Opnion: Did We Expect Too Much of AMD Bulldozer?

[dreamer77dd]

i thought this video showing how 8 cores is used. http://youtu.be/FgtzwU9iXnc
It might not be the fastest but it can carry a load. thinking truck. It also cheap for what it is. No new motherboard, gpu, and 8 cores for the future of software. People it in the hands of the people and so developers will develop more for 8 core chips. Soon Intel will have 50, 100 cores. ect. Well not really soon but working on it t be soon.
"People are mostly complaining because single threaded performance decreased from the Phenom II in many cases, so the gap to sandy bridge is huge.. software's limited to 1-4 cores usage are still VERY common, so that's why, but yes, BD can be interesting for some software" by @AustinKing09

 

[RogueKitsune]

Bulldozer makes for a great server part, but a pretty meh desktop part. It excels in very highly threaded applications or running a bunch of programs at once(using all of its cores). Which is the usual server environment. In a desktop environment most applications hardly use more than 2 threads, and those that do don't use more than 4. All is not lost for AMD(I hope) because this is a new architecture which, hopefully they can pull more power from with farther revisions, and as time advances maybe we will see more and more application take advantage of more than 4 cores.

But that could just be my wishful thinking.....

 

[Guest]

Read this article first, then went and read the review with benchmarks. Sandybridge doesn't just beat Bulldozer... it wipes the floor with it, I didn't see any mixed reviews, everything showed the i5 and i7 smoking it and that's before taking into account the i5's awesome overclocking ability.

 

[molo9000]

It's actually a pretty neat and innovative CPU.
The problem is software.
Writing multithreaded code is not trivial and programmers are used to ever increasing single thread performance, so why bother writing efficient multi threaded code unless u actually are writing some critical productivity app?

The FX8150 actually beats the i7-2600k in some benchmarks. Shame that most games can't use 8 cores.

 

[beenthere]

Opinions vary but criminal convictions are forever.

 

[Guest]

It was well known that Bulldozer would not compete with Sandybridge, well before the launch time, and if anyone saw the helium overclocking parade and automatically thought they were getting an 8ghz cpu at home, they were just delusional.

If Bulldozer didn't perform worse than it's predecessors in certain situations then it wouldn't had received the amount of flak it has.

At any rate, AMD needs to fire their pr and marketing department, this too has been knows for quite a while now.

 

[dalmvern]

As many people have said, I dont think we expected too much. We expected a chip from AMD that would at least be competitive with Sandy Bridge, hopefully outperforming it and setting the bar for Ivy Bridge.

Granted, the chip had a new architecture which will surely be improved on and revised but for now, Bulldozer is a bit of a disappointment.

 

[JohnnyLucky]

I did not expect anything from Bulldozer. On the other hand it appears gamers and enthusiasts expected too much. More than 30 hardware reviews of the new Bulldozer were published yesterday. That is a record number of reviews for one pc component in one 24 hour period. The reviews were not positive.

We have a somewhat similar situation with Intel's new 520 Series ssd that will be released on November 4th. The are unsubstantiated rumors and speculation that the new ssd will use a SadForce controller. As usual Intel has remained silent.

 

[captaincharisma]

[citation][nom]beenthere[/nom]Opinions vary. That's why people should learn what performs best for their use and then buy AMD instead of supporting the convicted crinals at Intel.[/citation]

sorry but criminal (as the fanboi troll calls them) or not i am not going to go for the loser when the winner outperforms them in almost every way

 

[dgingeri]

The big expectations came from this being a new architecture. Why make a new architecture if not to improve performance? Why bother spending billions on engineering and research, only to take a step back from where you are? AMD did exactly this. They stepped backward, lost performance per clock, and overall barely stayed right where they were. This is the most upsetting thing with this chip.

 

[mrmaia]

Now even if AMD comes out with the world's fastest CPU for real, any marketing it does on that will have absolutely NO effect at all. AMD is getting discredited with all that bull... yes, bullshit.

 

[rjq]

I think people was expecting way too much, almost a unrealistic scenario for this processors, but it was until I saw the Newegg's video that I realized that they are going to use this processor (and the price was a big clue) as part of a Three pieces product that the called "scorpio platform??" which is necessary to GET the desired performance.... in the mean while we are getting some phenoms for the budget rigs and that would do for two-three years hopping the architecture had mature enough.

 

[Guest]

Interlagos Opteron 16 cores at 1.8ghz is a fail also since you cannot use this chip as a workstation way to low for Single threaded use.

 

[Guest]

Well my two cents on the subject are as follows, I own a T1090 and see no reason to up grade to BD.

 

[killerclick]

I knew exactly what BD would turn out to be years ago and I've been shouted down by fanboys who said I don't know what I'm talking about.

Fanboys should realize there are no shortcuts in this business, there are no magic breakthroughs and the product/tech maps are drawn years in advance. Intel has been making all the right moves since 2006 while AMD has been stumbling. The reverse would have to be true (for another 5 years) for AMD to even hope of catching up to Intel.

