CNET: Is AMD still relevant?

[Puppies04]

Since this thread is going to hell in a handbasket. I'll say it.

LOL_Wut_Axel stops posting and futurefields takes over at that time. I wish I had access to the IP access lists.

Or maybe it is me bwah bwah bwah, oh actually it isn't.

 

[futurefields]

I don't even know what that means. But please, anybody that wants to debate me on one of my actual points, please do so.

My points are as follows:

1) AMD is aimed at bottom feeder market who want the quickest cheapest way to play games like WoW and Diablo 3.

2) If you spend a little bit more initial investment, you get an Intel machine that is not only higher total performance, but higher performance per dollar.

 

[futurefields]

Consumers could care less about benchmarks, they are meaningless. They're only relevant to the shrinking online benchmarking addicts. Consumers care about user experience and how they interact with their devices on a daily bases. If their system gives them off the chart SuperPi numbers but stumbles and stutters in a game they like to play, do you honestly believe they would give a rats ass about the benchmark? If benchmarking is your holy grail, run some OpenCl or C++ AMP compiled benchmarks. Benchmarking can give you any number of different results.

Why did you only respond to the first part of my post? How about the next part, this part --

Download a free to try audio program like REAPER and compare how many audio plugins you can run on the slowest i5 compared to the fastest AMD bulldozer before your cpu starts to choke.

You will see what a joke AMD's architecture is

That's not a benchmark. That's a real world scenario I encounter on a daily basis, where AMD just doesn't cut it. Talk to audio and video editors around the world about what brand of CPU they prefer, and why.

 

[WhoBeDaPlaya]

Why should I care about integrated graphics again?

That's all you're gonna be able to afford with that kid on the way

 

[borisvodofsky]

AMD is still very relevant up till the $500 budget price range.

If you're willing to spend more than that, then Intel has the better performance to price ratio.

 

[borisvodofsky]

That's all you're gonna be able to afford with that kid on the way

GT 240 bro, GT 240, I'll donate one for free w/ proof of poorness caused by the baby

 

[borisvodofsky]

Why did you only respond to the first part of my post? How about the next part, this part --

That's not a benchmark. That's a real world scenario I encounter on a daily basis, where AMD just doesn't cut it. Talk to audio and video editors around the world about what brand of CPU they prefer, and why.

Bah... Yea, memory controller isn't as snappy on the AMD cpus. Still, the CHOKE you're talking about doesn't really happen on anything >$600.

If these editors are choking, then it's cuz they're poor, it's not really AMD's fault.

 

[piesquared]

Why did you only respond to the first part of my post? How about the next part, this part --

That's not a benchmark. That's a real world scenario I encounter on a daily basis, where AMD just doesn't cut it. Talk to audio and video editors around the world about what brand of CPU they prefer, and why.

Why are you only responding to the merits of Bulldozer, and ignoring APU's, OpenCl, C++ AMP and HSA? Heterogeneous computing has emerged as a very viable and potent solution to accelerate applications beyond what legacy x86 compute can do.

Here's Robert Rodriguez' take: "if the cloud is the wild west, then the APU is the bigger gun and I need that"

http://www.youtube.com/watch?feature=player_detailpage&v=kuWbgYcBUyQ

And the recent CS6 release supports OpenCl and OpenGl accelerators where Trinity outperforms the fastest ivy bridge APU by 2x. @ ~27:00

http://www.youtube.com/watch?v=uIYWFej-_wg&feature=player_detailpage

There are over 100 OpenCl accelerated applications and growing daily.

 

[ShintaiDK]

I always love the empty promises for the future with already failed technologies.

Its always..we just have to wait...wait...and then wait some more. The solution to fix the failed product is always just around the corner. Just always seems to slip around the next corner. Soon everyone will use it. Soon just aint coming.

And before you say x companies are backing the technology. Less than 1 out of 100 of those techs all the companies back tend to get somewhere besides a tiny niche.

 

[piesquared]

Nothing to wait for. That's just a marketing meme propagated by the dishonest. The software is out there and the benefits are real and tangible.

 

[ShintaiDK]

OpenCL is basicly mimmicking CUDA now. OpenCL apps are doing nothing more that CUDA aint already doing. How did that go? Oh ye....

But if we really hope alot it might just happen. Lets sit down and pray ./\.

