AMD Zambezi news, info, fans !

[LowRun]

Incredible how many times people have been told not to post OBR stuff but still they go on

[bamtan2]

that is very interesting. the cores bebop and skedattle like they're on spring break, until the benchmark is set to 8 cores, when the cores clock speed goes down so much that each one is pegged at 100% doing work. it seems like that benchmark is a sketchy one for this testing purpose. I think you would prefer something that pegged the chip at 100% even when at 4.2ghz. but still, a very interesting video, even if it seems to be benchmarking the benchmarking app... or the operating system or the motherboard or the chip or whatever is doing the load switching

[Evantaur]

the guy who posted that vid has as much credibility as a klingon
never trust a klingon!

[The Stilt]

that is very interesting. the cores bebop and skedattle like they're on spring break, until the benchmark is set to 8 cores, when the cores clock speed goes down so much that each one is pegged at 100% doing work. it seems like that benchmark is a sketchy one for this testing purpose. I think you would prefer something that pegged the chip at 100% even when at 4.2ghz. but still, a very interesting video, even if it seems to be benchmarking the benchmarking app... or the operating system or the motherboard or the chip or whatever is doing the load switching

I suggest to take a look at this: http://blogs.amd.com/work/2011/01/31...er-goes-to-11/

Power consumption increases as the amount of threads used increase.
Once the TDP has been reached, the turbo feature will stand down (-> return to the next highest performing Pstate which does not breach the TDP) until "headroom" is again available.

I recon a properly multithreded game such as Dirt 3 would be perfect to demonstrate the turbo feature on BD.

[Olivon]

New AMD Opteron 6200 models sighted - CPUWorld



Thanks to Ronny145, once again !

[JumpingJack]

Do you have any documentation on this? Not denying just would like to know exactly how Win7 & Vista are able to discern between logical and physical cores.

Windows XP is SMT aware, as is Windows 2003:

This white paper provides information about support for the Hyper-Threading Technology (HT) in the Microsoft® Windows® family of operating systems. It provides an overview of HT, details of dependencies on BIOS, a description of the Windows operating system license model for HT, details of the support features in Windows XP and the Windows Server 2003 family, and guidelines for application developers on how to take advantage of the features and the performance benefits provided. This paper is intended for BIOS developers, OEM system manufacturers, and Independent Software Vendors (ISVs) that produce multithreaded Windows applications, particularly those that use processor affinity

http://download.microsoft.com/downlo...ad_Windows.doc

This does not mean they were any good at it, Vista is also SMT aware but it also sucked. In windows 7, MS introduced SMT parking (a variation on core parking used in server class OSes): http://www.ditii.com/2009/07/23/wind...on-with-intel/

What this means is, that in lightly threaded applications, windows 7 scheduler attempts to schedule threads per physical core rather than scheduling threads that would share a core. The effect is actually pretty pronounced, as much as 20% performance can be reclaimed in some situations, at least in those that I have been able to measure. How effective this is in the larger scheme of things is uncertain to me, but it is a measurable effect.

[ReKcOlNu]

In windows 7, MS introduced SMT parking (a variation on core parking used in server class OSes): http://www.ditii.com/2009/07/23/wind...on-with-intel/

I get the feeling SMT parking was just a needed bug fix for Intel CPUs with hyper-threading running Windows all along, but MS/Intel never got around to fixing it until Win7..... so of course the marketing dept will just call it a feature! LOL!

[duron]

Did You Know?

http://www.amd.com/us/aboutamd/corpo...fun-facts.aspx

[BeepBeep2]

I get the feeling SMT parking was just a needed bug fix for Intel CPUs with hyper-threading running Windows all along, but MS/Intel never got around to fixing it until Win7..... so of course the marketing dept will just call it a feature! LOL!

Well, during Pentium 4, choosing the SMT thread over a core was never a problem...

But when you have 12 threads and 6 real cores, I can see it becoming a problem.

[-Boris-]

Not much news, rumors or speculations in this thread anymore so close to launch.

[muziqaz]

Did You Know?

http://www.amd.com/us/aboutamd/corpo...fun-facts.aspx

''Did you know that AMD's new 16-core processor codenamed "Interlagos" will offer the world's highest x86 core count?''

*sigh* Would have been nicer to see: highest performance, not core count as the bullet point

[-Boris-]

''Did you know that AMD's new 16-core processor codenamed "Interlagos" will offer the world's highest x86 core count?''

*sigh* Would have been nicer to see: highest performance, not core count as the bullet point

I agree, would be better if they didn't say anything at all. Now it just sounds like they don't have good performance despite the core count.

[AzureSky]

Well, during Pentium 4, choosing the SMT thread over a core was never a problem...

But when you have 12 threads and 6 real cores, I can see it becoming a problem.

this is true but only up to a point, for example, try running windows 2000 on a p4 with HT enabled like many companies did back then, you have to disable HT to get the performance to not suck, because windows 2000 tries to use both cores as real cores, and thus overloads the poor little intel netburst chip.

I know, most people back then wouldnt use 2000, but the fact is, I still to this day see companies holding onto 2k over xp or vista, and some are even holding out for windows 8(stupid cheap bastages, 7 would be a better move if you ask me....)

I have also seen 2k used with modern HT enabled proc's and its never pretty, mind this was for testing, but if you try and treat the 6 core cpu as a 12core, things dont work out to well in high load situations.

I have myself found as high as a 37% drop in perf when encoding using the x64 ogg vorbis encoder on all "Cores" even under 7, disable HT and your perf comes back....