 

[Pixelgrease]

AMD needs a CPU that is designed for the future. Given current semiconductor technology that future is more CPUs on a die, not faster clock speed. Bulldozer allows rapid turnaround on new CPU designs by simply adding more cores. Previous designs did not have quite so scalable an intraCPU data transfer architecture.
To beat Intel with bulldozer modules, AMD must lead with number of Bulldozer modules cores versus Intel cores. For example: 4 modules (8 cores under Windows) versus 4 Intel cores plus hyperthreading (8 cores under Windows) -- AMD needs a "10 core" CPU to beat a 4 core Intel CPU. Sounds bad, but if they are the same price I would go AMD.
At 32nm, AMD should be able to scale beyond 4 module (8 cores) and it looks like they have done 8 modules (16 cores).
Single-threaded speed is not worth what it used to be; the future is multi-threaded and being able to scale logical cores efficiently is key -- not single-threaded execution.

 

[peroludiarom]

There are too many variables - is it windows fully compatible, intel compiler reads wright CPUID, MoBos fully compatible with the new hardware. After all this is new architecture, and we cannot expect to be perfect (somebody may be don't remember Core Is issues with first mobos).

And for me guys, all day around I'm reading post like "such a disappointment, i just order my I5" - this is acting like a herd of sheeps (sorry but this is truth) - see one - two benches, not even wait a week to see more opinions - and one sheep goes, and others after her.

At last place, marketing was go too far, and somebody has to be fired.

 

[back_by_demand]

I never expected too much, seeing as Intel are selling the dozen or so CPUs that are faster for the best part of $1000+ each, with the odd notable exception. It is a very powerful CPU and would make a great upgrade from a previous generation, AMD will do great business in the mid-range. Perhaps attempting to take the performance crown was something to wish for but the smart money is on steady sales in a known market.

 

[thesalopian]

"Unfortunately, AMD has never been meant to take over the high end spectrum since its creation..."

AMD Athlon vs. Intel Pentium 4 anyone?? (I guess it's what you call high end though... so my apologies for being annoying here)

Talking about the Pentium 4 it does seem like the architecture Bulldozer is AMD's netburst in terms of failing to deliver. I really hope they something special in the bag but I think the problem is this should have come out when they first said. Maybe then it would have stood a chance. It's a real shame. I've used both Intel and AMD products and go for whatever gets me the best performance for the money I have available to spend at that time. We need at least one serious competitor in the x86-64 CPU market to drive prices down and, most importantly, innovation!

 

[bison88]

I think AMD expected more from AMD. Even a lot of AMD fanboys I know had that gut feeling that it would be a step in the right direction on the competitive and design with a competitive price. I'm not ecstatic nor upset with the benchmarks I've seen. Kind of what I expected. You have to remember if this chip came out years ago before the delays after delays we probably wouldn't be having this conversation, those delays ultimately feel like it's already an aged product for people that waited for so long.

I feel more upset for the people that waited so long and with all the talk and hype from AMD, that they are getting what they're getting. But the price is nice, the cores can come in handy in the future or depending on what you're doing, and things should get better. AMD just needs to meet there deadlines as best as possible. They don't have the backing, R&D, or finances like Intel, but they really do need to do something before they are considered a distant competitor. Right now Intel hasn't really messed with the clock speed or adding more cores, but on the architect and design of the chip to boost functionality. AMD needs to get back into a Core-to-Core performance like Intel and stop thinking adding more cores is going to make there problems go away.

 

[singemagique]

It's a brand news architecture and it brings 8 cores to the mainstream. Now if they can get the price down about $50USD I think it's a great addition. Sure it's not the monster all the advertising had hyped, for that AMD needs to re-examine their marketing strategies. However, I think it can definitely find a niche of consumers that can make use of the increased amount of cores (ie. one of these would be great and cheap for molecular modeling).

 

[Guest]

It consumes far too much power for a 32nm chip design, making the "performance per watt" argument invalid for most desktop users. It also uses far too much power overclocking for too little gains. AMD should have told us Bulldozer would be a niche product for those that use 8 threads constantly that want something cheaper than an i7-2600 and don't mind having to pay extra on their electric bill. Even for most gamers, a Sandy Bridge i3 will be better in most cases because of superior single thread performance and higher frame-rates in games that are CPU bound (the i3 generally outperforms Phenom IIs at 3.7ghz in gaming fps). I just don't see a place for bulldozer in desktop computing right now.

 

[afrobacon]

I was really expecting Zambezi to be not quite as fast as Nehalem but make a valiant effort that would force Intel to lower their prices. Though there is still time for this to happen I can't imagine it making a huge difference.

 

[masterasia]

By releasing Bulldozer, AMD is telling the world that it is now more than 3 years behind Intel because the i7 920 still kicks ass and that was released in late 2008.

 

[Prey]

You were right in your expectation, it was like Duke Nukem Forever, talk and talk and talk, and when you see it.. ehh.

After all that hype, only to say, "ehh, it's alright." I am by no means a fan boy, I wish AMD would have rocked em, better competition means lower prices ! AMD fell short of the mark.