Why is the response to Bulldozer so overwhelmingly negative?

[LOL_Wut_Axel]

People got their hopes set way high. With all the hype, nothing short of a miracle that would blow Intel out of the water would have been good enough.

Disappointment seems to have taken overhand a bit, and it does get blown out of proportions. I think it'll be interesting to see how the BD architechture fares on the longer run.

Clearly expecting for it to be overall better than the Phenom II after it had been two years in the market and was based on an eight-year-old architecture was asking for too much.

 

[woolfe9999]

The "high expectations" isn't the most relevant thing here. AMD could have said literally nothing about this chip other than that they were going to release it. They could have left the public to form their own expectations as a matter of pure conjecture. And this chip would still not have satisfied any of those expectations, varied as they would have been. Sure, the high expectations didn't help, but something is wrong with this architecture. This was an engineering gaff more than a marketing gaff.

 

[Obsoleet]

Because people are shortsighted. Only a CPU enthusiast, one who uses Extreme Series CPUs, has a legitimate complaint against the FX series.

CPUs have been overpowered for 90% of us since the C2D launch, and the Phenom II was in that ballpark, leaves the FX series in good company. The power use is a little high, and the price is a little high as well. But overall it's equal to the Phenom II stuff, and over time since the design is still more advanced and has longer legs, it'll be a better chip than sticking with Phenom.

 

[Farmer]

If you are interested in massively parallel computation in a research environment, and you are interested in a specialized solution: why not a Tesla/Fermi cluster?

What type of work is this? Do you use a well known code or your own homebrew? The research I do involves classical atomistic simulation, amongst other things. The type of performance scaling improvement we would see from using Tesla blows CPU-focused computation out of the water, far outshadowing whatever performance increase you would obtain from switching from, say, 24 cores of a particular architecture to 24 cores of a slightly better architecture.

Obviously, no home user in their right minds would purchase a Tesla. However, a research-oriented number cruncher will cost you $10K average, and going Tesla is not totally out of the question.

 

[Arg Clin]

Clearly expecting for it to be overall better than the Phenom II after it had been two years in the market and was based on an eight-year-old architecture was asking for too much.

Asking for something that doesn't make sense. Asking for it to surpass Phenom II in /lightsinglethreaded applications when it was clearly designed with multithreading in mind.

AMD has a long history of thinking ahead and not get rewarded for it. Reminds me of the K6-2 with 3Dnow!/SIMD debate.

If history repeats itself we'll likely see a Bulldozer-like design from Intel in the future. (focus on more cores+less performance pr core) But they won't do it untill the market is ripe for it. Feet-dragging is rewarded again because it gives the users a quick fix for their old and tired lightly threaded/single threaded apps.

Does that mean that Bulldozer is the best buy right now? I many cases no - and ultimately that is the plight of AMD once again. Consumers do not look that much ahead and great new technology/fresh thinking isn't really rewarded untill the apps support it. And that might takes years, develops doesn't really bother with it untill Intel backs it, because of Intel volume.

Again - this is not a Bulldozer sales pitch. Bulldozer wouldn't be the chip of my choice is many situations. But the bashing of AMD for having a longer perspective than singlethreaded performance in yesteryears apps is uncalled for IMO.

 

[StrangerGuy]

Asking for something that doesn't make sense. Asking for it to surpass Phenom II in /lightsinglethreaded applications when it was clearly designed with multithreading in mind.

AMD has a long history of thinking ahead and not get rewarded for it. Reminds me of the K6-2 with 3Dnow!/SIMD debate.

If history repeats itself we'll likely see a Bulldozer-like design from Intel in the future. (focus on more cores+less performance pr core) But they won't do it untill the market is ripe for it. Feet-dragging is rewarded again because it gives the users a quick fix for their old and tired lightly threaded/single threaded apps.

Does that mean that Bulldozer is the best buy right now? I many cases no - and ultimately that is the plight of AMD once again. Consumers do not look that much ahead and great new technology/fresh thinking isn't really rewarded untill the apps support it. And that might takes years, develops doesn't really bother with it untill Intel backs it, because of Intel volume.

Again - this is not a Bulldozer sales pitch. Bulldozer wouldn't be the chip of my choice is many situations. But the bashing of AMD for having a longer perspective than singlethreaded performance in yesteryears apps is uncalled for IMO.

They should have seen it coming, all i can say is: suck it up. Blaming consumers for not being future-minded is not going to work in principle and practice.

 

[mosox]

Because people are shortsighted. Only a CPU enthusiast, one who uses Extreme Series CPUs, has a legitimate complaint against the FX series.

