Why is the response to Bulldozer so overwhelmingly negative?

[Ferzerp]

^ unimportance of frame of reference comprehension fail!

(hint: speed has 0 effect on the human body, you're hurtling through space compared to the sun at over 30KM/s right now, and 10 times that relative to the center of the galaxy. Don't feel the speed? That's because speed exerts no force!)

 

[Arg Clin]

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

Where did I say Bulldozer is "good"? I think it's good at some things and not very good at others. Balanced means that I don't think it's quite as simple as just saying that it's a bad chip and "suck it up". It has some merits, and the concept is interesting.

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.

Since when did performance/watt become king? Why are we not looking at absolute performance? Quite a lot of enthusiasts overclock - a sure killer of "performance/watt". Power consumption is one of the areas AMD (or maybe in fact GloFo) need to adress - so performance/watt charts will make BD look bad in particular. At least I'd be interested to hear why this particular aspect is so interesting that it makes or breaks a chip..? Again - quite unbalanced way to place the focus I think.

How about that FX8150 beats 2600K in Handbrake - a fairly cpu hungry application a lot of enthusiasts use. http://www.legitreviews.com/article/1751/6/

Cyberlink Media expresso 6.5 - another good example: http://www.legitreviews.com/article/1751/5/

...and then Bulldozer looses in others. But to claim it's a failure is out of proportions.

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

Source? Does Bulldozer not in fact hold the record in raw GHz
overclock..?

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.

Roadmaps are subject to change and should be taken with a huge grain of salt. AMD didn't stay in business for 40+ years by not being able to adjust to the current market situation.

 

[Hypertag]

...and then Bulldozer looses in others. But to claim it's a failure is out of proportions.


Source? Does Bulldozer not in fact hold the record in raw GHz
overclock..?

That was done with liquid helium (liquid nitrogen was literally too hot) and with only a single module on the most cherry picked chip AMD had access to.

As far as a source on power consumption on bulldozer limiting overclock ability, http://www.bit-tech.net/hardware/cpus/2011/10/12/amd-fx-8150-review/10 . I would love to see how much power usage bulldozer was pulling @8GHz.

 

[Arg Clin]

That was done with liquid helium (liquid nitrogen was literally too hot) and with only a single module on the most cherry picked chip AMD had access to.

Still a valid indicator that BD is capable of high clockspeeds. The claim that "Bulldozer ramps up clock speeds no better than Sandy Bridge" is unbacked by facts, since noone has clocked a SB to the same level.

As far as a source on power consumption on bulldozer limiting overclock ability, http://www.bit-tech.net/hardware/cpus/2011/10/12/amd-fx-8150-review/10 . I would love to see how much power usage bulldozer was pulling @8GHz.

Limiting overclock ability - perhaps. But that's not the same as saying it will not ramp clockspeeds well in future iterations.

 

[Ajay]

Since when did performance/watt become king? Why are we not looking at absolute performance? Quite a lot of enthusiasts overclock - a sure killer of "performance/watt". Power consumption is one of the areas AMD (or maybe in fact GloFo) need to adress - so performance/watt charts will make BD look bad in particular. At least I'd be interested to hear why this particular aspect is so interesting that it makes or breaks a chip..? Again - quite unbalanced way to place the focus I think.

High power consumption (most notably when overclocked) relative to SB is due to the fact that AMD/GF are sticking with strained SIO as their substrate material (hope that's changing with the 22/20nm nodes) and the large number of transistors (~2 Billion!).

Neither of these will change significantly over the next two years (unless BD is going to GF's 28nm). If GF has/is developing a 28nm HP sub-process for CPUs, it will take AMD ~18 months to do the die shrink. A 1/2 shrink will only be worth it is AMD has already started, else they will need all their manpower for the full node die shrink.

 

[frostedflakes]

All else being equal, SOI should actually have lower leakage and lower power consumption. Intel's process technology is just that good. It's really quite amazing how much they've been able to accomplish with bulk Si.

 

[LOL_Wut_Axel]

Where did I say Bulldozer is "good"? I think it's good at some things and not very good at others. Balanced means that I don't think it's quite as simple as just saying that it's a bad chip and "suck it up". It has some merits, and the concept is interesting.

Since when did performance/watt become king? Why are we not looking at absolute performance? Quite a lot of enthusiasts overclock - a sure killer of "performance/watt". Power consumption is one of the areas AMD (or maybe in fact GloFo) need to adress - so performance/watt charts will make BD look bad in particular. At least I'd be interested to hear why this particular aspect is so interesting that it makes or breaks a chip..? Again - quite unbalanced way to place the focus I think.

