AMD Speed problems with the new Bulldozer/Zambezi cpus

[haxartus]

Where is the comparison that you're talking about ? Most of the released screenshots are fake.

[Deleted]

Give me your source so I can debunk it into oblivion.

http://wccftech.com/2011/05/21/amd-b...nch-benchmark/

Not the source I read it from, but the only one I had at hand atm.

[vesseblah]

It's one of those supposedly leaked screenshots/comparisons where 8 cores was faster than 4+HT in multithtread test. Honestly, it would be pathetic if that wasn't the case...

edit:

Cinebench referred above is currently the best real world benchmark for multiple cores, so it's possilble and totally believable that higher number of cores would and should win it.

[haxartus]

That result is not possible for an 8 core Bulldozer. If we had an 8 core Phenom II it would score better than that.

[Skarsguard]

I would rather them fix all the bugs and release it later then release it now and not be up to par to the sandy bridge but if they wait to long Intel's new bridge will be coming out and it could just be the end for amd. If your thinking about waiting to get the new bulldozer like i was going to do i would just get a Sandy Bridge now and enjoy it.

[Nightbliss]

I would rather them fix all the bugs and release it later then release it now and not be up to par to the sandy bridge but if they wait to long Intel's new bridge will be coming out and it could just be the end for amd. If your thinking about waiting to get the new bulldozer like i was going to do i would just get a Sandy Bridge now and enjoy it.

How is it the end for AMD?
I seriously have a hard time getting these statements.

AMD has once rulled the CPU market and only a couple of years ago, Intel claimed that title.
AMD CPU's will always have a market share because not everyone is ready to pay 200€ for a CPU when they can pay 100€ and get something that does the same, just not that fast.

Or do you see everyone riding Ferrari's and other exotic/sports cars?
Different people have different needs and if it wasn't for these sort of ideology, Intel CPU's would cost 5x more and same goes for Nvidia GPU's, if AMD wasn't there to release cheaper yet powerful GPU's, Nvidia prices would pump up like an erect dick..

[Deleted]

I would rather them fix all the bugs and release it later then release it now and not be up to par to the sandy bridge but if they wait to long Intel's new bridge will be coming out and it could just be the end for amd. If your thinking about waiting to get the new bulldozer like i was going to do i would just get a Sandy Bridge now and enjoy it.

For AMD to release something that cannot beat Intel would sure be the death of them for the high end CPU market. So they won't. I believe AMD knows their CPU will beat the competition and wants to make sure it does it by miles.

[Asera]

Really starting to see AMD ditch high end and focus on servers, SOC's, and low power chips now. Especially with the success of Zacate so far (I love mine!).

[Nightbliss]

Starting to believe the news on Xbit Labs were total bullcrap.

Source & More

Schedules it for "late summer"
During its big conference at Computex, AMD officially broke the news that the Bulldozer is delayed to "late summer".

Although the new 9 Series chipset is now officially announced, it looks like that FX Series, codename Zambezi CPUs based on the Bulldozer architecture will have to wait a bit. AMD didn't provide the official date but just rather said that the FX Series is coming in "late summer", which means that it might arrive in late July or even in August.

In case you somehow managed to miss it, the FX Series is made in 32nm and will be available in 4-, 6- and 8-core flavors. It is made for AMD's latest AMD 9-series AM3+ chipsets, but will also work on some AM3 motherboards as well, without AMD's blessing of course.

More

[Deleted]

Starting to believe the news on Xbit Labs were total bullcrap.

A Chinese user on Zol.com.cn forums has posted a CPU-z Screenshot of him running an Bulldozer based Zambezi-FX Processors on a Core clock of 6Ghz. The Processors used was “FX-8130P” which will be AMD’s top offering in its Bulldozer Lineup and the overclock was achieved using Liquid Nitrogen.

No shit at nearly 2v.

