AMD to start Bulldozer AM3+ production by March 2011, launch in April 2011

AMD to start Bulldozer AM3+ production by March 2011, launch in April 2011

Quote:

---

According to a roadmap leaked by a Swedish forum, the desktop version of Bulldozer comes by April. Engineering samples of Zambezi AM3+ CPU are expected in December 2010 and hopefully this has already happened.

Production candidate 32nm silicon should be ready by February 2011 and the CPU should be production ready by March 2011.

The launch should take place in April, which is the first month of Q2 2011. Many people expected this CPU at a later date, and this up to 8-core processor series might be something to get AMD much needed traction and attention.

The Scorpius platform also comes with AMD 990FX Northbridge that can take care of 2x16 PCIe Crossfire mode and it comes matched with SB950. The new SB950 southbridge supports Raid 5, 14 USB 2.0, 6 SATA ports at 6Gb/s as well as AMD overdrive. Unfortunately there is no native USB 3.0 support.

Scorpius also supports AMD 990X chipset with two 8X PCIe Crossfire capable slots, AMD 880G in AMD3+ version that has one 16X PCIe slot as well as integrated DX10.1 graphics card at 560MHz. The last chipset to support Scorpius needs is AMD 970 and this little one has only one PCIe16X slot, but the vast majority of people don't need more anyway.

---

the pics are clickable
http://morepic.ru/thumbs/llano1.png http://morepic.ru/thumbs/llano2.png http://morepic.ru/thumbs/llano3.png

http://morepic.ru/thumbs/llano4.png http://morepic.ru/thumbs/llano5.png http://morepic.ru/thumbs/llano6.png http://morepic.ru/thumbs/llano7.png

Source - www.amdforum.se forum

Also, the slides say that Llano is going to be out in July. Which means that xbitlabs was right about this

It also says these slides are october 2010 update, so outdated information.

What are "Husky" cores? I know bobcat and bulldozer...whats husky?

I really am hoping that AMD pulls out a home run on this one. Intel hasn't impressed me with the initial Sandy Bridge Reviews. They're like 10% max better than the i3/i5/i7's that are already out. If AMD can finally close that gap, I'll be happy to wait until Bulldozer hits the market.

Quote:

---

Originally Posted by Tenknics

What are "Husky" cores? I know bobcat and bulldozer...whats husky?

---

husky is a new revision of the K10 stars core used in deneb and thuban; it's intended to make Llano more cost efficient as it's smaller than bulldozer because it doesn't need a very powerful memory and cache system

Looks a bit faked to me, especially the fifth picture, looks like something made in paint. It's quite bad work, some of the lines are a pixel or two to long.
There is something with the language used that don't look that proffesional too. And what about the longevity charts? Why don't they say anything about longevity?
Haven't seen any official Bulldozer charts mention Zosma either.

if they release zacate under 4.5W, we might even see systems powered by AA batteries next year.

How much likely do you concider an April launch? With the somewhat luckluster release of SB with limited bclk possibility versus AMD's "all-in" attempt to raise from the shadows of Intel in order to try and break out from the current scenario when Intel controls the performance and sets the pricing it's rather tempting to sit back and wait and see but I would hate to find out that Bulldozer would get delayed to say June/July for example.

First, I don't know if those slides are real because I don't work in that group, but where are you seeing launch in April? All it says is IP in April.

I'm not sure what IP stands for in this case, but if Brazos has IP in november and is supposedly launching in January then wouldn't June be a better guess for Zambezi?

Quote:

---

Originally Posted by Darakian

I'm not sure what IP stands for in this case, but if Brazos has IP in november and is supposedly launching in January then wouldn't June be a better guess for Zambezi?

---

Actually AMD started to ship first Zacate chips to its customers in November,on the day when analyst day was held.

Quote:

---

“As of 4 am Central Time, our facility in Singapore, our test and manufacturing facility, shipped the very first Fusion APUs - production units - to customers worldwide," said AMD senior vice president and general manager Rick Bergman.

---

link

Wooo, if they do come out in april, I can hold off on building a new PC until then... can't wait to see some #'s since SB hasn't been as awesome as I thought

Will AM3+ mobos be available before Zambesi launch? Since they're bc with current AM3 chips folks like me can upgrade to a futureproof system skipping AM3 motherboards.

