SB is here. When will Bulldozer be released?

[edplayer]

Pure BS . Anand said between 10 and 30%.

10% clock for clock, in his summary. Maybe you are thinking of a different Anand?

 

[edplayer]

Your going to have to prove the 10% increase is real . Because every app needs to be run not selected ones. Than AVX is a factor that is discounted, Your not including the IGP improvements your flat out picking numbers to fit your world . The consumer who will use these and companies that use these are going to see a much larger improvement than what your willing to admitt to . You have no defense here because the SB chip tested has IGP ondie. Like it or not its part of the INTEL improvement . Your GPU standards don't change the facts at all. The consumer that uses SB and its IGP just got a huge performance boost . Bigger than C2D, If ya view the market as a whole. Intels 1st qt results will be stunning. Get over yourselves.

What is wrong with you man?

I don't have to prove anything. I'm not making any claim of how fast SB is or is not. I wrote that Anand made a guess. Look at it again.

 

[IntelUser2000]

10% clock for clock, in his summary. Maybe you are thinking of a different Anand?

About the 3rd time now:

http://www.computerbase.de/artikel/p...lierungsrating

2500: 15% faster per clock in average
2600: 17%

 

[magomago]

I have been going out of my way NOT to hype the product. The reality is that until you have shipping production silicon in your hand, any benchmark data is just speculation.

I would think as enthusiasts this crowd would appreciate people not coming in from the vendors to hype their product as the best thing since sliced bread.

And I continue to be amazed at the folks (on both sides) that want to see the other guy fail. Can't we all just agree that if you want the best products at the lowest prices in the market, you need good strong compeition?

No, I agree with you, and in BD's case, we don't hvae any benchmark data. You shouldn't be hyping it - very rarely does any hype match whatever is being pushed, and even a strong showing would be a failure in people's eyes if things were hyped too much PRIOR to any real performance information being released.

I was just saying, "Why do you guys think BD is over hyped, when the reality it really isn't! Hype, from my POV, needs at least some real details or performance numbers, and we know next to nothing about BD performance. Just sit down and wait for the day that they say something....or buy a sandybridge Even if you disagree with me on what hype is, then surely don't tell me that BD is hyped because the vast majority of people here are expecting it to be underwhelming or a lukewarm attempt by AMD to come close to SB performance"

 

[edplayer]

About the 3rd time now:

Maybe you are not understanding what is going on here. I'm not saying how fast Sandy Bridge is, I'm repeating what Anand said back in his August Sandy Bridge preview.

 

[IntelUser2000]

Maybe you are not understanding what is going on here. I'm not saying how fast Sandy Bridge is, I'm repeating what Anand said back in his August Sandy Bridge preview.

I see, I apologize. Still, that 10% is being thrown around alot.

 

[edplayer]

No problem. Thats one reason I like Anand's CPU bench comparisons. You can directly compare two specific cpus. Its not perfect, but its a lot better than using only one number.

 

[ydnas7]

P.... The consumer that uses SB and its IGP just got a huge performance boost . Bigger than C2D, If ya view the market as a whole. Intels 1st qt results will be stunning. Get over yourselves.

Nemesis is basically correct here, particularly for the mobile market. Same effect will occur with LLano also. But for the general consumer market, SB is a huge performance boost for same $$$, Intel's lowest IGP now (temporarily until LLano etc) matches ATI's highest IGP...

I don't know where LLano will be priced, but I expect LLano (with all cores/shaders active but not BE) will be between the core i3-2100 and the core i5-2300
this too will be a big boost for the general public, don't expect to be repeated until we get on package memory...

 

[magomago]

Yeah ydnas, you may be right about IGP, but Intel's tock hasn't been matched by AMD yet. AMD's 'tock' is coming soon. Of course it may only be equal to what Intel does, or it may go further. We'll see.

I gotta say I agree with Anand's assessment - we'll see bigger benefits once we see these integrated GPUs matching performance of 70 dollar cards.
Will definitely be interesting for GPU companies - no more bottom feeder cards in a segment that sells a lot. I'd be curious to see breakdown of card price vs % of revenue for each company. Where is the most money made?

