AMD's Andrew Feldman admits that Bulldozer was a failure

[MadDias]

http://www.pcworld.com/article/20400...ver-chips.html
I thought this was news, but since the fukushima catastrophe I'm not allowed to post in the news section anymore.
maybe someone else can do that for me.

Bulldozer was without doubt an unmitigated failure. We know it,? Feldman said.

?It cost the CEO his job, it cost most of the management team its job, it cost the vice president of engineering his job. You have a new team. We are crystal clear that that sort of failure is unacceptable,? Feldman said.

lets see what AMD has learned from this failure.

[AMDforME]

Bulldozer was a failure because it was four years late. Had it been delivered when it was suppose, it would have been just fine. IMO, GloFo execs are the ones who should have all been fired.

FWIW, AMD has already resolved the shortcomings of BD and consumers will see that with Steamroller.

[MadDias]

Bulldozer was a failure because it was four years late. Had it been delivered when it was suppose, it would have been just fine. IMO, GloFo execs are the ones who should have all been fired.

FWIW, AMD has already resolved the shortcomings of BD and consumers will see that with Steamroller.

a cpu with 90% of the ipc of it?s predecessor is what u call just fine?
the word "unmitigated" worries me a little and make?s me doubt that steamroller will resolve all the shortcomings in bulldozer architecture.

[FlanK3r]

BD had few good ideas... Maybe in future...Look example at Vishera, clock to clock is more better and still is it Bulldozer uniqarchitecture. We will see, what brings Steamroller and Excavator (bioth are based at BD )

[haylui]

a cpu with 90% of the ipc of it?s predecessor is what u call just fine?
the word "unmitigated" worries me a little and make?s me doubt that steamroller will resolve all the shortcomings in bulldozer architecture.

IPC has to be coupled with clock speed.
The idea of BD is to have high clock speed but the foundry failed to delivery high speed chips for AMD.

[[XC] Oj101]

a cpu with 90% of the ipc of it?s predecessor is what u call just fine?
the word "unmitigated" worries me a little and make?s me doubt that steamroller will resolve all the shortcomings in bulldozer architecture.

Had it been four years earlier it would have replaced the original Phenom and been a much bigger success

[FlanK3r]

hehe... But now with Vishera, Im really happy. Performance in practice is very good, in many real apliactaion leave behind i7-2600k. OC ability is very huge with good board (or Im lucky for Vishera). My first 5 GHz PC station . Im happy for reasonable power consumption after OC. Bulldozer was in this horror story.

We will see what coming next from AMD...

[imamage]

http://www.pcworld.com/article/20400...ver-chips.html
I thought this was news, but since the fukushima catastrophe I'm not allowed to post in the news section anymore.
maybe someone else can do that for me.
lets see what AMD has learned from this failure.

Well , they bluff it too much hype with Bulldozer and the end product really disappointed fans
I was so upset that I almost switch to Intel (Sandy Bridge) for my upgrade , but in the end
I enjoy the "6" thread (multi-tasking) gain by my trusty 1090T, I grab a FX-6300 as Piledriver wasn't that bad.

[imamage]

I don't know but do GF have yield / production issue with 32nm also harm AMD ?

[AbortRetryFail?]

The first released stepping after a die-shrink, especially at the high-end, has always been a struggle for AMD.

I'm not really sure why folks would think otherwise. We're just AMD beta-testers

The only true success I can think of was the T-Bird with minimal architectural changes. I believe Deneb to Bulldozer can be viewed successful simply because it subsequently brought forth Vishera.

The good news is ... the shrink from Vishera to Steamroller involves minimal changes to the arch and the cores design will essentially be first marketed via the Kaveri APU.

I suspect that first APU stepping at 28nm may be the one on which to hold our collective breaths. If it goes (relatively) smoothly for AMD ....

Steamroller should be just dandy.

[vario]

Well, these kind of folks never admit that somethings bad, except when their upcoming thing is much better.Grasping at straws here but from everything i hear and read looks like steamroller is gonna be at least OK.Hope their gonna stick with am3+ for steamroller FX .

[rintamarotta]

The first released stepping after a die-shrink, especially at the high-end, has always been a struggle for AMD.
I'm not really sure why folks would think otherwise. We're just AMD beta-testers
The only true success I can think of was the T-Bird with minimal architectural changes. I believe Deneb to Bulldozer can be viewed successful simply because it subsequently brought forth Vishera.
The good news is ... the shrink from Vishera to Steamroller involves minimal changes to the arch and the cores design will essentially be first marketed via the Kaveri APU.
I suspect that first APU stepping at 28nm may be the one on which to hold our collective breaths. If it goes (relatively) smoothly for AMD ....
Steamroller should be just dandy.

Kaveri Apu = Finnish and means, Friend Assistance

Hopefully AMD gets back into it (performance) within next 2 years, i want to switch from I-camp to A-camp but A-camp performance is keaping me from doing it (alot of H264 encoding and A-camp is slow in it).
High-end performance havent been allways struggle for AMD, remember good old AMD Athlon and Duron days when AMD whooped Intel in allmost every single task?