 

[2cents]

I guess i'll give my two cents. I believe AMD is still relevant in many regards, their GPUs for one. Piesquared is right, the general population doesn't care about benchmarks and we can't forget what AMD has to offer, a very capable processor for a good day's worth of work. Sure Intel has AMD beat when it comes to performance, but honestly selling that idea to the average consumer doesn't seem very plausible. And true, you're going to get a few that are willing to dish out some extra cash for the best. We also have to remember that we are enthusiasts in a niche market. This is our field of expertise and hopefully we get the main idea. Honestly though if you approach Joe and offer him two different types of computers (one that is AMD and one that is Intel), it's safe to say the first question he will ask is, “How much?”. This is where AMD might shine. In my opinion, AMD has a great market strategy and honestly it appears to be working, even though it might take awhile for the numbers to reflex.

 

[piesquared]

OpenCL is basicly mimmicking CUDA now. OpenCL apps are doing nothing more that CUDA aint already doing. How did that go? Oh ye....

But if we really hope alot it might just happen. Lets sit down and pray ./\.

Ignorance is bliss. The software is there and the big ISV's are behind it, not quite the same thing CUDA isn't already doing.

But if we really hope a lot, it might just not happen. How's that praying going? ./\.

 

[ShintaiDK]

Come back when OpenCL can do more. Might be never

 

[piesquared]

Can do more what? GPGPU already outperforms x86 CPU compute in many applications. Microsoft apparantely thinks there's merit in heterogeneous computing, they invested heavily in C++ AMP.

 

[Vesku]

DirectX is basicly mimmicking Glide now. DirectX apps are doing nothing more that Glide aint already doing. How did that go? Oh ye....

OpenCL is basicly mimmicking CUDA now. OpenCL apps are doing nothing more that CUDA aint already doing. How did that go? Oh ye....

But if we really hope alot it might just happen. Lets sit down and pray ./\.

 

[ShintaiDK]

DirectX is basicly mimmicking Glide now. DirectX apps are doing nothing more that Glide aint already doing. How did that go? Oh ye....

API for games still only used for games. Exactly, you prove my point.

 

[futurefields]

Why are you only responding to the merits of Bulldozer, and ignoring APU's, OpenCl, C++ AMP and HSA? Heterogeneous computing has emerged as a very viable and potent solution to accelerate applications beyond what legacy x86 compute can do.

Here's Robert Rodriguez' take: "if the cloud is the wild west, then the APU is the bigger gun and I need that"

http://www.youtube.com/watch?feature=player_detailpage&v=kuWbgYcBUyQ

And the recent CS6 release supports OpenCl and OpenGl accelerators where Trinity outperforms the fastest ivy bridge APU by 2x. @ ~27:00

http://www.youtube.com/watch?v=uIYWFej-_wg&feature=player_detailpage

There are over 100 OpenCl accelerated applications and growing daily.

That is hilarious.

 

[futurefields]

edit- nevermind

 

[sm625]

AMD's target market is the bottom feeder.

"I just want the cheapest computer that will run WoW and D3 @ 30-40fps"

And that's fine.. that's what they are there for. I just think it's hilarious when people try to act as though they are still somehow competitive with Intel.

The fact is Intel could put AMD out of business. They could easily price their chips even less, and just completely annihilate the price/performance benchmark, more than they do already. Intel *let's* AMD survive.

Intel would never survive trying to operate at AMD's margins. AMD lives on margins that would kill Intel. The shareholders would revolt. The payoffs and the bribery would stop. The company would fall apart within 4 quarters of AMD style margins. The CEO would be gone and the battle would stop.

As a curious side note, Intel makes its money by providing payoffs such as the ultrabook bribes. Those payoffs dont count against its margins. So what Intel is basically doing, in addition to the obvious crimes (for which they will be found guilty of probably 3-4 years from now) is cooking their books. The concept is simple: Say you make widgets for $100 apiece. You sell them for $300 leaving you a nice wonderful 67% margin. Everything is good, everyone is happy. But what do you do if you cannot sell your widgits for $300 anymore? Easy. You create an UltraWidget standard and hand out a bunch of money to people to encourage them to build systems using your widget (and only your widget) for a lower cost. Now with the total systems costs lowered, you have just inflated the demand for your $300 widgets. So you can keep your prices at $300 per widget and thus your margins remain at 67%. Obviously this is a cut and dry accounting scam worthy of a JP Morgan bankster swine Wall Street Wall of Shame. But this is exactly, exactly what Intel is doing if they do not count these ultrabook bribes against their margins. They would be attempting to cover up falling margins. That would be very bad news for the entire sector! These are the kind of games that result in market moves like what happened in 2008. But that of course is another story.

 

[MisterMac]

One could argue that if 14nm tapes out without issues - then possibly AMD will be come irrelevant\broke or become multi x86\ARM designer minor player.
Providing HSA designs to samsung etc.

While Intel tries to argue antitrust issues are with ARM and in general anyone who manufacturers a ISA chip, instead of by x86 alone.

(22 nm delay was mostly due to apple wanting more gpu power on the mobile chips, no?).

