AMD Zambezi news, info, fans !

[RussC]

My test of success, is whether its good enought to show up in the 1page ad's that are in PCWorld by IBUYPOWER and CyberPower. If AMD FX is in those ad's, then its a success. It would appear to me that the leaked performance so far would be able to get AMD FX into those along side INTC I7 and I5. Especially at the proposed price of ~$240 8150. JMHO

RussC

[xVeinx]

There is a lot that doesn't add up, and the fact that certain individuals are excited by bulldozer is enough for me to remain confident that the rumors being pumped out right now are not the full story. On the other hand, it is rather confusing to see the varied performance in these rumors. None of it looks particularly wonderful as of yet..

[informal]

I was saying the same thing when this point was brought up initially. It doesn't make sense to release a flawed product to bad press.

I still say something doesn't add up, unless the FX-6100 ( I assume 4 cores 6 threads) I'm eyeing is only slightly better than a 1090t; but anyone expecting a SB killer is going to be "disappointed."

FX6100 with the leaked specs (3.3Ghz stock/3.9GhzT),and if these 8150 performance number are true, has no chance against Thuban 1100T. In MT apps it will lose hands down to 1100T,now that 8150 barely manages to beat it, while in ST it will score lower than 1100t @ 3.7Ghz (which is its max Turbo). Potential buyers are left with higher OC headroom though. If FX6100 can go to ~5Ghz on air than it might be a solid alternative to 1100T,but even at close to 4.8-5Ghz it will be slower or on par with 4Ghz Thuban.

edit: I have no clue how AMD can ask 190-200$ for FX6100 versus 200$ for 1100T. Thuban is being EOLed ,true,but we should expect that its price will go down after the FX launch so it makes even less sense to price FX6100 at the present level of 1100T now that we "unofficially" know how 8150 performs...

[PerryR]

There is a lot that doesn't add up, and the fact that certain individuals are excited by bulldozer is enough for me to remain confident that the rumors being pumped out right now are not the full story. On the other hand, it is rather confusing to see the varied performance in these rumors. None of it looks particularly wonderful as of yet..

I'm still optimistic. Unless AMD underestimated the performance of their own product.............Some of these "leaks" make it sound like a prettied up Thuban on 32mm.

Who knows, maybe Lee's a paid Intel shill! LOL! I kid, I kid!

FX6100 with the leaked specs (3.3Ghz stock/3.9GhzT),and if these 8150 performance number are true, has no chance against Thuban 1100T.

I kinda knew that, but didn't want to outright indicate that the 6100 would be a fail in comparison to older arch. Damn, what the hell is going on!? Where's JF with a "that's BS" comment when you need him? LOL.

[Leeghoofd]

Who knows, maybe Lee's a paid Intel shill! LOL! I kid, I kid!

darn I'm busted now

plz ignore all of the previous comments I made before :p

[PerryR]

darn I'm busted now

LOL! I knew it!

[EniGmA1987]

unless the FX-6100 ( I assume 4 cores 6 threads) I'm eyeing is only slightly better than a 1090t

Its 3 modules and 6 threads for the FX-6100. A single module is disabled from the 8's

[SSJVegeta]

single threaded performance on CB10 looks to be around a 3.2GHz Conroe or something right?

According to leaked tests, pretty much yes.

My E4500 @ 3.3GHz in CB R10: Rendering (Single CPU): 3747 CB-CPU

Comes within 300 points of the single threaded CB R10 BD 8150 score.

[Apokalipse]

According to leaked tests, pretty much yes.

My E4500 @ 3.3GHz in CB R10: Rendering (Single CPU): 3747 CB-CPU

Comes within 300 points of the single threaded CB R10 BD 8150 score.

Yeah, I'm not convinced the numbers are really "right".

[SSJVegeta]

Yeah, I'm not convinced the numbers are really "right".

I agree. If they are right, AMD have just commited seppuku :x

[cal_guy]

More slides

http://www.techbang.com.tw/posts/720...87%E7%AB%A0%29

[TheBlueChanell]

Yeah. If they are going to do that then just spend a few extra weeks and fix it so the launch goes smoothly. We don't need a Sandy Bridge-esque fiasco and AMD can't really afford that, though I doubt it's of that caliber.

My point is we've waited years, whats a few more weeks. :p

[PerryR]

More slides

http://www.techbang.com.tw/posts/720...87%E7%AB%A0%29

Weird how the slide says 8mb l3 for all processors, but that review that was posted earlier states otherwise.

[duron]

I guess we should call in the "Myth Busters" to do an extreme test

[radaja]

Weird how the slide says 8mb l3 for all processors, but that review that was posted earlier states otherwise.

they just worded it wrong or the translation did it?
it looks like they confused the L2 with L3

The linked modules share L3 cache. With a quad core is 4MB, 6MB and one gets a hexacore octacore get 8MB. As much as the L2 memory is. Obviously the modules share the memory controller. This is however still dual channel, but can now officially 1866 MHz DDR memory along.

they did get the sizes right 4MB, 6MB, 8MB they just called it the L3 when its the L2 that changes with lower models

[Manicdan]

More slides

http://www.techbang.com.tw/posts/720...87%E7%AB%A0%29

one slide shows win 8 gives more gaming performance, either it means BD is being held back, or it means this could be the first windows to be better for gaming

[demonkevy666]

how do we know pre-caching is working or not ?

[PerryR]

they just worded it wrong or the translation did it?
it looks like they confused the L2 with L3

They indicated that it was the same for L2 (specifically), so I assume they were mistaken.