CPUs have been overpowered for 90% of us since the C2D launch, and the Phenom II was in that ballpark, leaves the FX series in good company. The power use is a little high, and the price is a little high as well. But overall it's equal to the Phenom II stuff, and over time since the design is still more advanced and has longer legs, it'll be a better chip than sticking with Phenom.

Yep. The C2Q, i5 750/760, i7 950, Phenom II X4, Phenom II X6, i5 2300-2500, Bulldozer, are all adequate for the regular user or a regular gamer. Even lesser CPUs (i3, Athlon II X3/X4) are good enough for non enthusiasts.

Modern fighter planes (F35, Rafale, Eurofighter, etc) are slower than the 40-50 y.o. fighters like the F104, Mirage III, F4 Phantom. But those modern planes are much better in multitasking.

The problem is the software. The software developers want a big market and since currently there are 1+ billion machines in use, many of them obsolete, they will dumb down the software in order to be able to work on old Pentium CPUs and the likes. And this means single threaded performance.

That's the problem, you may have a very good car and nobody to make the fuel that's the best for it, only some generic fuel that works on all the cars, even in the Ford Model T that was last produced in 1927.

 

[TakeNoPrisoners]

Because people are shortsighted. Only a CPU enthusiast, one who uses Extreme Series CPUs, has a legitimate complaint against the FX series.

CPUs have been overpowered for 90% of us since the C2D launch, and the Phenom II was in that ballpark, leaves the FX series in good company. The power use is a little high, and the price is a little high as well. But overall it's equal to the Phenom II stuff, and over time since the design is still more advanced and has longer legs, it'll be a better chip than sticking with Phenom.

"Only"

 

[exar333]

Because people are shortsighted. Only a CPU enthusiast, one who uses Extreme Series CPUs, has a legitimate complaint against the FX series.

CPUs have been overpowered for 90% of us since the C2D launch, and the Phenom II was in that ballpark, leaves the FX series in good company. The power use is a little high, and the price is a little high as well. But overall it's equal to the Phenom II stuff, and over time since the design is still more advanced and has longer legs, it'll be a better chip than sticking with Phenom.

I call shens on this. How do you know this?

 

[exar333]

Yep. The C2Q, i5 750/760, i7 950, Phenom II X4, Phenom II X6, i5 2300-2500, Bulldozer, are all adequate for the regular user or a regular gamer. Even lesser CPUs (i3, Athlon II X3/X4) are good enough for non enthusiasts.

Modern fighter planes (F35, Rafale, Eurofighter, etc) are slower than the 40-50 y.o. fighters like the F104, Mirage III, F4 Phantom. But those modern planes are much better in multitasking.

The problem is the software. The software developers want a big market and since currently there are 1+ billion machines in use, many of them obsolete, they will dumb down the software in order to be able to work on old Pentium CPUs and the likes. And this means single threaded performance.

That's the problem, you may have a very good car and nobody to make the fuel that's the best for it, only some generic fuel that works on all the cars, even in the Ford Model T that was last produced in 1927.

You are right, but it's AMDs responsibility to make something that works with existing software AND future software. Constraints are in every industry, and those who ignore the market are doomed to fail, as 'good' as the product might be if used in the ideal application.

 

[evident]

at this point, anyone who can't admit that bulldozer is pure trash and disappointment is an amd fanboy. the benchmarks and numbers are there. there's no other way to spin it. who cares about multicores when the single thread performance is so terrible. I've just been reading the cpu forum briefly for the last few months because i was looking to upgrade and i'm shocked that people have these crazy brand loyalties.

 

[Arg Clin]

They should have seen it coming, all i can say is: suck it up.

What is this.. a football game?

Blaming consumers for not being future-minded is not going to work in principle and practice.

Like I said; "this is not a Bulldozer sales pitch. Bulldozer wouldn't be the chip of my choice is many situations. But the bashing of AMD for having a longer perspective than singlethreaded performance in yesteryears apps is uncalled for IMO." Consumers should buy exactly what chip they feel suits their needs the best. "Consumers do not look that much ahead and great new technology/fresh thinking isn't really rewarded until the apps support it. And that might takes years, develops doesn't really bother with it until Intel backs it, because of Intel volume." From a consumer point of view it makes perfect sense that AMD, despite being highly innovative is behind. 3Dnow! never took off, but SSE did.. why.. because Intel backed it. Fair? - not by a long shot - but that's how it works. (we'll take the debate on whether or not the anti-competition authorities are doing a good job some other time)

I was NOT in any way taking a stab at consumers for being short sighted for not praising Bulldozer. I was taking a stab at those who is bashing AMD and judging Bulldozer as failure just because it doesn't give an instant performance fix in yesterdays singlethreaded applications.

