More Bulldozer Benchmarks (Supposedly from AMD)

[Idontcare]

All these slides are fake. They're damn awful with terrible typesetting and it's impossible that they've been made by any professional in the matter. Even lines are broken in some of these images.

Stop looking at the scores and think a bit about the container ffs.

Wait, you mean to tell me that a world-renowed reputable site the likes of donanimhaber.com would stoop so low as to peddle questionably authentic photoshopped slides? oh the horrors

 

[RussianSensation]

Two other oddities mentioned .....

How is AMD's 2600k 25% faster than 2500k in WinRAR 4.0 and yet real world tests show about an 8-10% difference?

Computerbase.de also found an 8% difference between a 2600k and 2500k in WinRAR 4.0.

Also, some of those benchmarks already fly on the X6 1100T, such as 7-Zip:

So looking at their chart, FX-8150 is still barely faster than the 2600k in the same 7-Zip bench? Shouldn't it easily beat the 2600K in 7-Zip?

 

[Chinoman]

I'm definitely down to get BD if those slides are accurate.

AMD's very recent fusion roadmap leaks format very similar to these slides (the AMD logo on the bottom right corner), so I'd say there's a good chance of these figures being legitimately from AMD.

But that's all I can conclude. Whether or not they're ACCURATE representations made by AMD is a whole different story.

 

[Silverforce11]

If it turns out to be ~2600K in encoding and gaming, i think my next build will finally be an AMD system (haven't had AMD cpus for over a decade!).

 

[NostaSeronx]

Wait, you mean to tell me that a world-renowed reputable site the likes of donanimhaber.com would stoop so low as to peddle questionably authentic photoshopped slides? oh the horrors

"OF Course, all results was FAKEs, i know it, ive faked them No real numbers before NDA "

Flashback moment

 

[Imouto]

Wait, you mean to tell me that a world-renowed reputable site the likes of donanimhaber.com would stoop so low as to peddle questionably authentic photoshopped slides? oh the horrors

The irony meter just imploded. You sure are enyoing yourself.

Anyway lol at anyone who took these as legit. Broken lines, awful layout, pixeled logos, nosense charts, unrelated benches, borked and hilarious sentences and more. And some ppl look at this and say that AMD marketing guys are earning their money. Just frigging lol.

 

[Patrick Wolf]

Why will it do anything at all to intel? They'll just introduce the 2700/2800/2900k to replace the 2600k's pricepoint and drop 2600k down to take 2500k's spot. They've been planning to do so for months, they're just waiting for BD to finally arrive.

Good point I think. Intel has had quite some time to cook up new chips to compete with BD. Reminds of this:

 

[NostaSeronx]

 

[Lonbjerg]

So you are saying Bulldozer will run like a dog? :sneaky:

 

[NostaSeronx]

So you are saying Bulldozer will run like a dog? :sneaky:

No, when compared to a dog and cat

That is what AMD FX looks like and if you click the picture that is its theme song

 

[Madmanden81]

If benchmarks are from AMD, they will be on AMD.com.

Not if the slides are leaked, I imagine. Not that I think this is legit anyway.

 

[psolord]

So you are saying Bulldozer will run like a dog? :sneaky:

...on steroids!

 

[lifeblood]

Lets make two assumptions. 1) these slides are not fake (questionable), and 2) AMD are not idiots. Given that, what do they mean?

AMD is positioning its top BD chip against the 2500k. This is good as the 2500K is a popular chip so their is good money to be made here. According to the charts it will perform a little better but cost about the same. Performance can/will vary with the BD doing better in heavily threaded apps and the 2500k doing better in less threaded apps, but when you look at the big picture, they're roughly equivalent.

I would think this is a good thing. Given the other rumors of BD being a total failure, at least now it is competitive. I was looking at upgrading to a 2500k so now I actually have a choice.

Of course, this assumes the prices are equivalent. I don't care about IPC, MHz, solid gold connectors, pretty boxes, or any of that crap. All I care about is when I slap my money down how much performance can I buy in the apps I care about.

 

[CHADBOGA]

I don't care about IPC, MHz, solid gold connectors, pretty boxes, or any of that crap. All I care about is when I slap my money down how much performance can I buy in the apps I care about.

There is a clear connection between the two bits I have bolded, even if you can't see it.

 

[exar333]

Hey! 8 cores can almost 4 with SMT! Awesome results!!

