AMD To Give Up Competing With Intel On x86?, CPU Prices Already Shooting Up? Options

[nobody813]

Yep, Bulldozer did use an automated process, which meant millions more transistors were used (can't remember the exact percentage)

[juggtron]

Aren't all the current intel processors still based to a degree on the old P6 designs? With a bit of netbust thrown in from nehalem onwards? And I cant believe that they wouldn't be using some form of automation in their design. As for the whole 'just die shrink it' I remember reading somewhere that you pretty much have to redesign the whole processor for the new process - the ideas usually stay the same with a little bit of tweaking.

there are probably some good idea's tied up in the BZ architecture, maybe if AMD keeps phenom/athlon alive in the budget segment we'll see some of the ideas filter through.

[nobody813]

All processors from Pentium M onwards are based off the Pentium III architecture (Pentium M had some aspects from Netburst I believe)

[Jeruselem]

AMD missed the boat with mobile platform too. nVidia are already putting out Tegras into new tablet designs.

[meowkitty]

If Bulldozer was decent people wouldn't be as negative. People feel burned (For some strange reason??) so they don't really care anymore.

With that being said i read a few weeks ago that AMD went for a more automated approach to chip design which more or less slowly killed their chips. So it looks like a decision made years ago is having a major impact on the copanies performance.

yea I read that as well, something about the processors not being hand drawn to save on R&D costs (how one hand designs a CPU is unfathomable to me though)

I too read that. if it wasn't linked in the main AMD thread it was a link at a link location.
it seemed to comedown to a belief that an automated process based on simultion could produce a better chip, but as I recall, the simulation could only handle a small section of the chip and only at sub 50Hz of simulated clock speed. in the end buldozer could not capitalise on aquired human knowledge and also fell victim of a foible of the simulations that they make them a little bigger and a little less eligant in their execution.
Quite a professional writeup and amaizingly insiteful of the chip creation process.

it may have been this article, death by a thousand cut.
http://semiaccurate.com/2011/10/17/why-did...zer-underwhelm/

it's natural to be disapointed when you are given info that a new item will be better and you want it to be better. it may just take a few days to get over it and get perspective.
I know I was and I am really holding our for a revision or a Phenom II sytle step up to goodness. but I am a couple years from my next build so may end up with an AMD APU with hybrid crossfire by then.

the 6100 is $10 less than the 1100T and under $200 on PCCase. that's not too shabby, in isolation.

[nobody813]

That article was a good read back when I read it a while ago. Still is good now

This post has been edited by nobody813: Dec 1 2011, 12:52 PM

[mark84]

If Bulldozer was decent people wouldn't be as negative. People feel burned (For some strange reason??) so they don't really care anymore.

With that being said i read a few weeks ago that AMD went for a more automated approach to chip design which more or less slowly killed their chips. So it looks like a decision made years ago is having a major impact on the copanies performance.

yea I read that as well, something about the processors not being hand drawn to save on R&D costs (how one hand designs a CPU is unfathomable to me though)

I too read that. if it wasn't linked in the main AMD thread it was a link at a link location.

You mean this post?
http://forums.atomicmpc.com.au/index.php?s...st&p=918165

Specifically this:
http://www.xbitlabs.com/news/cpu/display/2...zer_Fiasco.html

Cliff A. Maier, an AMD engineer who left the company several years ago, the chip designer decided to abandon practice of hand-crafting various performance-critical parts of its chips and rely completely on automatic tools. While usage of tools that automatically implement certain technologies into silicon speeds up the design process, they cannot ensure maximum performance and efficiency.
Automated Design = 20% Bigger, 20% Slower

"The management decided there should be such cross-engineering [between AMD and ATI teams within the company] ,which meant we had to stop hand-crafting our CPU designs and switch to an SoC design style. This results in giving up a lot of performance, chip area, and efficiency. The reason DEC Alphas were always much faster than anything else is they designed each transistor by hand. Intel and AMD had always done so at least for the critical parts of the chip. That changed before I left - they started to rely on synthesis tools, automatic place and route tools, etc.," said Mr. Maier in a forum post noticed by Insideris.com web-site.