 

[LOL_Wut_Axel]

Asking for something that doesn't make sense. Asking for it to surpass Phenom II in /lightsinglethreaded applications when it was clearly designed with multithreading in mind.

AMD has a long history of thinking ahead and not get rewarded for it. Reminds me of the K6-2 with 3Dnow!/SIMD debate.

If history repeats itself we'll likely see a Bulldozer-like design from Intel in the future. (focus on more cores+less performance pr core) But they won't do it untill the market is ripe for it. Feet-dragging is rewarded again because it gives the users a quick fix for their old and tired lightly threaded/single threaded apps.

Does that mean that Bulldozer is the best buy right now? I many cases no - and ultimately that is the plight of AMD once again. Consumers do not look that much ahead and great new technology/fresh thinking isn't really rewarded untill the apps support it. And that might takes years, develops doesn't really bother with it untill Intel backs it, because of Intel volume.

Again - this is not a Bulldozer sales pitch. Bulldozer wouldn't be the chip of my choice is many situations. But the bashing of AMD for having a longer perspective than singlethreaded performance in yesteryears apps is uncalled for IMO.

Even if you're looking at JUST multi-threaded applications it's barely better than the i5-2500K.

Well, would you look at that. 1% faster than the i5-2500K overall in important multi-threaded benchmarks. WOOT!

Here's a look at performance/watt in multi-threaded workloads:

The i5-2500K has 2x higher performance/watt even in multi-threaded.

Now let's look at overall performance/watt:

2.2x higher performance/watt. Simply unacceptable.

Bulldozer is a huge failure, and anyone that says otherwise is a moron. It's overall not even better than the i5-2500K in multi-threaded benchmarks and costs more.

And going by AMD estimates future iterations will be turds, too. AMD is predicting a 10-15% performance and power consumption improvement each year. That won't get them far as Ivy Bridge will overall be a ~10% performance improvement (5% from IPC, 5% from slightly higher clock speeds) and a 25% or higher improvement in power consumption.

 

[Ferzerp]

No, you don't get it. It will magically become better over time! (that's the current claim, right?)

 

[Arg Clin]

Even if you're looking at JUST multi-threaded applications it's barely better than the i5-2500K.

2.2x higher performance/watt. Simply unacceptable.

Bulldozer is a huge failure, and anyone that says otherwise is a moron. It's overall not even better than the i5-2500K in multi-threaded benchmarks and costs more.

And going by AMD estimates future iterations will be turds, too. AMD is predicting a 10-15% performance and power consumption improvement each year. That won't get them far as Ivy Bridge will overall be a ~10% performance improvement (5% from IPC, 5% from slightly higher clock speeds) and a 25% or higher improvement in power consumption.

We all know the benchgraphs, so no need to cherry pick further I'm sure you are well aware that BD is built to ramp clockspeeds quite well and that some software favours a BD style like architecture.

Not the first time someone has been called 'a moron' for having a balanced view vs. the mainstream one-liners ala "it suxors!!!" - won't be the last either.

But do carry on and tell us from the magic crystal ball what AMD will be doing with the BD architechture in further iterations

 

[nyker96]

I have also heard a recent news that AMD lowered trinity core improvement expectations compare to Llano from 50% down to about 20%, since Trinity is based on Piledriver which is the next iteration of BD, this isn't good news at all. AMD seems to be hitting rough patch for its CPU lines. Maybe they should just cut their losses and drop BD, just improve on K10 more and die shrink it.

 

[Arg Clin]

No, you don't get it. It will magically become better over time! (that's the current claim, right?)

Absolutely. Obviously since CPU's are made by Gandalf. Engineers tweaking their designs are a myth.

 

[LOL_Wut_Axel]

We all know the benchgraphs, so no need to cherry pick further I'm sure you are well aware that BD is built to ramp clockspeeds quite well and that some software favours a BD style like architecture.

Not the first time someone has been called 'a moron' for having a balanced view vs. the mainstream one-liners ala "it suxors!!!" - won't be the last either.

But do carry on and tell us from the magic crystal ball what AMD will be doing with the BD architechture in further iterations

Yes, because clearly thinking Bulldozer is good is being "balanced". I'm being objective; I overall like AMD's GPUs more than NVIDIA's.

Nothing is being cherry picked. Those are eight multi-threaded REAL-WORLD programs: Blender, HandBrake, MainConcept, After Effects, Photoshop (MT filter), Adobe Premiere, Matlab, 7-Zip. 8 applications, and it ended up being 1% faster overall while costing significantly more and having significantly worse performance/watt.

Bulldozer ramps up clock speeds no better than Sandy Bridge. As a matter of fact, its high power consumption limits it.