How about that FX8150 beats 2600K in Handbrake - a fairly cpu hungry application a lot of enthusiasts use. http://www.legitreviews.com/article/1751/6/

Cyberlink Media expresso 6.5 - another good example: http://www.legitreviews.com/article/1751/5/

...and then Bulldozer looses in others. But to claim it's a failure is out of proportions.


Source? Does Bulldozer not in fact hold the record in raw GHz
overclock..?

Roadmaps are subject to change and should be taken with a huge grain of salt. AMD didn't stay in business for 40+ years by not being able to adjust to the current market situation.

LOL, AMD fanboys these days. Sandy Bridge is great when overclocking because performance/watt actually increases until the point most people get: 4.5-4.6GHz. The minimum additional voltage and the fact the die is so small means that little additional heat is produced, and performance increases so much while power consumption increases so little that efficiency is HIGHER at those speeds than running at stock.

Bulldozer doesn't "win some, lose some". In 80% programs it'll be outright slower, and in others it can match or beat the 2500K by a small margin. That you're able to find one perticular benchmark where it performs well and use it as "evidence" that it has its merits proves how much of a crappy CPU it is. It needs to be shown in a good light by people that tell you to forget about all other twenty benchmarks and focus on one exception instead.

Performance/watt is and has always been an extremely important part of computing, especially in servers. That's why Intel will still dominate there as well; they give everyone from small businesses to large enterprises much more performance at the target power they want to achieve, and the configurable TDP in Ivy Bridge is another idea that will cement their lead in that market.

On desktops it's less relevant, but once you get to a certain point it's definitely something to take into account. It's not just the extra power consumption, but also the heat that it outputs. All your other components will run hotter because of it and it could quite literally become a room heater. Then there's the fact it's so expensive for the performance it gives, not to mention that it's an example of engineering laziness: you have a CPU with over two times the number of transistors and yet it's convincingly beaten, not to mention the other CPU was launched three quarters earlier.

Bulldozer holding a record for highest overclock means NOTHING. It does not apply to any real-world situation. It was done using extreme methods of cooling and the only reason why it got there is because the multiplier and base clock is completely unlocked unlike unlocked Sandy Bridge CPUs where you can only get a few MHz from BCLK overclocking and the Turbo functionality-enabled higher multipliers only let you go to 57x. The stupid CPU department at AMD thought it would be a good idea to make the CPU with a longer pipeline even though they should've seen from the beginning they weren't gonna hit their target 4.5GHz+ frequencies so BD could be competitive.

 

[bryanW1995]

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

Um, what? The human body can handle any speed short of light just fine, it's the acceleration that gets you . And jets actually are constrained by the melting point of materials used in their construction. SR-71's, even though they fly at extremely high altitudes/extremely thin atmospheric conditions, were almost at the limit of the materials used to make them (though that was 40+ years ago admittedly). Having said all of that, the straight line speed of jets is more constrained by the materials used to build the plane, while the abiltiy to change direction rapidly is contrainted by the human body's inability to handle the extremely high G forces caused by the huge/rapid velociy changes (aka "acceleration").

^ unimportance of frame of reference comprehension fail!

(hint: speed has 0 effect on the human body, you're hurtling through space compared to the sun at over 30KM/s right now, and 10 times that relative to the center of the galaxy. Don't feel the speed? That's because speed exerts no force!)

Sheesh, lots of fail in this thread. Speed exerts a very small but finite amount of force on an object that grows quite rapidly as you approach the speed of light. Ever think about the "force" required to keep your relative speed at the speed of light even though you're traveling at .99c "north" while another spaceship travels at .99c "south" of the galactic plane?

 

[bryanW1995]

Where did I say Bulldozer is "good"? I think it's good at some things and not very good at others. Balanced means that I don't think it's quite as simple as just saying that it's a bad chip and "suck it up". It has some merits, and the concept is interesting.

Since when did performance/watt become king? Why are we not looking at absolute performance? Quite a lot of enthusiasts overclock - a sure killer of "performance/watt". Power consumption is one of the areas AMD (or maybe in fact GloFo) need to adress - so performance/watt charts will make BD look bad in particular. At least I'd be interested to hear why this particular aspect is so interesting that it makes or breaks a chip..? Again - quite unbalanced way to place the focus I think.