 

[Ferzerp]

Intel would never survive trying to operate at AMD's margins. AMD lives on margins that would kill Intel. The shareholders would revolt. The payoffs and the bribery would stop. The company would fall apart within 4 quarters of AMD style margins. The CEO would be gone and the battle would stop.

As a curious side note, Intel makes its money by providing payoffs such as the ultrabook bribes. Those payoffs dont count against its margins. So what Intel is basically doing, in addition to the obvious crimes (for which they will be found guilty of probably 3-4 years from now) is cooking their books. The concept is simple: Say you make widgets for $100 apiece. You sell them for $300 leaving you a nice wonderful 67% margin. Everything is good, everyone is happy. But what do you do if you cannot sell your widgits for $300 anymore? Easy. You create an UltraWidget standard and hand out a bunch of money to people to encourage them to build systems using your widget (and only your widget) for a lower cost. Now with the total systems costs lowered, you have just inflated the demand for your $300 widgets. So you can keep your prices at $300 per widget and thus your margins remain at 67%. Obviously this is a cut and dry accounting scam worthy of a JP Morgan bankster swine Wall Street Wall of Shame. But this is exactly, exactly what Intel is doing if they do not count these ultrabook bribes against their margins. They would be attempting to cover up falling margins. That would be very bad news for the entire sector! These are the kind of games that result in market moves like what happened in 2008. But that of course is another story.

 

[Lonyo]

Well, since some manufacturers are still planning to use Atom processors in next gen Win 8 tablets, I would say there's an opportunity for AMD, since Atom is horrible.

 

[KompuKare]

Intel would never survive trying to operate at AMD's margins. AMD lives on margins that would kill Intel. The shareholders would revolt. The payoffs and the bribery would stop. The company would fall apart within 4 quarters of AMD style margins. The CEO would be gone and the battle would stop.

As a curious side note, Intel makes its money by providing payoffs such as the ultrabook bribes. Those payoffs dont count against its margins. So what Intel is basically doing, in addition to the obvious crimes (for which they will be found guilty of probably 3-4 years from now) is cooking their books. The concept is simple: Say you make widgets for $100 apiece. You sell them for $300 leaving you a nice wonderful 67% margin. Everything is good, everyone is happy. But what do you do if you cannot sell your widgits for $300 anymore? Easy. You create an UltraWidget standard and hand out a bunch of money to people to encourage them to build systems using your widget (and only your widget) for a lower cost. Now with the total systems costs lowered, you have just inflated the demand for your $300 widgets. So you can keep your prices at $300 per widget and thus your margins remain at 67%. Obviously this is a cut and dry accounting scam worthy of a JP Morgan bankster swine Wall Street Wall of Shame. But this is exactly, exactly what Intel is doing if they do not count these ultrabook bribes against their margins. They would be attempting to cover up falling margins. That would be very bad news for the entire sector! These are the kind of games that result in market moves like what happened in 2008. But that of course is another story.

While I do love a good conspiracy theory, I think you're reading a bit too much into the ultrabook marketing kickback thing though...

BUT having said that, I did wonder how these new 'elite' ultrabooks can get to their mythical $600 price point if Intel sells the ULV CPUs for $200+ (and the + is rather a lot since Intel seems to charge more for their chipsets each generation despite having almost integrated everything into the CPU).

Doesn't leave much left over for SSD, chassis etc. guess the new 'elite' ultrabooks will have to have a cheap cr*p TN panel and other compromises since Intel wants to create a new market, refuses to reduce their 60%+ margins, but expects all the OEMs to cut theirs and is creatively looking at (almost) any ideas to reduce BOM. Seems both Intel and Microsoft love the idea of high volume sellers as long as their margins remain intact. But OEMs might have different ideas and most would probably be glad to see either company go.

Both MS and Intel might eventually get dethroned but not, for example, because ARM and Android / ChromeOS etc. will perform better or be better but rather since ARM players and Google can survive on margins lower than 60%. Or to put the margin thing an other way: if on a $600 ultrabook Intel makes $200, Microsoft makes $50 and everyone makes $10 (disclaimer: made up numbers), it doesn't take a genius to see that Intel and MS are not making too many friends.

For us computer enthusiasts ARM may not hold much interest but then again nobody listens to enthusiast. With Win8 Microsoft seems to positively trying to alienate PC users, so if even Microsoft no longer cares for computer enthusiasts and power users jumping to another OS may not seem such a bad idea in a few years time. That is, as long it's not some other wannabe tablet OS like Ubuntu Unity...

 

[futurefields]

...Intel makes its money by providing payoffs such as the ultrabook bribes.

That's funny. I thought they made their money by making microprocessors etc...

Learned something new on Anandatech today.