Quote:

---

Originally Posted by RPGWiZaRD

How much likely do you concider an April launch? With the somewhat luckluster release of SB with limited bclk possibility versus AMD's "all-in" attempt to raise from the shadows of Intel in order to try and break out from the current scenario when Intel controls the performance and sets the pricing it's rather tempting to sit back and wait and see but I would hate to find out that Bulldozer would get delayed to say June/July for example.

---

Well, I personally don't doubt that AMD can release their next generation CPU in April for the masses, but I am in doubt that they can match Intel's progress in CPU tech.
Which means that I am looking forward to build a bulldozer rig but I am not looking forward its performance to be on par with SB.

With SB 4 core, and don't see why Zambesi can't be on his level or even surpass it. SB 8 core will be faster, and more expensive, same as 980X vs X6.

SB 4 core is slower than current i7 six-core, there are no core increase just a bit more ipc and other tweaks. From Thuban --> Zambesi we have core increase, ipc, more cache, new memory controller, updated optimizations (SSE4.1-4.2 which benefit i7 in some programs over the X6 such as the adobe suite, and also AVX -available for both intel&amd-).

Maybe that's a reason why people are not too impressed with sb launch (they wanted 6-8 core version from the get go).

Well for the mainstream SB i guess they match and even surpass them, but with the S2011 they will have a hard time.

I think it is reasonable to have conservative expctations. Realistically we can't say with any certainty how these chips will perform. To many factors changed between architectures.

I suspect it may be another situation where it depends on what you are using it for. It may be slow at single thread tasks, but have strong multithreaded performance.

Quote:

---

Originally Posted by zerazax

Wooo, if they do come out in april, I can hold off on building a new PC until then... can't wait to see some #'s since SB hasn't been as awesome as I thought

---

This version is Sandy Bridge isn't in the same league as Bulldozer. Wait until the cpus meant for socket 2011 to make that comparison.

Sandybridge that's out right now should be compared to Bobcat.

Frankly I don't see Bulldozer or Sandybridge EX as comparable. They are both filled with tradeoffs which will make choosing either one for my future build incredibly difficult to pull of especially since no review out there will cover the issues I would need to be better informed of my suspicions on what I should expect depending on cpu platform.

While I'm looking forward to Bulldozer the chipset is completely underwhelming based on those slides. Hopefully they will relinquish whitepapers on their chipset when asked for it.

Quote:

---

Originally Posted by mutantmagnet

While I'm looking forward to Bulldozer the chipset is completely underwhelming based on those slides. Hopefully they will relinquish whitepapers on their chipset when asked for it.

---

Underwhelming in what way? The specs seem reasonable to me and i'm looking forward to iommu support. Should give VMs a good performance boost.

Quote:

---

Originally Posted by mutantmagnet

Sandybridge that's out right now should be compared to Bobcat.

---

Please tell me you are not serious here... These 2 are made for 2 different segments.Zacate(since Bobcat is just a x86 core isnide it) is a tiny chip compared to SB.Just because both have CPU+IGP doesn't mean they fit in the same product or price segment. Llano is SB's competitor in mainstream desktop and notebook segments.
8C Bulldozer is SB 6/8 core competitor in high performance/enthusiast segment.

Quote:

---

Originally Posted by Pantsu

It also says these slides are october 2010 update, so outdated information.

---

a roadmap doesn't change just like that, sure there might be a push-pull from a certain part but mainly they stay the same for several months unless they have some severe issues. So no they are not outdated, those roadmaps are fixed for 2-3months, they don't change every week, expect a new update early januari.

Quote:

---

Originally Posted by -Boris-

Looks a bit faked to me, especially the fifth picture, looks like something made in paint. It's quite bad work, some of the lines are a pixel or two to long.
There is something with the language used that don't look that proffesional too. And what about the longevity charts? Why don't they say anything about longevity?
Haven't seen any official Bulldozer charts mention Zosma either.

---

perhaps change the question a bit, have you ever seen real official slides which you can compare to?

Quote:

---

Originally Posted by Hornet331

Well for the mainstream SB i guess they match and even surpass them, but with the S2011 they will have a hard time.