I really hope amd's new design avoids such issues

[johnieboy]

http://translate.google.com/translat...ldozer-fx.html

For those looking for some advice before the time about Bulldozer FX, here's one of our colleagues CanardPC over two weeks yet before the end of the NDA. How is this possible? Now comes the newsstands the new magazine CanardPC Hardware , with inside the test Bulldozer FX ... or almost!

In fact the test is not yet available, NDA forces, and readers can actually login from October 12 on a web page to discover the full review. The scores are not directly present in the magazine, but the buying guide of CPU if and CanardPC it evokes the FX-8150 and 8120, both versions expected octocores respectively 245 and $ 205.

Here are some selected pieces from the magazine of our colleagues to read it all for yourself and have access to complete test CanardPC from October 12, it will cost you € 5.90 at your newsagent.

"In the more upscale, FX processors based on the Bulldozer architecture we also disappointed: while they are still generally more efficient than their predecessors and allow AMD to approach much of the last Core i5 and i7 but their performance remains below expectations. Besides, as we announced already in our previous issue, if they can sometimes compete with Sandy Bridge in the applications of rough calculations, the results are in video games very far behind. Only overclockers (and fanboys) will find them a great interest given their predisposition in this area. "

"The AMD FX-8120 is probably the model of the new series" Bulldozer "that offers the best price / performance ratio. It is able to compete with the i5-2500K in most computing applications even if gross lags behind in video games. are nonetheless substantial overclocking capabilities and available at no additional cost. Faced with the old Phenom X4 980, this is a very good alternative. "

"Offered at a price slightly lower than the Core i7-2600K Intel, the FX-8150 is currently the most powerful model from the new architecture" Bulldozer "for AMD. Unfortunately, he fails, at best, that to match its direct competitor in some media processing applications and is always behind in games. "

[tifosi]

http://translate.google.com/translat...ldozer-fx.html

Last time Dresdenboy checked and reported that these chappies don't have the latest AGESA code. God knows what else is amiss. In short, i'm not holding my breath to read their review and may be so shouldn't you. Not yet!

Any which ways you cut it, 8150 competing with Gulftowns and keeping it right up with SB is no small feat. Yes, it may come out as people are saying it is, i.e. close but not faster than SB. Mind you, i have no chips, nor a clue, i'm just assuming people talking out of their rear ends may be correct. However, don't forget that the leap from Phenom II to SB is quite huge.
__________________________________________________ __________________________________________________ _________________

Note for some "will buy Intel" posters. This is a thread in AMD forum. Comparisons are inevitable, but trolling isn't! I think that sums up my rant!

[-Boris-]

Any word on native quads? If quads isn't to far behind octas at few threads I might buy one of those and overclock them. I think native quads could be quite cool, and be capable of high frequencies.

[Manicdan]

do you mean the 4100, or do you mean where the literally rebuild it without 2 modules, or where they turn off the extra "core" for each module?

[muziqaz]

Last time Dresdenboy checked and reported that these chappies don't have the latest AGESA code. God knows what else is amiss. In short, i'm not holding my breath to read their review and may be so shouldn't you. Not yet!

Any which ways you cut it, 8150 competing with Gulftowns and keeping it right up with SB is no small feat. Yes, it may come out as people are saying it is, i.e. close but not faster than SB. Mind you, i have no chips, nor a clue, i'm just assuming people talking out of their rear ends may be correct. However, don't forget that the leap from Phenom II to SB is quite huge.
__________________________________________________ __________________________________________________ _________________

Note for some "will buy Intel" posters. This is a thread in AMD forum. Comparisons are inevitable, but trolling isn't! I think that sums up my rant!

Not to mention the price difference between fx-8150 vs 2600K (rumoured $245 vs $315 at newegg). So it's by far not the same price range.

[rog]

Last time Dresdenboy checked and reported that these chappies don't have the latest AGESA code. God knows what else is amiss. In short, i'm not holding my breath to read their review and may be so shouldn't you. Not yet!

There is no performance difference between 0.9x and 1.x Agesa code at all.

[EniGmA1987]

There is no performance difference between 0.9x and 1.x Agesa code at all.

How do you know?

[Smartidiot89]

do you mean the 4100, or do you mean where the literally rebuild it without 2 modules, or where they turn off the extra "core" for each module?

There's supposed to be a native Quad-Core (2 modules) with only 4 MB L3 cache.

[bamtan2]

http://translate.google.com/translat...ldozer-fx.html

good info, thank you. I think that sounds legit. it competes well but falls behind in games, where intel is usually the leader anyway.

now the final question becomes: can the normal overclock of a bulldozer make it faster than a normal overclock of a sandy bridge. probably 99% of 2500k can run at 4.2ghz. if 99% of bulldozer can run 5ghz or more on air, there might be a chance for bulldozer yet...

[demonkevy666]

There's supposed to be a native Quad-Core (2 modules) with only 4 MB L3 cache.

it that the only native lesser die ?

[radaja]

There's supposed to be a native Quad-Core (2 modules) with only 4 MB L3 cache.

cool that is good news they should be lower TDP right,i thought there was only the x4 with two modules disabled like the FX4100 with 4Mb L2 and 8Mb L3?

[Manicdan]

the problem with a 2 module only chips is that you cant unlock it, for some of us that would have been the main selling point.
for AMD though it would be a great price/mm2 rather than turning perfectly fine chips into half working chips simply due to demand. and to be honest up to 4 cores (or threads, however people want to perceive it) is still all thats really desired from most people, and anything more just isnt worth the extra price.

i do hope that because they are built starting with just 2 modules, it can have some with a good yield that OCs like crazy and thus offering 4.5ghz or more in stock forms.