 

[tincart]

Hey! 8 cores can almost 4 with SMT! Awesome results!!

Is this a game where we try to figure out which verb to use?

 

[bronxzv]

Hey! 8 cores can almost 4 with SMT! Awesome results!!

it was predictable since the disclosure of the Bulldozer memory hierarchy :
http://www.realworldtech.com/page.cfm?ArticleID=RWT082610181333&p=8

one common bottleneck for high throughput multithreaded code is the L2$ bandwidth, on paper a Bulldozer module has more or less the same L2$ bandwidth than one SNB core (+ more L2$ usage due to the two small L1D write-through caches vs. the write-back shared L1D on SNB) so it's predictable that 4 BD modules will have roughly the same speed than 4 SNB cores on benchmarks that saturate the L2$ bandwidth (x264 come to mind).

Now, IMHO both designs have a lackluster L2$ throughput, simply not enough bandwidth to handle well optimized AVX-256 code (+ FMA4 on Bulldozer)

Anyway, in other words one BD "core" is more or less half a SNB core. If AMD really intents to advertise the fact that they have "more cores" I hope reviewers will resurect the (pretty good) AMD's "True Performance Initiative" paper.

 

[toolbag]

I'm going to buy a bulldozer now and kill everyone in my FPS games.

 

[nyker96]

1st they showed mostly games and high threaded app, missing office work app. If these slides are real and data represents reality, then we can at least say BD caught up w/ intel on games, of course threaded it better be since w/ so many cores. but single, low thread, office and media apps probably still behind Intel as they didn't present anything in those areas.

 

[nyker96]

I'm going to buy a bulldozer now and kill everyone in my FPS games.

that depends on ur gaming skill more than your computer ;]

 

[trollolo]

Hey! 8 cores can almost 4 with SMT!

sorry i accidentally the whole test results.

 

[aigomorla]

I'm beginning to have second thoughts on the FX chips from "joke" to "possibility". I just talked with a NewTek rep who strongly suggested I prepare to replace my 12U of Lynfield Xeons with AM3+.

Daimon

he is going to get fired real soon... :biggrin:

No IT department would allow this, unless your guy favored AMD.
Which is financial suicide (waste in IT resources) unless it was Magoney that was replacing the intel rigs.

Lynfield = Not that old in technology that would merit needing an upgrade.
Lynnfield > AM3+
Lynnfield -> AM3+ = your IT guy is spending unneeded resources in a time where resources is very scarce.

Hence, when someone in your IT department finds out and tells the Boss, your IT guy is going to be the first one to get a pink slip from the company.

 

[JBT]

I want gaming benchmarks vs. i5 2500k and i7 2600k at stock and then both OC'd to max with air.

gogogogogo

Same here.

 

[dac7nco]

he is going to get fired real soon... :biggrin:

No IT department would allow this, unless your guy favored AMD.
Which is financial suicide (waste in IT resources) unless it was Magoney that was replacing the intel rigs.

Lynfield = Not that old in technology that would merit needing an upgrade.
Lynnfield > AM3+
Lynnfield -> AM3+ = your IT guy is spending unneeded resources in a time where resources is very scarce.

Hence, when someone in your IT department finds out and tells the Boss, your IT guy is going to be the first one to get a pink slip from the company.

Lightwave is easy to build clusters for. I have six 2U quad-core Xeons on my desk, which cost about $700 per node with ECC memory. Lightwave also scales linearly with core count. Cheap AMD boards accept non-regestered ECC. Xeons on a server board do not overclock. 48 decent cores > 24 Gen.1 Nehalem cores. I don't pretend expertise in watercooling, but I do with clusters. I can see Zambezi being the basis of a cheap farm. I'd likely gain performance AND money right now by selling off my Xeons and Supermicro boards, and that includes the price of six Intel NICs.

Daimon

Edit: I'm pretty sure a guy doesn't get fired for having new guidelines for existing customers looking to build cheap nodes, and I don't subscribe to the idea that I should be locked into an Intel platform, as I have been since 2006.

I am my IT department; rendering is a hobby.

 

[BudaBomb]

Yeah but even if these test results are true and it doesnt quite keep up with intel in every aspect you can still get a processor that is comparable for less and spend more money elseware on your pc. Plus arent all bulldozer chips supposed to be unlocked so that you can overclock then and you dont need to get the most expensive chip to do that right.