---

what a hardtime with what? You really think AMD will only be able to scale zambezi against 1155 and 2011???

Quote:

---

Originally Posted by zerazax

Wooo, if they do come out in april, I can hold off on building a new PC until then... can't wait to see some #'s since SB hasn't been as awesome as I thought

---

We haven’t seen the AVX performance advantage yet

Quote:

---

Originally Posted by informal

Please tell me you are not serious here... These 2 are made for 2 different segments.Zacate(since Bobcat is just a x86 core isnide it) is a tiny chip compared to SB.Just because both have CPU+IGP doesn't mean they fit in the same product or price segment. Llano is SB's competitor in mainstream desktop and notebook segments.
8C Bulldozer is SB 6/8 core competitor in high performance/enthusiast segment.

---

Even Llanos CPU cores are half the size compared to SBs. Llano cannot compete directly to SB (Core to Core) but a 4-core Llano could be competitive (Price and Performance) to a 2 Core SB.

I'll be pissed if they launch before April, I *really* need a break :ROTF:

Quote:

---

Originally Posted by Aten-Ra

We haven’t seen the AVX performance advantage yet

Even Llanos CPU cores are half the size compared to SBs. Llano cannot compete directly to SB (Core to Core) but a 4-core Llano could be competitive (Price and Performance) to a 2 Core SB.

---

this time around Husky will also feature the turbocore 2.0... so they will compete much more into the single IPC parts and will totally destroy the HT cores.

Quote:

---

Originally Posted by mutantmagnet

Sandybridge that's out right now should be compared to Bobcat.

---

huh

Maybe true in ULV field but not so in portable or desktop. In ULV the Intel SB's will have more CPU power but lack IGP punch of a Zacate in most case "Even if the SB's are with 12 EU's they will be on a diet"

Quote:

---

Originally Posted by duploxxx

a roadmap doesn't change just like that, sure there might be a push-pull from a certain part but mainly they stay the same for several months unless they have some severe issues. So no they are not outdated, those roadmaps are fixed for 2-3months, they don't change every week, expect a new update early januari.

---

There's a difference between public and confidential information, quite a huge one at that. Public information is almost like a promise, while confidential information is only an estimate to help partners anticipate the coming of a new product. Confidential information is much more likely to change - granted, not just like that, but you definitely can't pin AMD down on it as this info isn't meant as such a promise.

And, again, it's about IP, not launch! I believe IP stands for "Industrial Production" (or otherwise 'initial production'), which indicates when they start producing. CPUs aren't done in a couple of hours, they need a couple of runs through the fab, then they need packaging (which is probably done at a completely different location), then distribution, the reviewers need some time... Then AMD still needs to approve launch. Expect May at the very earliest - if these slides are still up to date (and also not fake), that is.

Quote:

---

Originally Posted by informal

Please tell me you are not serious here... These 2 are made for 2 different segments.Zacate(since Bobcat is just a x86 core isnide it) is a tiny chip compared to SB.Just because both have CPU+IGP doesn't mean they fit in the same product or price segment. Llano is SB's competitor in mainstream desktop and notebook segments.
8C Bulldozer is SB 6/8 core competitor in high performance/enthusiast segment.

---

Right.. and the Sandybridge that exist now isn't meant for the performance/enthusiast segment which Bulldozer is going to cover. That's my point and not mentioning Llano along with Bobcat shouldn't have thrown you off.

Quote:

---

Originally Posted by BrowncoatGR

Underwhelming in what way? The specs seem reasonable to me and i'm looking forward to iommu support. Should give VMs a good performance boost.

---

Iommu was the only thing I didn't know the exact details of when I made my statement. After some preliminary research that is something I'm actually excited for and see as an important feature being implemented. Everything else didn't impress me.

Quote:

---

Originally Posted by Nintendork

With SB 4 core, and don't see why Zambesi can't be on his level or even surpass it. SB 8 core will be faster, and more expensive, same as 980X vs X6.

SB 4 core is slower than current i7 six-core, there are no core increase just a bit more ipc and other tweaks. From Thuban --> Zambesi we have core increase, ipc, more cache, new memory controller, updated optimizations (SSE4.1-4.2 which benefit i7 in some programs over the X6 such as the adobe suite, and also AVX -available for both intel&amd-).