I hope 28nm shrink goes well for AMD, so in future (within 2years) we might see AMD back in equal performance versus intels top products.

Now, lets wait to see 5Ghz AMD FX chip that can be overclocked to 6Ghz with air, Piledriver and Bulldozer is made for high clocks but GF just cannot keap up with production quality requirements for those clocks.

[FlanK3r]

i think, there is not much OC ability for new FX...Maybe 200 MHz up, not more (so 5200 MHz stable maybe with good cooling)

[stangracin3]

If that 5 Ghz chip is true, I would lie to see what they include with it as a cooler.

[64NOMIS]

Long live the righteous.

[FlanK3r]

If that 5 Ghz chip is true, I would lie to see what they include with it as a cooler.

I think AMD wattercooling setup (rebrand Antec 920 Kuhler). I think, it can be true...My FX-8350 can rendering at 5300 MHz and I have not special cooling, only H100 Corsair. At 4900 Mhz I can benchmarking all with 1.4V only . Yes, I have better chip than average 8350s, but if this monsters will be binned, why not?

[mockingbird]

Bulldozer was a failure because it was four years late. Had it been delivered when it was suppose, it would have been just fine. IMO, GloFo execs are the ones who should have all been fired.

FWIW, AMD has already resolved the shortcomings of BD and consumers will see that with Steamroller.

No. It failed because Intel was able to play the underdog to AMD after Pentium 4. While Intel was eliminating the bloat in their corporate structure and outsourcing production to a very successful Jewish team in Israel, AMD was adding bloat to their own infrastructure and ignored market trends and consumer demands while it was pre-occupied with delusional notions that it would inevitably overtake Intel and add to its empire indiscriminately and with impunity.

During this time, Intel sheared its fat, and delivered great products to the market at great prices. Intel went unchallenged in the Laptop market as well as the medium-range desktop market, and by the time AMD finally put out Kuma to compete with Core2, it was far, far too late.

Now, AMD is impractical for even the ultra-low end market, as high-quality, inexpensive Intel platforms for office environments cost less than equivalent FM2 platforms. The rationale for purchasing AMD chips for better graphics performance for the low-end has now disappeared since Intel chips, while not posessing as impressive GPU cores as AMD chips, are still very powerful and more than sufficient for your average office user.

All AMD has left is the PS4 and Xbox One contracts. With the emerging console markets and upcoming competition from new console platforms, this is not enough to sustain them.

AM's experimentation with a multiculturalist corporate environment and Euro-socialist ideas of how a company should function is what lead to their demise. Serves them right.

[MadDias]

If that 5 Ghz chip is true, I would lie to see what they include with it as a cooler.

seems to be true: http://www.xtremesystems.org/forums/...=1#post5191503

what cooling they pack with that beast? maybe one of those aio watercooling gimmicks, which i?m using on a i5 right now. lousy thing...

[Darakian]

I think pretty much anyone on AM3+ is going to stay AM3+ so long as there is a simple upgrade path that performs decently. If steamroller is AM3+ then great, I'd love to jump from my old phenom 2 onto something more modern. If not, I'll be looking at haswell whenever my phenom becomes a real hindrance.

[stangracin3]

I'm not sure they will put an AIO in for a 5ghz chip. unless it was one of the 240's because i doubt a 120 could handle it

[leo5111]

that chip sounds like Bull Rumor that tdp would be mandatory water, GOOD WATER

[FlanK3r]

I told it...My FX-8350 5 GHZ has 240W at CPU in hard load. And I have not good watter....Only H100 Corsair.

[Mechanical Man]

Please flanker. It is water. Only one t not twice like in wasser. It just always gots into my eye. Thank you for you remembering that in future

[FlanK3r]

hehe, thx.

[Tomasis]

No. It failed because Intel was able to play the underdog to AMD after Pentium 4. While Intel was eliminating the bloat in their corporate structure and outsourcing production to a very successful Jewish team in Israel, AMD was adding bloat to their own infrastructure and ignored market trends and consumer demands while it was pre-occupied with delusional notions that it would inevitably overtake Intel and add to its empire indiscriminately and with impunity.

During this time, Intel sheared its fat, and delivered great products to the market at great prices. Intel went unchallenged in the Laptop market as well as the medium-range desktop market, and by the time AMD finally put out Kuma to compete with Core2, it was far, far too late.

Now, AMD is impractical for even the ultra-low end market, as high-quality, inexpensive Intel platforms for office environments cost less than equivalent FM2 platforms. The rationale for purchasing AMD chips for better graphics performance for the low-end has now disappeared since Intel chips, while not posessing as impressive GPU cores as AMD chips, are still very powerful and more than sufficient for your average office user.

All AMD has left is the PS4 and Xbox One contracts. With the emerging console markets and upcoming competition from new console platforms, this is not enough to sustain them.

AM's experimentation with a multiculturalist corporate environment and Euro-socialist ideas of how a company should function is what lead to their demise. Serves them right.

so many exaggerations in your comment.

In history, Intel has been seen as "single player" in most people's eyes.

Btw look at Kaveri, Jaguar in ultra low power section. I think you are lagging behind current market times.

The last sentence made me chuckle!