AMD's Bobcat and Bulldozer

[Sn0wm@n]

@JF-AMD: would it be possible to get us access to the lead client side marketing guru ???? since xs is mostly about client side cpu's it would make more sense ... or could we get some info from time to time that is relayed by the client side

[Mats]

These Perf Numbers suggest an i7 2600k will smash an i7 880..

You obviously haven't read the preview, right?

[Mats]

Still not a new socket? Well they've already said they had to do AM3+ and make bulldozer incompatible with AM3 to ensure a performance boost(EDIT: More performance boost compared to sticking to AM3, I mean). No LGA, no PCI-E, is that something to complain? Do tell me if there's a performance difference beyond the margin of errors in putting the PCI-E lanes on the chipset or on the processor. Memory channels, yes, that's only one thing though.

You're missing my point. I'm not the one who's missing LGA and PCIe, I'm talking about AMD. They're going to add it, so why not add the pins now?
AMD wouldn't upset many people by doing this since the most anticipated compatibility is gone anyway.

Replacement probably comes when the next generation comes.

Exactly! If BDs first socket was made the right way, it could continue to work with the next gen BDs, but now we got AM3+ which is neither enough compatible with old boards, nor is modern enough to stay for some generations.

Heck, do we know if AM3+ is the same socket as AM3? We know that AM2+ and AM2 are the same sockets. Nothing that's stopping them from adding a wee bit more pins around the corners here and there and still maintain compatibility like how AM3 processors can fit into AM2+.

Been there.. . http://www.xtremesystems.org/forums/...8&postcount=43

[Sn0wm@n]

the next socket would most likely arive with the 22nm node ....

[Mats]

the next socket would most likely arive with the 22nm node ....

Which would make AM3+ relatively shortlived just because AMD won't make a completely new socket this time, and that's my point.

It's not the end of the world, upgrading CPU doesn't always work well for various reasons. You can put a X6 in a four year old 690G motherboard, but you probably won't..

[nn_step]

But when will the details of Bulldozer's decode finally be revealed

[Sn0wm@n]

Which would make AM3+ relatively shortlived just because AMD won't make a completely new socket this time, and that's my point.

It's not the end of the world, upgrading CPU doesn't always work well for various reasons. You can put a X6 in a four year old 690G motherboard, but you probably won't..

i tought your point was that making it non compatible in a way but still keep backward compatibility was bad ???

[blindbox]

Been there.. . http://www.xtremesystems.org/forums/...8&postcount=43

Right back at ya... exactly!

That means future updates CAN have PCI-E lane on the processor. They're just sticking to pins that's all

[mAJORD]

You obviously haven't read the preview, right?

I have, and that's why I said it.

[savantu]

Which would make AM3+ relatively shortlived just because AMD won't make a completely new socket this time, and that's my point.

Llano will have a different socket from BD.
Mainstream parts will have a socket that enables PCI-E. Performance parts will not. Just like Intel ( with LGA 1155/1366 )

[terrace215]

So in order avoid giving intel info on the competition of their unreleased SB chips, AMD is keeping quiet.

The design of those parts is fixed. Silicon samples are out there already. AMD is keeping quiet because BD silicon is not yet in a state to be sent out for review. They just taped the thing out last quarter... you can't expect them to be sending out A0 silicon for review..

[Sn0wm@n]

The design of those parts is fixed. Silicon samples are out there already. AMD is keeping quiet because BD silicon is not yet in a state to be sent out for review. They just taped the thing out last quarter... you can't expect them to be sending out A0 silicon for review..

agreed .... how much more time do you think before amd sends out review sample ??? 2 3 months ???

[JumpingJack]

agreed .... how much more time do you think before amd sends out review sample ??? 2 3 months ???

It is typically 9 months to a year after first tape out that products finalize and have enough volume to lanch. My guess is that we will see first leaked BD numbers around marchish timeframe of next year.

For example, Barcelona was demonstrated in Decemberish of 2006 (taped out like August-septemberish of 06) and we saw first products around novemberish of 07...