Maybe that's a reason why people are not too impressed with sb launch (they wanted 6-8 core version from the get go).

---

Whats more important, SB with 8 cores will be not relase (its latest info what I have)...


Zambezi in April/May is think really possible and good point for AMD. Look here SB 2500k in CInebench-AIRcooled, its dangerous score.

would it delay again

rather in may

http://www.techpowerup.com/img/10-11-18/133a.jpg

dont panic, AMD is last year stroneger and stronger. Really think, Zambezi will leader with Gulftowns in multithreading - in 2011 year

wow Bulldozer... I first heard of Bulldozer when i was a little kid. I always thought DNF would be released before Bulldozer, i was wrong apparently.

BD codename was mentioned,for the first time, in 2007 so you couldn't possibly hear it when you were a kid(or maybe you could,are you a kid now :) ?)
BD on 45nm was scheduled for late 2009 but was canceled. This BD version,done on 32nm and scheduled for Q2 2011,is much improved design based on the first(canceled) version. So a delay is 1.5 years or so,from late 2009 to early 2011. I can't see any DNF theme in BD development,at all.

Zambezi 8 cores have a release date, What about 8 Core SB's? All we hear about is 1155.. Anything?

Quote:

---

Originally Posted by Johnny87au

Zambezi 8 cores have a release date, What about 8 Core SB's? All we hear about is 1155.. Anything?

---

September supposedly?

Quote:

---

Originally Posted by zerazax

September supposedly?

---

so late after zambezzi ???

No 8 cores SB for 1155, it's only for 2 cores and 4 cores
It's only for 1356 with 6 cores and 8 cores for Extreme Edition

Quote:

---

Originally Posted by Sn0wm@n

so late after zambezzi ???

---

Q3 2011 apparently

Quote:

---

Originally Posted by zerazax

Q3 2011 apparently

---

september it is then

Quote:

---

Originally Posted by xlimit

No 8 cores SB for 1155, it's only for 2 cores and 4 cores
It's only for 1356 with 6 cores and 8 cores for Extreme Edition

---

From what i herd they ditched 1356 and moving on with LGA2011 :shrug:

six-, octal-core SB - socket 2011 only (autumn 2011)

1366 has no future

Has no future? They're releasing the 990x surely cant be that bad.. No reviews or tests as with the LGA2011..

what price range will Zambezi 8 cores be? around 600$ maybe ?

Quote:

---

Originally Posted by MAS

six-, octal-core SB - socket 2011 only (autumn 2011)

1366 has no future

---

1356/66 has a future, theres nothing wrong with the socket. Instead LGA 2011 is just a rumor, and its been rumors about LGA2011 will come with ivy bridge in 2012. Theres also been rumors about SB with 4-8 cores will use 1356/66 socket. NOTHINGS written in stone, its all rumors.

Quote:

---

Originally Posted by EagleRock

what price range will Zambezi 8 cores be? around 600$ maybe ?

---

I would say where or a little above where 1090T is today. Bulldozer is a very compact chip and the 8 core model will prolly be smaller then todays phenom II X4 and just a bit bigger then SB with 4 cores.

Quote:

---

Originally Posted by Kristers Bensin

1356/66 has a future, theres nothing wrong with the socket.

---

who will buy expensive 1366 mobo and slow 1366 cpu after SB launch?

for instance, i930 + X58 mobo have the same price as 2600K + 1155 mobo. Plus in soc.1366 case you have to buy 50 per cent more memory

no...SB is only 1155 LGA...At 4Q 2011 or 1Q 2012 we will see new LGA 2011 with IB CPUs...Its a bit change...
AND NOW BACK TO BULLDOZER CHIP :).

Quote:

---

Originally Posted by FlanK3r

no...SB is only 1155 LGA...At 4Q 2011 or 1Q 2012 we will see new LGA 2011 with IB CPUs...Its a bit change...
AND NOW BACK TO BULLDOZER CHIP :).

---

No ivybrdige will come first for S1155, S2011 will get SB. Its the same as it was with nehalem/westmere. The only difference is, that this time consumer products are released first and servers come later.