And it's not me that's saying what AMD will be doing with Bulldozer--it's AMD themselves. I'm just stating what they've told us, and what Intel has told us. Going by those facts AMD is in a worse position, as Ivy Bridge will be more of an improvement overall than Piledriver will.

 

[Edrick]

I was taking a stab at those who is bashing AMD and judging Bulldozer as failure just because it doesn't give an instant performance fix in yesterdays singlethreaded applications.

Yea, but it doesnt provide much of an "instant performance fix" in todays multithreaded applications either. I think that is the main issue.

 

[hooflung]

at this point, anyone who can't admit that bulldozer is pure trash and disappointment is an amd fanboy. the benchmarks and numbers are there. there's no other way to spin it. who cares about multicores when the single thread performance is so terrible. I've just been reading the cpu forum briefly for the last few months because i was looking to upgrade and i'm shocked that people have these crazy brand loyalties.

Because, like you, people are entitled to their opinion. And there are facts that say otherwise that this is not a trash CPU. You are looking for a reason to justify your rage. Others are trying to find the silver lining. And they have it on the workstation and server spaces despite anything you try to cram down their throat.

 

[Medu]

I have also heard a recent news that AMD lowered trinity core improvement expectations compare to Llano from 50% down to about 20%, since Trinity is based on Piledriver which is the next iteration of BD, this isn't good news at all. AMD seems to be hitting rough patch for its CPU lines. Maybe they should just cut their losses and drop BD, just improve on K10 more and die shrink it.

The problem with that is that K10.5 isn't very good either. AMD have had to sell bigger CPU's than Intel for far less and these has been true for over 5 years now. The Phenom II is a fairly poor chip but it had 1 redeeming quality- performance per $. Performance per mm²(important to AMD's bottom line) is poor, as is PPW which a lot of people stupidly overlook. So people spend 10% less on an AMD machine but probably pay far higher energy bills(less of a problem in the US where energy is far cheaper than most of the world) and end up with a machine that is 30-40% slower than the 10% more expensive Intel machine.

So to answer the original question: BD has been received so poorly because AMD has positioned it poorly. If it was been sold at $150 then it would of been received better- even though it still would of been a dreadful CPU(just like the 2900XT rubbish but ATi/AMD had the sense not to put it in the same price bracket as the 8800GTX.)

Because, like you, people are entitled to their opinion. And there are facts that say otherwise that this is not a trash CPU. You are looking for a reason to justify your rage. Others are trying to find the silver lining. And they have it on the workstation and server spaces despite anything you try to cram down their throat.

Such as? Benchmarks that show a CPU with a die size of ~320mm2 marginally outpace a CPU with on-chip graphics and a die space of ~220mm2 isn't a good sign. Bottom line is that it's a VERY VERY poorly executed CPU that should never have gotten the green light. It's concept was flawed, the software isn't ready for it and AMD didn't have the resources to make it work....

 

[beginner99]

People forget that a lot of task just can't be multi threaded at all. You will always need good single-threaded performance. IMHO in a certain way ARM is showing the way to go. Use high-power cores only when you need them. Anything that can be parallelized, run it on a ton of "slow cores", eg. GPU. Once OpenCL, CUDA etc. mature, and more software makes use of it, many core CPU's will get useless. 4 high power cores will probably still be enough.

 

[Chiropteran]

People forget that a lot of task just can't be multi threaded at all.

Like what? What tasks can't be multi-threaded and can't be handled by a 1.6ghz e-350?

 

[lol123]

Modern fighter planes (F35, Rafale, Eurofighter, etc) are slower than the 40-50 y.o. fighters like the F104, Mirage III, F4 Phantom. But those modern planes are much better in multitasking.

That's a ridiculous analogy. The reason that modern fighter jets aren't faster than fighters were 40-50 years ago isn't that they are fast enough, it's that we can't make them any faster with respect to maneuverability and fuel consumption even with modern technology.

And we have the same situation in the computer industry. The reason that clock frequencies aren't increased more than they are isn't that further increases wouldn't be welcome, it is that we have hit a thermal wall where increasing clocks has become a lot harder to do than before. Jacking up frequencies and improving single-threaded IPC, for example by improvements to cache latency, branch prediction, prefetching etc., is still the best and most flexible way of increasing performance (witness the success of IBM's POWER7 and Intel's Sandy Bridge), it's just that at the present time it takes a great deal of resources to do so.

 

[Olikan]

That's a ridiculous analogy. The reason that modern fighter jets aren't faster than fighters were 40-50 years ago isn't that they are fast enough, it's that we can't make them any faster with respect to maneuverability and fuel consumption even with modern technology.
.

off topic:
jets aren't faster because the human body can't handle highter speeds.