How about that FX8150 beats 2600K in Handbrake - a fairly cpu hungry application a lot of enthusiasts use. http://www.legitreviews.com/article/1751/6/

Cyberlink Media expresso 6.5 - another good example: http://www.legitreviews.com/article/1751/5/

...and then Bulldozer looses in others. But to claim it's a failure is out of proportions.


Source? Does Bulldozer not in fact hold the record in raw GHz
overclock..?

Roadmaps are subject to change and should be taken with a huge grain of salt. AMD didn't stay in business for 40+ years by not being able to adjust to the current market situation.

This is similar to gtx 480's problem. BD uses WAY too much power for the performance that it gives. If an 8150 was 10-15% faster than 2600k then the power usage would still be an issue, but not a deal killer. For 10-15% SLOWER performance, however, you need to offer something better in another area; ie, lower power consumption.

 

[lol123]

Um, what? The human body can handle any speed short of light just fine, it's the acceleration that gets you . And jets actually are constrained by the melting point of materials used in their construction. SR-71's, even though they fly at extremely high altitudes/extremely thin atmospheric conditions, were almost at the limit of the materials used to make them (though that was 40+ years ago admittedly). Having said all of that, the straight line speed of jets is more constrained by the materials used to build the plane, while the abiltiy to change direction rapidly is contrainted by the human body's inability to handle the extremely high G forces caused by the huge/rapid velociy changes (aka "acceleration").

That's all very true. However there was the MiG-25 as an example of a fighter (or more accurately an interceptor) that could fly at Mach 3+, and which used titanium on the leading edge of the wings to withstand the temperatures created at those speeds, although if it went much faster than that the turbine blades in the engines would actually melt. The major problem with that design though was that it was practically impossible to change its flight direction at or anywhere near its designed max velocity.

This is an interesting discussion, although I would say it's rather off-topic.

 

[Ferzerp]

Um, what? The human body can handle any speed short of light just fine, it's the acceleration that gets you . And jets actually are constrained by the melting point of materials used in their construction. SR-71's, even though they fly at extremely high altitudes/extremely thin atmospheric conditions, were almost at the limit of the materials used to make them (though that was 40+ years ago admittedly). Having said all of that, the straight line speed of jets is more constrained by the materials used to build the plane, while the abiltiy to change direction rapidly is contrainted by the human body's inability to handle the extremely high G forces caused by the huge/rapid velociy changes (aka "acceleration").



Sheesh, lots of fail in this thread. Speed exerts a very small but finite amount of force on an object that grows quite rapidly as you approach the speed of light. Ever think about the "force" required to keep your relative speed at the speed of light even though you're traveling at .99c "north" while another spaceship travels at .99c "south" of the galactic plane?

You're missing a bit of the intricacies of special relativity here.

 

[Cerb]

At least I'd be interested to hear why this particular aspect is so interesting that it makes or breaks a chip..? Again - quite unbalanced way to place the focus I think.

You can't reasonably sell a CPU that uses more power than a fast Prescott. I don't mean cherry-picked unlocked models, but normal CPUs for people that don't have $100 of cooling equipment in their cases.

If each CPU technically can handle 5GHz on air, let's say, but uses >140W or so doing it, it can't be sold as that fast. Period. It needs to offer good performance at speeds that don't push the voltage limits.

Major OEM demand does affect prices, and I don't see them clamoring for anything much hotter than desktop Athlon II CPUs, these days.

Portable computers matter more than desktops. Llano is good, but it needs follow-ons that can do more processing with less power. The more power it uses, the more likely a more expensive Intel part will be chosen, both by big vendors and we who buy 99% of our PC parts at Newegg.

To a point, not having the raw performance, nor performance/Watt, of the competition, is OK. Athlon IIs were cheap enough. Most Phenom IIs were cheap enough, to the right users (the L3 needs certain workloads, the speedy no-L3 ones beg for overclockers, and the X6 need killer applications). Idle power sucks, but S3 makes that a lesser issue, for a non-server.

The GTX 480 analogy is apt; as are nVidia's prior GT2xx efforts. In those cases, however, the raw performance at least knocked your socks off, until you removed your ear plugs.