Quote:

---

Originally Posted by Hornet331

No ivybrdige will come first for S1155, S2011 will get SB. Its the same as it was with nehalem/westmere. The only difference is, that this time consumer products are released first and servers come later.

---

Well servers and high end desktop. I guess with BD being more or less an unknown it made sense for Intel to play it safe and give themselves time to adjust their strategy if BD turns out better than expected. Anyone know if S2011 will have IOMMU support?

Quote:

---

Originally Posted by FlanK3r

no...SB is only 1155 LGA...At 4Q 2011 or 1Q 2012 we will see new LGA 2011 with IB CPUs...Its a bit change...
AND NOW BACK TO BULLDOZER CHIP :).

---

well, check your sources again, many people already have 6c & 8c SB LGA2011 engineer samples

yes, ES at 32nm SB are, but one man said me, for production at 32nm is LGA2011 cancelled....(and at October 2010 roadmap I seen not info about LGA2011 :-) )

why is it always the same on this forum that i come into a thread titled amd bulldozer, and leave because all thats in it is :banana::banana::banana::banana: about intels next :banana::banana::banana::banana:... what is it with the mods here?

Quote:

---

Originally Posted by nunchucka

why is it always the same on this forum that i come into a thread titled amd bulldozer, and leave because all thats in it is :banana::banana::banana::banana: about intels next :banana::banana::banana::banana:... what is it with the mods here?

---

Fugger + Intel perhaps? Don't know, but it would be nice for Intel users to stay out of AMD threads and vice versa.

yeahh more about bulldozer ... less about sandy bridge ....

When do you escape some result with ES Zambezi?

It's going to be a very tight seal on BD info until launch in Q2.

Quote:

---

Originally Posted by informal

It's going to be a very tight seal on BD info until launch in Q2.

---

Just like there release of the 6xxx and probably the same with Antilles

Quote:

---

Originally Posted by Johnny87au

Just like there release of the 6xxx and probably the same with Antilles

---

considering the performance preview of bobcat i expect a preview in march (maybe earlier) together witht he 900 series chipset ;)

any leaked prices?

lol, prices and now? about 5 months before launch?:-D

All I care about is 1) IPC, 2) Overclockability, and 3) thermals.

Until that is known, all this is fluff.

add 2) seems good from design...But, time show more...

Quote:

---

Originally Posted by mstp2009

All I care about is 1) IPC, 2) Overclockability, and 3) thermals.

Until that is known, all this is fluff.

---

The realworldtech article gives us some hints to some of these questions. The indicators are for medium IPC and high frequency. Of course thermals probably won't be known until launch.

Quote:

---

Originally Posted by mstp2009

All I care about is 1) IPC, 2) Overclockability, and 3) thermals.

Until that is known, all this is fluff.

---

IPC is useless without frequency. What would you like the most? 1: 500MHz at 125W with a performance equal to a X6 1055T? Or 2: 8000MHz at 95W with a performance 40% higher than i7 980X?
You want performance, thermals and overclockability. IPC and frequency are secondary. ;)

Quote:

---

Originally Posted by -Boris-

IPC is useless without frequency. What would you like the most? 1: 500MHz at 125W with a performance equal to a X6 1055T? Or 2: 8000MHz at 95W with a performance 40% higher than i7 980X?
You want performance, thermals and overclockability. IPC and frequency are secondary. ;)

---

IPC * clockspeed = performance [except in cases explicitly independent of system performance]

What is wanted is an optimal mix of performance, power, and efficiency.

Quote:

---

Originally Posted by FlanK3r

lol, prices and now? about 5 months before launch?:-D

---

I predict the prices will be inline, with intels equivalent products. THere may possibly be a premium on server chips, if the thermals for multi-core processing are under control.

damn I'm good.

Quote:

---

Originally Posted by nn_step

IPC * clockspeed = performance [except in cases explicitly independent of system performance]

What is wanted is an optimal mix of performance, power, and efficiency.

---

As I said:
"IPC is useless without frequency."

And how do you define efficiency? Isn't a good efficiency an optimal mix of performance and power? Well and maybe die size.