I don't think it matters really, regardless of when it comes it is going to be fun to dissect it.

[qcmadness]

It is typically 9 months to a year after first tape out that products finalize and have enough volume to lanch. My guess is that we will see first leaked BD numbers around marchish timeframe of next year.

For example, Barcelona was demonstrated in Decemberish of 2006 (taped out like August-septemberish of 06) and we saw first products around novemberish of 07...

I don't think it matters really, regardless of when it comes it is going to be fun to dissect it.

I don't think AMD will let the chips to be leaked.

Shanghai / Deneb has nearly zero leaked info vs official info.

[JumpingJack]

I don't think AMD will let the chips to be leaked.

Shanghai / Deneb has nearly zero leaked info vs official info.

I am not so sure -- if they have a strong winner on their hands, they may resort to the same guerilla marketing approach as Intel. Freeze purchasing decisions so to speak.

One thing you can bank on though, the BD launch will not be like Barcelona.

[PatRaceTin]

Both Intel & AMD need new board???

[JF-AMD]

@JF-AMD: would it be possible to get us access to the lead client side marketing guru ???? since xs is mostly about client side cpu's it would make more sense ... or could we get some info from time to time that is relayed by the client side

I am sure they read this site, but they really don't comment on things in public.

I live in a different world. Server people need to know about technology 12-18 months before they deploy. There are decisions that they need to make about the technology in order to put their plans in place. And when they buy, they buy thousands of servers on a very regular cadence. Telling a server customer what is coming a year or two from now causes them to say "great, let me put that into my plan, I will start buying them 90-120 days after they are available because I have to run them through qual and testing first."

Giving a consumer benchmarks today can stall the market. It is a different world.

[JF-AMD]

I don't think AMD will let the chips to be leaked.

Shanghai / Deneb has nearly zero leaked info vs official info.

I can't speak for the client side, but on the server side I have a very strong policy on no leaks, period. I cannot control 100% of the samples that we send out, but if some parts end up in the wild it is not our doing.

[god_43]

lol they are the apple of the silicon world.....though i do not care much for apple.....

[ajaidev]

I can't speak for the client side, but on the server side I have a very strong policy on no leaks, period. I cannot control 100% of the samples that we send out, but if some parts end up in the wild it is not our doing.

Intel has sent out about 500 samples of sandy bridge 8c in march - april and had no leaks but about a month or so back about 1100-1200 more sample were sent out along with documentation and there were leaks.

There were a whole tray of 32nm 6 cores stolen from Intels R&D center just months ago for which everyone had a inspection done.

The thing is that when small quantities are used the security is better and overall accountable but when the quantity increases security decreases. Also ES cpu's fetch high prices in the market, people who want them are willing to pay heavily not only in terms of money but also with system stability and performance because of the ES nature of the cpu.

Most samples i get form Intel are given with NDA and under a strict one piece per load, i cant have more than one cpu at a time. This makes leaks very hard in the early schemes of things.

[MAS]

http://blogs.amd.com/work/2010/08/23...ge-4/#comments

AMD guy promised bulldozer review next week (see comments)

when is round 2?


John Fruehe August 26, 2010

In legal review now, probably early next week.

[generics_user]

http://blogs.amd.com/work/2010/08/23...ge-4/#comments

AMD guy promised bulldozer review next week (see comments)

i think you REALLY misunderstood his post

Round 2 of the questions on Bulldozer are UNDER legal review to be posted; there will be no product review only more questions on Bulldozer

[superrugal]

http://blogs.amd.com/work/2010/08/23...ge-4/#comments

AMD guy promised bulldozer review next week (see comments)

He is definitely saying the "question round 2" not bulldozer review.

[ohnoitseddy]

Review? Far from it, just more answers to questions, akin to this: http://blogs.amd.com/work/2010/08/23...ons-round-one/

Info can get twisted so easily on the internet. And next AMD would be blamed for not giving "a full review as promised". :P

[JF-AMD]

It is round 2. They are actually through legal and will be posted on monday. I post them in the "AMD" forum because I see it as information, not "news".