The 32Nm based AMD FX-8130P comes at a stock clock of 3.8Ghz which can turbo upto 4.2Ghz holding a total of 8Cores/8 Threads with 8Mb L3 Cache, The processor used was an early engineering sample and part of the older revision, AMD announced new Revisions due to performance issues in its current Bulldozer lineup so final clocks of the model would be changed. The clock achieved by the user was 6021.2mhz(223.0×27) on an insane 1.984 VCore, The CPU only consumes 125w TDP while the one in the CPU-z shows 186W.

[Asera]

No shit at nearly 2v.

Not to mention they were getting K10.5's to >7ghz on Liquid Nitrogen. This only got to 6?

[Nightbliss]

No shit at nearly 2v.

Yes and? I'd like to see where are these problems that prevent BD to stick above 3.5Ghz people talk about.
Hence my post, its not like I'm 10 and new to Informatics and Harware.

And by the way, some more nice info from a PDF released by AMD for their partners (not supposed to have come out public).

Here

Need to have a PDF Reader installed.

Not to mention they were getting K10.5's to >7ghz on Liquid Nitrogen. This only got to 6?

Number of Cores plays a huge part in Voltage requirement, overclock stability and achievement.
Ofcourse you know this already.

[Deleted]

Yes and? I'd like to see where are these problems that prevent BD to stick above 3.5Ghz people talk about.
Hence my post, its not like I'm 10 and new to Informatics and Harware.

Rumors are rumors you know. I'll believe they're delayed when AMD comes out and tells us in the face. The CPU-Z you posted there however looks dangerously legit.

[Erous]

AMD's original Phenom line was hit with bugs that impacted performance by 20%+. I thought they really came through and learned their lessons with the new revisions. It stinks that their new and upcoming line is lacking performance compared to their rival, again

[Deleted]

AMD's original Phenom line was hit with bugs that impacted performance by 20%+. I thought they really came through and learned their lessons with the new revisions. It stinks that their new and upcoming line is lacking performance compared to their rival, again

How can you possibly know it's lacking performance? There's no official release of the chip yet, even if it explodes during testing there's no way you can judge the chip yet.

[Nightbliss]

How can you possibly know it's lacking performance? There's no official release of the chip yet, even if it explodes during testing there's no way you can judge the chip yet.

So now its ok to say this but when I said it, people said I missed the point blablabla?
They are engineering samples so it doesn't matter if they can get past 3.5Ghz or not.

[Deleted]

So now its ok to say this but when I said it, people said I missed the point blablabla?
They are engineering samples so it doesn't matter if they can get past 3.5Ghz or not.

I never really stated it was a problem they couldn't, and honestly I don't care either. If they run a lower clock speed and beat the competition, who cares? I'm not an e-peen freak that wants over 9000MHz to be cool.

[Asera]

Number of Cores plays a huge part in Voltage requirement, overclock stability and achievement.
Ofcourse you know this already.

Yeah. But technically the 8core bulldozers are still quads, since it's just part of the execution line thats split in two. It's still 4 "modules", as they call them. I'd expect it behaves similar to a quad in terms of stability in that respect.

[Nightbliss]

I never really stated it was a problem they couldn't, and honestly I don't care either. If they run a lower clock speed and beat the competition, who cares? I'm not an e-peen freak that wants over 9000MHz to be cool.

Same really.
But I thought it was funny that people "bashed" me for not understanding the top logic but then use the very same arguments to kill other people's theories

Anyway, theres too many sources leaking info that you don't know which is real and which isn't.

[dicertification]

All we have atm is speculation and very few facts.

The only thing that I'm believing at this point is that there is a delay because there are a number of sources claiming such. As far as performance goes, we will have to wait for production silicon to be sure. I'm keeping my hopes up.

But if there's to much of a delay people will get on the "wait for Ivy Bridge" bus, and that would be harsh for AMD's sales.

---------- Post added 2011-06-02 at 01:41 PM ----------

Yeah. But technically the 8core bulldozers are still quads, since it's just part of the execution line thats split in two. It's still 4 "modules", as they call them. I'd expect it behaves similar to a quad in terms of stability in that respect.

?

I thought each "core" had 2 int units on it. A quad core would be a 8 int unit. But a 8 core would be a 16 in unit so the 8 core would still be 8 cores as each has it's own FPU along with the 2 int units on it.