Quote:

---

Originally Posted by Solus Corvus

The realworldtech article gives us some hints to some of these questions. The indicators are for medium IPC and high frequency. Of course thermals probably won't be known until launch.

---

wait then what is you base for IPC ? cpu wise it must have a control you getting that from?

?:confused:?

Quote:

---

Originally Posted by demonkevy666

wait then what is you base for IPC ? cpu wise it must have a control you getting that from?

?:confused:?

---

I mean probably higher then P2 but lower than SB. They look to be making up for it with clocks and cores.

Quote:

---

Originally Posted by Solus Corvus

I mean probably higher then P2 but lower than SB. They look to be making up for it with clocks and cores.

---

Yeah, I have a feeling that IPC will be very similar to Nehalem.

Quote:

---

Originally Posted by -Boris-

As I said:
"IPC is useless without frequency."

And how do you define efficiency? Isn't a good efficiency an optimal mix of performance and power? Well and maybe die size.

---

efficiency is how well the processor fits the task at hand. For example, massively parallel tasks tend to be far better suited towards simple in-order cores. Were as single threaded but compute heavy tasks tend to be far better suited towards super heavy OoO cores.

No, there's different kinds of efficiency

From the product perspective, it's how much performance can you get for the given constraints

Quote:

---

Originally Posted by Darakian

I'm not sure what IP stands for in this case, but if Brazos has IP in november and is supposedly launching in January then wouldn't June be a better guess for Zambezi?

---

Initial Production?

AMD keep silence. Four months before launch - no tests. The most likely scenario is that zambezi has single-thread performance level compared to nehalem (clock-for-clock), in other words zambezi is faster than K10.5 by the same per cent K10.5 is faster than K8 (10 per cent)

The main advantage of bulldozer is therefore its eight (pseudo-)cores.

Quote:

---

Originally Posted by MAS

AMD keep silence. Four months before launch - no tests. The most likely scenario is that zambezi has single-thread performance level compared to nehalem (clock-for-clock), in other words zambezi is faster than K10.5 by the same per cent K10.5 is faster than K8 (10 per cent)

The main advantage of bulldozer is therefore its eight (pseudo-)cores.

---

No tests, no previews = Sandy bridge road cleaning ...

Quote:

---

Originally Posted by MAS

The main advantage of bulldozer is therefore its eight (pseudo-)cores.

---

Pseudo? What do you mean by "pseudo"?

Quote:

---

Originally Posted by MAS

The main advantage of bulldozer is therefore its eight (pseudo-)cores.

---

God this terminology inconsistency is frustrating. Bulldozer should have 16 cores in 8 modules. If it came out with 8 cores it wouldn't be competitive with Sandybridge.

Quote:

---

Originally Posted by mutantmagnet

God this terminology inconsistency is frustrating. Bulldozer should have 16 cores in 8 modules. If it came out with 8 cores it wouldn't be competitive with Sandybridge.

---

4 modules - 8 cores on desktop
8 modules - 16 cores on server only for now

so amd has pseudo cores now ??? LOLL

Quote:

---

Originally Posted by Sn0wm@n

so amd has pseudo cores now ??? LOLL

---

So does Intel with HT on. Welcome to 2002. :p:

Quote:

---

Originally Posted by zalbard

So does Intel with HT on. Welcome to 2002. :p:

---

since when does amd has pseudo cores???? intel has ht and yes i know about that ...


but please help me figure out why amd has pseudo cores

http://www.pcper.com/images/reviews/985/bd_slide3.jpg

http://www.dvhardware.net/news/2010/...hips_slide.jpg

Also, Google is your friend... Same for the search function on these forums. :p:

not this again :rolleyes:

Quote:

---

Originally Posted by zalbard

http://www.pcper.com/images/reviews/985/bd_slide3.jpg

http://www.dvhardware.net/news/2010/...hips_slide.jpg

Also, Google is your friend... Same for the search function on these forums. :p:

---

and those slides proves your right??? LOLL



i thought that it showed the full physical hardware of 2 real cores ...

AMD won't use module terminology when it ships Zambezi. It will be cores only(since it *really* has 8 real cores).

zalbard
Modules consistes of two cores, not the virtual ones found on HT.

