Did Bulldozer Damage the AMD brand?

[pelov]

It's slated for an early 1H 2013 release. At this point it's impossible for them not to have an engineering sample in hand.

I'm hoping it's a case of them learning from their mistakes, which is to say: put up or shut up. Their estimates for Trinity were pretty much spot on with no overestimation. I don't think they overhyped the release either. It was represented in the same way it performed, which was mostly GPU increase and HSA advancements. Vishera seems much the same, as AMD has said that we should expect a 15% improvement which sounds accurate when comparing Husky to Trinity or Trinity to FX4xxx.

Unfortunately, they can come out with a competitive chip on the desktop/laptop and I don't think it would turn their sinking ship around. The enthusiast ship has set sail already and attempting to regain that end of the market isn't the recipe to long term success. I know I'm parroting myself, but unless AMD have a very good Jaguar core that scales down well or an SoC brewing for the past couple of years, things aren't looking good for them. I just hope this isn't another Bobcat, in that they have a great chip but they were late in execution and brought it to a market that's already changed its course

 

[Ancalagon44]

Unfortunately, they can come out with a competitive chip on the desktop/laptop and I don't think it would turn their sinking ship around. The enthusiast ship has set sail already and attempting to regain that end of the market isn't the recipe to long term success. I know I'm parroting myself, but unless AMD have a very good Jaguar core that scales down well or an SoC brewing for the past couple of years, things aren't looking good for them.

I think this is the problem.

Look, we enthusiasts are a small portion of actual consumers. Actual consumers probably wont care either way between Intel and AMD. So, if AMD can have a number of well priced Fusion laptops available at the lower end of the market, they might actually sell some.

As for enthusiasts, we are a much tougher sell. Not only do we check benchmarks before buying, we also remember things like BD. AMD can forget about using marketing to sell its processors now - only good benchmarks will sell processors in the enthusiast community. And I honestly dont know if AMD can catch up to Intel in the enthusiast space - they screwed up so badly with BD that I just dont know. Had they instead pushed for improving and re-architecting Thuban, they might have been able to do it.

The other thing that I think is a pity is that they through away the momentum that they had A) after the Athlon 64 days, B) after the lawsuit with Intel was settled resulting in a large payout for AMD and billions of dollars of fines levied against Intel. They had all of this good momentum, and goodwill from gamers like me who supported the underdog, and they through it away. Through bad management.

What were they doing in the Conroe days? How is they never caught up?

 

[MrColin]

This thread is the first I've heard of BD. I'm going to have to say this hurts the brand terribly.in my view.

 

[Ajay]

It's slated for an early 1H 2013 release. At this point it's impossible for them not to have an engineering sample in hand.

I'm hoping it's a case of them learning from their mistakes, which is to say: put up or shut up. Their estimates for Trinity were pretty much spot on with no overestimation. I don't think they overhyped the release either. It was represented in the same way it performed, which was mostly GPU increase and HSA advancements. Vishera seems much the same, as AMD has said that we should expect a 15% improvement which sounds accurate when comparing Husky to Trinity or Trinity to FX4xxx.

Even if SR comes out 2H13, AMD would need an engineering run in hand, otherwise they won't likely hit even that target. They seem to be sticking to the 15% performance/watt increase between generations, at least officially. On paper Steamroller looks better than that (anywhere from 30-50% has been bandied about). Given IDC's comments on the apparent failure of their simulations predictive capabilities (poor), I plan on tamping down my enthusiasm - unless some tech site gets and ES sample and preliminary benches actually look good.

 

[taltamir]

I think the Bulldozer design is excellent in theory, but was far more difficult in practice to actually design and implement

So was the p4 on which BD is based on. Intel dropped this design because all those "in theory" ideas did not work well IRL.

 

[BenchPress]

On an igpu, yes. On a real GPU though, I'd say it will make perfect sense for a long time to come, but then again, only on the specific workloads that it already makes sense for.

Consumer applications that use OpenCL are close to non-existent, let alone ones offering a sizable advantage worth buying an expensive discrete AMD graphics card for.

It's really an insignificant market, so it doesn't offer AMD any noticeable advantage. Meanwhile Intel is more than doubling the SIMD peformance for nearly every market segment, at practically no extra cost (on the order of 10% more transistors, compensated by denser mature process technology). So the reasons for getting Intel hardware increase a lot more than the reasons for getting AMD hardware.

AMD buying (all of) ATI was a huge mistake. It made a big dent in their finances, which led to cutting back on R&D, which led to the disaster called Bulldozer. And it all stems from the fairy tale that GPU silicon is somehow faster than CPU silicon. It's not. Intel is quadrupling the CPU throughput per core over just a few years' time with little effort. No GPU expertise required.

 

[Ferzerp]

Consumer applications that use OpenCL are close to non-existent, let alone ones offering a sizable advantage worth buying an expensive discrete AMD graphics card for.

It's really an insignificant market, so it doesn't offer AMD any noticeable advantage. Meanwhile Intel is more than doubling the SIMD peformance for nearly every market segment, at practically no extra cost (on the order of 10% more transistors, compensated by denser mature process technology). So the reasons for getting Intel hardware increase a lot more than the reasons for getting AMD hardware.

AMD buying (all of) ATI was a huge mistake. It made a big dent in their finances, which led to cutting back on R&D, which led to the disaster called Bulldozer. And it all stems from the fairy tale that GPU silicon is somehow faster than CPU silicon. It's not. Intel is quadrupling the CPU throughput per core over just a few years' time with little effort. No GPU expertise required.

Notice I said only the instances it makes sense for right now. I didn't claim it was more than a niche. Nor did I disagree that doing it on an igpu isn't particularly bright when massively faster gpus exist.

 

[Ajay]

AMD buying (all of) ATI was a huge mistake. It made a big dent in their finances, which led to cutting back on R&D, which led to the disaster called Bulldozer. And it all stems from the fairy tale that GPU silicon is somehow faster than CPU silicon. It's not. Intel is quadrupling the CPU throughput per core over just a few years' time with little effort. No GPU expertise required.

Well, probably allot of effort, but far less money than what AMD paid for ATI.

The sad part about it is that AMD was looking to jump start their combined CPU/GPU program for desktops and laptops by buying ATI, but the cost was so high that it had did too much injury to the CPU side of the business.

 

[taltamir]

AMD buying (all of) ATI was a huge mistake. It made a big dent in their finances, which led to cutting back on R&D, which led to the disaster called Bulldozer.

This is not related to reality. the problem with BD is that it is meant to be "Pentium 4 done right". Its fundamental issue with the architecture and design.
Hypothetical cuts in R&D are not the issue and even if they were the fact that so many years later BD derivatives still all suck show that extra R&D money wouldn't have solved that (as intel saw when they poured so much R&D money into the mistake that was P4)

 

[georgec84]

Poor leadership damaged the AMD brand. Poor leadership and awful decision making.

 

[ThatsABigOne]

Yes it did.