Funny how those slides have a comparison against the 980x in games, particularly when the 2600k already beats that for substantially less moo-la.

[slaveondope]

one slide shows win 8 gives more gaming performance, either it means BD is being held back, or it means this could be the first windows to be better for gaming

The L4D improvements are large. Whats Valves engine called again?

[PerryR]

It's the Source engine.

[BeepBeep2]

Source is multi-threaded...at least the Orange Box and more recent like the "L4D2 engine" variant.

It scales nicely with cores, and is also CPU intensive. It's nice to see a 10% improvement with just Win8's scheduler...taking me back to how 4 (of the) threads will have better IPC than 8 of course...

[flyck]

FX6100 with the leaked specs (3.3Ghz stock/3.9GhzT),and if these 8150 performance number are true, has no chance against Thuban 1100T. In MT apps it will lose hands down to 1100T,now that 8150 barely manages to beat it, while in ST it will score lower than 1100t @ 3.7Ghz (which is its max Turbo). Potential buyers are left with higher OC headroom though. If FX6100 can go to ~5Ghz on air than it might be a solid alternative to 1100T,but even at close to 4.8-5Ghz it will be slower or on par with 4Ghz Thuban.

edit: I have no clue how AMD can ask 190-200$ for FX6100 versus 200$ for 1100T. Thuban is being EOLed ,true,but we should expect that its price will go down after the FX launch so it makes even less sense to price FX6100 at the present level of 1100T now that we "unofficially" know how 8150 performs...

You can add Deneb 3.7GHz to be a burden for that FX also..
Hence the whole current FX4 lineup would simply be worse than the current deneb lineup. (they overclock better but thats it).
If the FX8150 barely edges the x6 1100T
FX8120 which performs up to 17% less in MT is going to be worse.
FX8100 wont be competitive with the X6 1055T..
Fx6100 won't be competitive with X6 1055T nor x4 3.7GHz
FX4170 will trade blows with x4 3.7GHz
FX41xx will be betweeen x4 3GHz and x4 3.7GHz.

llano @3.1Ghz > Fx4100 in cpu performance.

[-Boris-]

First indication:

AMD A8-3850
Radeon HD 6550D
100W
4
2.9GHz
400
600MHz


AMD A8-3800
Radeon HD 6550D
65W
4
2.4/2.7GHz
400
600MHz


TDP significant lower while the only constant is gpu resources and clockspeeds.

You know that you can't read TDPs like that? If one chip gets 66W and one gets 63W, the first will be labeled 100W and the other 65W.

second indications:

Reviews A8 3850. They include power figures. it is trading power draw blows with similar clocked 45nm x4 parts.
http://www.anandtech.com/show/4476/amd-a83850-review/9
POwer draw figure cpu only is right where it is for gaming.

As I read it the GPU intensive load maxes out the TDP more than the CPU intensive load does.

Like i mentionned, that is the ideal situation, but the same is true for BD. Ideally on a good process BD would be far better than what they can do now with the current issues. Your hypotetical BD replacement with thuban cores in llano form would simply not do what on this process node what you would expect them to do.

And you are making bad conclusions. It's reasonable to believe that the modifications in the silicon and transistor types needed for a working GPU isn't optimal for a CPU. GPUs are generally made for low frequencies but higher density on the most power consuming circuits. The design desicions made to fit a working GPU might be bad for a CPU. Llano is not a good example of CPU performance on 32nm. If a Thuban on 45nm isn't to far behind BD then what would happen with a Thuban made on 32nm? It would be almost half the size of Bulldozer, and be capable of higher frequencies than the original Thuban. And still have room for optimizations. And if your theory is correct and 32nm is botched and worse than 45nm, then maybe AMD should stick to 45nm for a while.

[dess]

@Leeghoofd: Are turbo modes (all-cores-on one and max one) enabled and working right? Better check for it directly (clocks during tests)! Perhaps it's not enough to enable it in the BIOS. There is a CPU driver to be installed, I mean...

This was posted by Dresdenboy some time ago:
http://semiaccurate.com/forums/showp...&postcount=824

Sounds interesting...

If there is any truth to this then it might be OS scheduling problem or a platform(bios?) problem. But don't get your hopes high.

Could be also a microcode update. AFAIK it's a very low-level (non-x86) code that is downloaded into the CPU at POST from BIOS and modifies the workings of certain parts.

(Wouldn't want to be in the place of those programmers now who are working on it 24/7... Even less if they are trying to catch some f...ing bugs... You know, the stress, almost no sleeping for days because of the adrenaline, and all... Brrr... Cheers to them!)

Win7's scheduling also needs some adjustments (to say the least), but I doubt an update for it will come in a week (if ever).

Its 3 modules and 6 threads for the FX-6100. A single module is disabled from the 8's

You mean from the 4.

I still say something doesn't add up

Sure... Wouldn't think at all that the current performance fits their intentions and plans. There must be some little flaw that's hindering it. Unless they've seriously missed something fundamental, but I wouldn't think that either.

Can't wait how it pans out in the end... Stirring...

and be capable of higher frequencies than the original Thuban.

How much? Isn't it uarch that basically determine frequency, rather than the process (in normal circumstances, I mean the latter being neither bad nor something wonderful).

[TESKATLIPOKA]

-Boris-

As I read it the GPU intensive load maxes out the TDP more than the CPU intensive load does.

you should think about that more
idle 44W
load CPU +79W in this case gpu should have the same consumption as in idle
load GPU +82W in this case gpu is working hard but cpu is working too but not as hard as doing encoding, that game doesn't use every core or something like that.