So by given an equal performance (asuming here and there)
1 core --> 100%
1 core + HT --> <100%, 120% (as long as the core are not fully loaded such as Linx bench)
2 core --> 180%~
1 module (two core) --> 160%?~


This was funny:

http://www.madboxpc.com/wp-content/u...lldozer_FX.jpg

8 cores Zambesi under Vision Black FX label
4-6 cores Zambesi under Ultimate Vision FX label

http://www.donanimhaber.com/islemci/...en-doguyor.htm

Dr Who pisses in his pants :D

:-d

Are there benchmark numbers leaked? :D

It's strange that we're 4 months before the launch and we don't have any number :confused:

At least these numbers keep me away from buying 2500K/2600K :yepp:

Zambesi am3+ only?

Quote:

---

Originally Posted by Hell Hound

Zambesi am3+ only?

---

yes ... its been know for months

Quote:

---

Originally Posted by MAS

AMD keep silence. Four months before launch - no tests. The most likely scenario is that zambezi has single-thread performance level compared to nehalem (clock-for-clock), in other words zambezi is faster than K10.5 by the same per cent K10.5 is faster than K8 (10 per cent)

The main advantage of bulldozer is therefore its eight (pseudo-)cores.

---

A. No benchmarks before launch, that is standard AMD policy.
B. They are real cores. Period.

Quote:

---

Originally Posted by Olivon

No tests, no previews = Sandy bridge road cleaning ...

---

See above.

Quote:

---

Originally Posted by Nintendork

zalbard
Modules consistes of two cores, not the virtual ones found on HT.

So by given an equal performance (asuming here and there)
1 core --> 100%
1 core + HT --> <100%, 120% (as long as the core are not fully loaded such as Linx bench)
2 core --> 180%~
1 module (two core) --> 160%?~

[/url]

---

Not quite. 1 module with 2 cores would be ~180%. 2 threads running on 2 different modules would be ~190-195%.

Quote:

---

Originally Posted by amdsempron_xs

Are there benchmark numbers leaked? :D

It's strange that we're 4 months before the launch and we don't have any number :confused:

At least these numbers keep me away from buying 2500K/2600K :yepp:

---

No, those are not benchmarks, benchmarks come out at launch.

Quote:

---

Originally Posted by Hell Hound

Zambesi am3+ only?

---

yes

Quote:

---

Originally Posted by JF-AMD

A. No benchmarks before launch, that is standard AMD policy.
B. They are real cores. Period.

See above.

Not quite. 1 module with 2 cores would be ~180%. 2 threads running on 2 different modules would be ~190-195%.

No, those are not benchmarks, benchmarks come out at launch.

yes

---

Your biggest problem will be the fud slung from the 'opposite' side. There are some so called experts on this forum, who struggle with the core/module type thing. I fully expect that to be exploited by the other camp.

It's a nice elegant solution to having extra cores with less die space, however intel on a smaller node will be able to compete core for core at some point, without the performance tradeoff.

The only question is can you keep the 'core' lead.

Quote:

---

Originally Posted by JF-AMD

A. No benchmarks before launch, that is standard AMD policy.
B. They are real cores. Period.

---

wrong policy. I'm quite sure Intel already has bulldozer chips and is aware of its performance level.

So nondisclosure bulldozer info doesn't help AMD

Your company needs brief zambezi testing, which can show strong sides of new cpu

Quote:

---

Originally Posted by Motiv

Your biggest problem will be the fud slung from the 'opposite' side. There are some so called experts on this forum, who struggle with the core/module type thing. I fully expect that to be exploited by the other camp.

---

Exactly. However I think it's pretty clear that the other camp is already trying to exploit it. A handful of well placed shills can have a pretty big affect. :D

Quote:

---

It's a nice elegant solution to having extra cores with less die space, however intel on a smaller node will be able to compete core for core at some point, without the performance tradeoff.

The only question is can you keep the 'core' lead.

---

Which tradeoff? I think it's the other way around, intel is trying to pseudo mock a dual core with dual threads. I see the tradeoff being slightly increasing performance while reducing die space.

