low memory subsystem values;yeah one SPi bench was posted too-apart from the fact that SPi has no merit at all)

What do you mean by the Crysis benchmark is not threaded at all?

The Crysis CPU test loaded my E6600 80% on one core and around 75% on the other.

[cky2k6]

right, but it doesn't load a quad core, cause the dual and quad core intels in the test scored within 2fps of each other. memory scores for these phenoms also seem low, they get less bandwidth than core 2, which is unexpected with an imc...

[informal]

right, but it doesn't load a quad core, cause the dual and quad core intels in the test scored within 2fps of each other. memory scores for these phenoms also seem low, they get less bandwidth than core 2, which is unexpected with an imc...

Exactly.BTW,i think that Phenom got even lower mem. scores than K8 in those tests.

[John600rr]

But it is not as if the Q6600 was running some 'different' version of Crysis which was better multi-threaded.

As many programs today are not completely multi threaded, especially for quad cores, I think it makes the results even MORE realistic.

[savantu]

Have you seen one with K7->K8 transition??Was it like 50% faster?No it wasn't.It was typical 15-20%,just as Phenom will be over K8.Although,since the doubled SSE resources,Phenom will be a lot faster than K8 (than those average 20%) in SSE intensive apps.

Is the difference between K10 and K8 greater than K8 vs. k7 ?
I doubt it.

As a result , K10 vs. K8 will impress less than K8 vs. K7 and your whole argument is in the air.

[uwackme]

First, dont even bother me with a test that runs the ram slower than 500Mhz. Using 375 on an AMD, wtf is that ?

My 5600+ gets 24s/22m SuperPi results. You gonna tell me an X4 is that slow... the test is flawed. Learn how to apply the scientific method to your work, then come back and let us know your tests and results. otherwise you are wasting electrons.

[informal]

Is the difference between K10 and K8 greater than K8 vs. k7 ?
I doubt it.

As a result , K10 vs. K8 will impress less than K8 vs. K7 and your whole argument is in the air.

Hmm,can you read right?

Let me quote myself:

Have you seen one with K7->K8 transition??Was it like 50% faster?No it wasn't.It was typical 15-20%,just as Phenom will be over K8.Although,since the doubled SSE resources,Phenom will be a lot faster than K8 (than those average 20%) in SSE intensive apps.

And yes,Phenom over K8 will be the same as K8 over K7.15-20% on average and a lot more in certain (SSE) scenarios.

[Epsilon84]

Funny thing is that there are no benchmarks of Phenom available(apart from one Crysis which is fishy as hell since it's not threaded at all and the system had some low memory subsystem values;yeah one SPi bench was posted too-apart from the fact that SPi has no merit at all)

You must have missed this?
http://www.expreview.com/news/hard/2...8206d6758.html




I think they mislabeled the X4 as a '7000' model when it should in fact be a '9000' model.

Comparing most of those numbers to Q6600 ones (such as PCMark05 or 3DMark 05/06) shows the 2GHz X4 to be slower somewhat, then again it is only clocked at 2GHz, and we don't know the memory timings etc.

[GoThr3k]

You must have missed this?
I think they mislabeled the X4 as a '7000' model when it should in fact be a '9000' model.

Comparing most of those numbers to Q6600 ones (such as PCMark05 or 3DMark 05/06) shows the 2GHz X4 to be slower somewhat, then again it is only clocked at 2GHz, and we don't know the memory timings etc.

you think that,but you dont know it for sure

and if you look at the memory bandwidth,you know something isnt right here.....

[informal]

You must have missed this?
http://www.expreview.com/news/hard/2...8206d6758.html




I think they mislabeled the X4 as a '7000' model when it should in fact be a '9000' model.

Comparing most of those numbers to Q6600 ones (such as PCMark05 or 3DMark 05/06) shows the 2GHz X4 to be slower somewhat, then again it is only clocked at 2GHz, and we don't know the memory timings etc.

No i haven't miss that one.Look at the memory subsytem scores for Phenom,aren't they strange?I call BS untill i see official review.
I think we should reserve our judgment 'till the parts officially launch.And i think that since some online stores list the ETA as early as tomorrow,some one will probably get Phenom and test it here at XS long before we hit Nov 19th.

[Epsilon84]

you think that,but you dont know it for sure

and if you look at the memory bandwidth,you know something isnt right here.....

Well, if someone is stupid enough to list a tri core as a quad core then I can't do much about that can I? Therefore I can only ASSUME they made a simple MISTAKE in labelling.

Somehow, I imagine if those results were actually favorable to AMD there wouldn't be all this skepticism?

[Epsilon84]

No i haven't miss that one.Look at the memory subsytem scores for Phenom,aren't they strange?I call BS untill i see official review.
I think we should reserve our judgment 'till the parts officially launch.And i think that since some online stores list the ETA as early as tomorrow,some one will probably get Phenom and test it here at XS long before we hit Nov 19th.

That I agree with. Can't wait, it's been too long already.

[informal]

Yep,i just hope some one grab it before the 19th.Since so many people here at XS shell out the good amount of $ for 790FX mobos,I therefore expect some of them managed to at least grab 9500.For it's price it should be great performer.Also i am looking forward to seeing how ordinary AM2 mobos cope with IMC clocking thing(shouldn't be a problem,but i am careful optimistic )

[Entity_Razer]

quick note:

My school has ... lets say a big "playground" with lots of fun hardware and my teacher today told me we have 3 barcelona's in school up and running with the needed motherboards/ram to support it (under NDA ofc) so all i can say/him confirm is that those benches are WAYYYYY of, and i really do mean WAY OFF

[accord99]

quick note:

My school has ... lets say a big "playground" with lots of fun hardware and my teacher today told me we have 3 barcelona's in school up and running with the needed motherboards/ram to support it (under NDA ofc) so all i can say/him confirm is that those benches are WAYYYYY of, and i really do mean WAY OFF

Barcelona is released so why would there be any NDA?

[Entity_Razer]

no idea all he said is that he can't openly discuss all the details/benches yet.

it isn't that surprising that they can't TBH, they are working in tight coop with both Intel as AMD and we constantly get the latest hardware to test/play around with (wel he and his specialised department does) so... prob some reason he can't give any long details yet

[Entity_Razer]

btw:

http://anandtech.com/cpuchipsets/sho...spx?i=3092&p=1

Not THAT extensive but i'm surpised i didn't see that link show up more. Barce tests & also a Opteron (barcelona Opterons) test linked in there. Shows barcelona to be 15-20% faster clock/clock then the x2 and the bandwidth to be a LOT better then any other benches i've seen linked here

[informal]

Entity_Razer,something is definitely going on behind the scenes until the 19th(can't dig out from local AMD guy what's really going on).I guess it's because of the Phenom launch and the fanfare.It's interesting where is AMD pricing the 2.2Ghz model tho.
Not long before 19th is over .

[Entity_Razer]

hehe...

well anand or johan de gelas said it in one of the reviews on Anand.... if Barcelona clocks 3Ghz it will pretty much dominate the market for months to come

[nemrod]

you think that,but you dont know it for sure

and if you look at the memory bandwidth,you know something isnt right here.....

It's not a tricore. If you go back to the source:
http://www.force.game.tw/
http://myevilprocessor.blogspot.com/...4-x2-6400.html
You will see that there is a name correction.

(Tests could be wrong, but this is about quadcore)