 

[JFAMD]

Now wait a minute. We had a topic here were you argued AMD said The year2011 is this not true? Was there a follow up topic I missed?

I sure hope AMd is earlier but if there late 2011 than this smells bad. I sure hope AMD doesn't rush the release and goof up again. Damage is already done . Do it right AMD do it right.

Yes, I was saying that prior to November, but in November we published this:

http://blogs.amd.com/work/2010/11/09/server-highlights-from-financial-analyst-day/

And since that time I have been saying Q2 for desktop, Q3 for server.

 

[VirtualLarry]

I don't necessarily think BD will be faster than SB, core for core. But it seems like ATI's "small die strategy" applied to CPUs. More cores for the same silicon real-estate, should increase price/performance ratio, even if it doesn't take the overall performance crown. Thus I think that BD will be successful.

Edit: PS. I think that we need more heavily multi-threaded benchmarks to give BD a fair comparison. Benching it with games that only take two cores, is going to show bad results.

 

[Elixer]

I don't necessarily think BD will be faster than SB, core for core. But it seems like ATI's "small die strategy" applied to CPUs. More cores for the same silicon real-estate, should increase price/performance ratio, even if it doesn't take the overall performance crown. Thus I think that BD will be successful.

Edit: PS. I think that we need more heavily multi-threaded benchmarks to give BD a fair comparison. Benching it with games that only take two cores, is going to show bad results.

That is going to require a paradigm shift in coding, there are so many things with the current way we think of doing things, that can't be done in parallel. Sure, it is great for the things that can be done that way, but, unless you are encoding or doing server stuff, then most core/threads will go unused.

For the Q2 release, at this time only MSI's bulldozer board has been spotted at CES, here is a pic http://techreport.com/discussions.x/20227 and that means that the boards are being produced. I am unsure why the other makers didn't show their bulldozer boards. Kinda reminds me of the days when everyone was afraid of intel, and they didn't want it known they made AMD boards as well...

Though, this does have me wondering, does Dirk's being canned, erm, resignation, from AMD, play a role in the fact that there have been no PR events to speak of?
Q2 is anywhere from April to June, that is a heck of a long time between now and then for AMD to be keeping quiet.--though, that still don't explain why the other board partners are quiet as well.

 

[grimpr]

First cpuz shots and clocks around, fake or not.

 

[Dadofamunky]

And I continue to be amazed at the folks (on both sides) that want to see the other guy fail. Can't we all just agree that if you want the best products at the lowest prices in the market, you need good strong compeition?

+1

 

[JFAMD]

That is going to require a paradigm shift in coding, there are so many things with the current way we think of doing things, that can't be done in parallel. Sure, it is great for the things that can be done that way, but, unless you are encoding or doing server stuff, then most core/threads will go unused.

For the Q2 release, at this time only MSI's bulldozer board has been spotted at CES, here is a pic http://techreport.com/discussions.x/20227 and that means that the boards are being produced. I am unsure why the other makers didn't show their bulldozer boards. Kinda reminds me of the days when everyone was afraid of intel, and they didn't want it known they made AMD boards as well...

Though, this does have me wondering, does Dirk's being canned, erm, resignation, from AMD, play a role in the fact that there have been no PR events to speak of?
Q2 is anywhere from April to June, that is a heck of a long time between now and then for AMD to be keeping quiet.--though, that still don't explain why the other board partners are quiet as well.

First, the paradigm shift in coding that you speak of started ~2003 when people found out that they weren't getting any more clock speed and that threads matter. This continues to gain more momentum with every released software program. You are right that people need to change their behaviors, but that started years ago.

Second, we are in the middle of quiet period so you wouldn't see any comments from AMD around bulldozer. We had to jump through huge loops to launch the APUs at CES. We have released a LOT of information about bulldozer to date, far more than any other product that I have been involved in. The only things that wait for launch are benchmarks, pricing, clock speed and SKUs. But that is the same with every launch of ours as well as our competitor.

 

[poohbear]

even if you are getting SB, waiting for BD might at least bring the price down if nothing else. but in the event that BD is great, now your patience can net you a better chip.

totally agree, that's the most sound strategy at this point for those of us in no rush to upgrade.