I'm not sure what the tradeoff of replacing 2 full cores with 2 full but faster cores would be. There is no evidence whatsoever to point to and say a SB die size is a good tradeoff over an Interlagos, Valencia or Zambezi die size.

Quote:

---

Originally Posted by MAS

wrong policy. I'm quite sure Intel already has bulldozer chips and is aware of its performance level.

So nondisclosure bulldozer info doesn't help AMD

Your company needs brief zambezi testing, which can show strong sides of new cpu

---

There is the Osbourne Affect, however I wonder how big that really is. CPU's are dirt cheap anyway.

Quote:

---

Originally Posted by MAS

wrong policy. I'm quite sure Intel already has bulldozer chips and is aware of its performance level.

So nondisclosure bulldozer info doesn't help AMD

Your company needs brief zambezi testing, which can show strong sides of new cpu

---

So that people can immediately stop buying current product and wait? This isn't about intel.

Quote:

---

Originally Posted by JF-AMD

So that people can immediately stop buying current product and wait? This isn't about intel.

---

Great. You get to piss off customers twice. The ones who will wait will be pissed if they find out that BD sucks and the ones who don't wait and buy SB will be pissed off if they find out that BD is great.

AMDs' policy is totally :banana::banana::banana::banana:ed up. Release a few numbers or loose customers.

Let me get this straight.....I'm average Joe who wants to purchase a computer one week from now. I have a look at Intel and discover this fantastic chip called Sandy Bridge, great performance for the value. I say to myself, "I wonder what AMD has to compete?" Looks around, nothing to compete agains SB. I think I'll buy SB. So how are you losing sales when the person has no intention of buying AMD chips anyhow? Don't you ever think to yourself that by releasing a few benches, you can stem the tide of people (even a half of 1%) from purchasing an Intel setup? Think outside the box man.

And I thought Hector ran the company badly!

Quote:

---

Originally Posted by freeloader

Let me get this straight.....I'm average Joe who wants to purchase a computer one week from now. I have a look at Intel and discover this fantastic chip called Sandy Bridge, great performance for the value. I say to myself, "I wonder what AMD has to compete?" Looks around, nothing to compete agains SB. I think I'll buy SB. So how are you losing sales when the person has no intention of buying AMD chips anyhow? Don't you ever think to yourself that by releasing a few benches, you can stem the tide of people (even a half of 1%) from purchasing an Intel setup? Think outside the box man.

And I thought Hector ran the company badly!

---

Your average joe wouldn't even know about BD, or be reading forums like these to look for some supposedly leaked benchmarks that might be true or false. They also wouldn't know what SB is. Your average joe would just walk into best buy and purchase a computer there. AMD can either release their CPU's at the same time as SB, or just keep doing what they are doing since it seems to work fine.

Quote:

---

Originally Posted by MAS

wrong policy. I'm quite sure Intel already has bulldozer chips and is aware of its performance level.

So nondisclosure bulldozer info doesn't help AMD

Your company needs brief zambezi testing, which can show strong sides of new cpu

---

if intel has acces to an amd cpu that is still withing nda ... dont you think it could be seen as corportate espionage ... a thing that is punishable with a prison sentense if its proven to be true .... come on ....


and the nda thing is the same with all big corporations .... saying too much hurts big corporations ...

Quote:

---

Originally Posted by freeloader

Great. You get to piss off customers twice. The ones who will wait will be pissed if they find out that BD sucks and the ones who don't wait and buy SB will be pissed off if they find out that BD is great.

AMDs' policy is totally :banana::banana::banana::banana:ed up. Release a few numbers or loose customers.

Let me get this straight.....I'm average Joe who wants to purchase a computer one week from now. I have a look at Intel and discover this fantastic chip called Sandy Bridge, great performance for the value. I say to myself, "I wonder what AMD has to compete?" Looks around, nothing to compete agains SB. I think I'll buy SB. So how are you losing sales when the person has no intention of buying AMD chips anyhow? Don't you ever think to yourself that by releasing a few benches, you can stem the tide of people (even a half of 1%) from purchasing an Intel setup? Think outside the box man.

And I thought Hector ran the company badly!

---

so an average joe that cant wait surely knows how to control a multi billion dollars corporation .... sure ...

this thread went lol