 

[exar333]

To anyone who doesn't believe BD is a pile of crap, read this review. Power consumption and efficiency review from TH. Very well done (I am hoping AT does one of these as well)

http://www.tomshardware.com/reviews/fx-power-consumption-efficiency,3060.html

 

[iXombie]

If you’re a huge fun of AMD buy the chip, an AM3+ board, 1200 watt (this is not sarcastic but is it appropriate for over clocking to 4.5 Ghz?) power supply and go nuts. But for the same money ($280 USD from microcenter) I could get a 2600k, 850 watt power supply and better performance with the exception of the twice a year I do media encoding. To defend this chip like it is your first born is just as ridiculous as people who are OS zealots. I use Windows 7 (web development), OSX 10.7 (music production) and BackTrack Linux (security consulting) and none of them are perfect. If you hate Intel, that’s ok by most of us, by all means purchase whatever you can live with. I hate getting spending MY money on inferior products. I could care less what anybody else does with THEIR money.
http://www.anandtech.com/bench/Product/287?vs=434

 

[Ajay]

All else being equal, SOI should actually have lower leakage and lower power consumption. Intel's process technology is just that good. It's really quite amazing how much they've been able to accomplish with bulk Si.

Good point!

 

[antisocialmunky]

You're missing a bit of the intricacies of special relativity here.

I am a neutrino, your reality is invalid +/- 64ns.

 

[bononos]

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.

Nonsense. Ordinary users who are thinking about a FX4100 or 6xxx have a perfectly legitimate complaint against BD. For the typical gamer, the BD performs poorly

CPUs may be overpowered for you if all you do is dabble with the wordprocessor. The design is more advanced and farsighted but it is also gimped so it won't perform as well as you think in the future.

 

[piesquared]

Nonsense. Ordinary users who are thinking about a FX4100 or 6xxx have a perfectly legitimate complaint against BD. For the typical gamer, the BD performs poorly

CPUs may be overpowered for you if all you do is dabble with the wordprocessor. The design is more advanced and farsighted but it is also gimped so it won't perform as well as you think in the future.

There's nothing typical about the gamers you're talking about. The consumers that buy CPU's and GPU's specifically for gaming are a pretty small crowd.

 

[zlejedi]

There's nothing typical about the gamers you're talking about. The consumers that buy CPU's and GPU's specifically for gaming are a pretty small crowd.

And those who buy it for web browsing and MS Word will be fine served with dual core SB celeron for 40 bucks.

 

[piesquared]

And those who buy it for web browsing and MS Word will be fine served with dual core SB celeron for 40 bucks.

As they will with the Brazos platform for even cheaper. I suppose Atom is cheaper yet but nobody wants it, as it's useless.

 

[bononos]

There's nothing typical about the gamers you're talking about. The consumers that buy CPU's and GPU's specifically for gaming are a pretty small crowd.

So?
It doesn't speak well of BD if that is all you and the other poster (I was replying to in my previous post) can offer. BD is gimped and theres no way around that fact.

 

[piesquared]

So?
It doesn't speak well of BD if that is all you and the other poster (I was replying to in my previous post) can offer. BD is gimped and theres no way around that fact.

And many don't agree with you, there's no way around that fact.

 

[bononos]

And many don't agree with you, there's no way around that fact.

Ok. Assuming you aren't merely being flippant, what was it that AMD was trying to achieve all these years with BD? It was definitely not some low end chip for casual consumers. AMD has tried its damndest for the past few years and produced BD is a 2B transistor monster of a cpu which was supposed to wipe up the competition handily but has problems outperforming its predecessor in some apps esp games to say nothing of SB.

So what is it about BD which is so superior (from what I gather from your tone)?

 

[NostaSeronx]

So what is it about BD which is so superior (from what I gather from your tone)?

Multithread Multitask

Cinebench(same as logical cores) or x264(1.5x CPU logical cores), Video Game(2+ logical cores), and other background tasks

But other than that it fails in

Singlethread Single task <-- people think these are important
Singlethread Multitask <-- When this and above Multithread Multitask are more important for desktop environments(Lossless WAV/FLAC/Redbook -> MP3/AAC encoding 8 songs at a time with all 8 threads Sandy Bridge is so much faster than Bulldozer)
Multithread Single task <-- people think these are important

 

[exar333]

Multithread Multitask

Cinebench(same as logical cores) or x264(1.5x CPU logical cores), Video Game(2+ logical cores), and other background tasks

But other than that it fails in

Singlethread Single task <-- people think these are important
Singlethread Multitask <-- When this and above Multithread Multitask are more important for desktop environments(Lossless WAV/FLAC/Redbook -> MP3/AAC encoding 8 songs at a time with all 8 threads Sandy Bridge is so much faster than Bulldozer)
Multithread Single task <-- people think these are important

Take a look at the link I provided a few posts up from TH. They test multi-thread and multi task scenarios, and BD struggles to keep up with the 2500K that has 1/2 the cores and 1/2 the threads.