 

[Phynaz]

AMD according to this is going to revive the FX label for its highest performing chips. If thats true, no doubt they won't slow.
http://www.xbitlabs.com/news/cpu/di...mance_of_AMD_s_Bulldozer_Microprocessors.html

Why do you say that? I'm not of the belief that marketing labels say anything about cpu performance.

 

[Phynaz]

Second, we are in the middle of quiet period so you wouldn't see any comments from AMD around bulldozer. We had to jump through huge loops to launch the APUs at CES.

I love it. Companies are still using this as a "we have nothing good to say" cover.

Unless AMD is doing a secondary offering in the next thirty days there is no mandated quiet period.

http://www.sec.gov/answers/quiet.htm

 

[Castiel]

First cpuz shots and clocks around, fake or not.

Worst Leak Ever

 

[smakme7757]

even if you are getting SB, waiting for BD might at least bring the price down if nothing else. but in the event that BD is great, now your patience can net you a better chip.

I sat on a Q9450 for 3 years with DDR2 ram and on the 9th of January 2011 i went to a 2600K. Massive increase in performance so my wait was well worth it in my oppinion.

If your sitting on a 1366/1156 then waiting until the end of this year would be your best bet. Those CPUs are still awesomely powerful, no need to upgrade to SB.

 

[exar333]

I don't necessarily think BD will be faster than SB, core for core. But it seems like ATI's "small die strategy" applied to CPUs. More cores for the same silicon real-estate, should increase price/performance ratio, even if it doesn't take the overall performance crown. Thus I think that BD will be successful.

Edit: PS. I think that we need more heavily multi-threaded benchmarks to give BD a fair comparison. Benching it with games that only take two cores, is going to show bad results.

On the CPU side, AMD is essentially completely opposite this methodology. They are placing a 8-core (4 module) BD against a 4 core SB (8 thread) CPU. Which one will be smaller? I am pretty sure it is not the BD core with double the cores....

 

[JFAMD]

I love it. Companies are still using this as a "we have nothing good to say" cover.

Unless AMD is doing a secondary offering in the next thirty days there is no mandated quiet period.

http://www.sec.gov/answers/quiet.htm

I never said it was an SEC requirement, but it is a company requirement. We launched Brazos at CES so obviously there is latitude, the question is how far up the legal chain do you want to push. There is extra scrutiny in the quiet period.

 

[JFAMD]

On the CPU side, AMD is essentially completely opposite this methodology. They are placing a 8-core (4 module) BD against a 4 core SB (8 thread) CPU. Which one will be smaller? I am pretty sure it is not the BD core with double the cores....

Hmmmmmm, what's the SB die size? We have publicly said that we will be smaller than Istanbul.

 

[Elixer]

First, the paradigm shift in coding that you speak of started ~2003 when people found out that they weren't getting any more clock speed and that threads matter. This continues to gain more momentum with every released software program. You are right that people need to change their behaviors, but that started years ago.

Second, we are in the middle of quiet period so you wouldn't see any comments from AMD around bulldozer. We had to jump through huge loops to launch the APUs at CES. We have released a LOT of information about bulldozer to date, far more than any other product that I have been involved in. The only things that wait for launch are benchmarks, pricing, clock speed and SKUs. But that is the same with every launch of ours as well as our competitor.

Not sure I agree about 2003 timeframe, but looking at all the applications released today, we have a long way to go before parallel programming becomes very useful in the non-server, non-encoding environment.

I am not sure I follow about the self imposed quiet period, if you have a product that will be a game changer, then it is in the companies own interest to start doing some marketing telling people 'don't abandon us yet, we have a great product coming along', instead of being quiet which is putting out signals that the product isn't as good as one would have hoped.

Think of the original phenom release, expectations were very high, and the people who actually waited for that were left with a more or less 'meh' feeling when it finally did arrive.
It didn't generate more sales, it just pushed those people to the intel camp.

I am thinking of doing something along the lines of how intel was doing skulltrail vids on youtube, AMD could be doing the same thing with BD.

 

[deimos3428]

Worst Leak Ever

3.835 Ghz=225x17, which would put the stock clock at 3.4Ghz -- if not a fake.