Apparently, automatically-generated designs are 20% bigger and 20% slower than hand-crafted designs, which results in increased transistor count, die space, cost and power efficiency.

[meowkitty]

Thank you, Mark84.

[nobody813]

AMD Still Committed To x86 - But Not In High End Desktop

[A Hitman]

Wait, nothing in that statement says to me that they are not continue with high end?
All it says to me is that they will expand into the mobile market.

[nobody813]

All I believe that is going to happen is all "high end" AMD processors will only compete with mainstream Intel chips. That's my opinion any way :-P

[bowiee]

Don't look now but the sky may actually be falling. If we lose competition in the main stream processor area, then we will be at the mercy of Intel who will no doubt grow fat and complacent (dare I say much more then they are now). Taken to its logical conclusion, having only one major player, desktop computing will become a niche market, when compared to mobile computing devices. And we gamers may be stuck with only consoles to play games, /shudder.

I mean lets face it most people these days use their mobile devices (including laptops) much more than a desktop (if they even own a desktop). With that in mind desktop computing must by weight of numbers of mobile devices be pushed to near extinction, Of course as we all know this would be bad for not only us but computing in general as CPU growth would then stagnate.

After all most mobile devices have very little processing power when compared to desktops, and yet most people seem to be happy to use a slower portable device. I use my laptop so I can access the net ect while out and about, but in terms of computing power compared to my desktop it just makes using it bearable.

This is how I as a gamer feel about mobile computing, but the average user out there does not care about computing power as long as the device of choice runs "angry birds". I hope I am wrong but I fear gaming computers and powerful desktop computing in general, will become such a small market as to not matter any more.

EDIT: As an after thought walk into any electrical store these days and the majority of computing on show is laptops or mobile devices of some type, desktops are rarely seen any more. But a few years ago they were prolific.

This post has been edited by bowiee: Dec 1 2011, 03:11 PM

[.:Cyb3rGlitch:.]

The sky isn't falling, people are moving on. Desktops are hold hat. They're bulky, expensive, inconsistent and unreliable.

[bowiee]

No its not true ! /Hugs my gaming computer. Yes the sky is falling for lovers of "Maximum Power Computing"

[mark84]

http://www.zdnetasia.com/amd-taps-servers-...th-62303056.htm

Greg Poole, director of field applications engineering at AMD, told ZDNet Asia in an interview Tuesday that making chips for desktops and laptops is its "bread and butter" business.

"If you look at the client computing business, which could be considered as our 'bread and butter', I'd say we have substantially more revenue generated from the client business including desktops and notebooks," he said.

However, this area is "largely a commodity business now" and is difficult for the company to hit the margins it wants to, the director added.

This is why the chipmaker will be focusing on its server-based products as these promises higher returns, Poole stated.

"The server space is extremely important to AMD and we are investing heavily to be successful...[With server processors,] you're providing mission-critical applications and solving the big problems of the world so you need more robust and reliable solutions. We know where our future needs to grow."

Poole's point was backed up by the company's third-quarter financial results. In the report released on Oct. 27, AMD stated that its computing solutions segment--which includes server processors--reached US$1.3 billion and was partially driven by "double-digit growth in server processor revenue driven by significantly higher ASP (average selling price)".

The Austin-based chipmaker will have to arrest the drop in server market share in the second quarter of 2011 though. According to IDC's findings, AMD slipped 0.6 percent to 5.5 percent in the global server and workstation processors segment during this period. Its main rival Intel, on the other hand, rose 0.6 percent to reach 94.5 percent, according to the ZDNet Asia report.

Quizzed how the company plans to increase its market share, Poole declined to comment saying he is not equipped to answer as he's "the engineering guy".

Working toward cooler servers
The executive did point out that enterprises' server requirements are evolving. In the past, the main consideration for server chips was "performance, performance, performance", but this has changed to include total cost of ownership (TCO) as well, he stated.

In an effort to help IT managers decrease their overall datacenter TCO and justify expenditure, Poole pointed out that AMD's latest Opteron server processor offers 16 cores and include more compute capability in a smaller footprint which, in turn, leads to lower power consumption in one's data center.

The chip also has power saving features such as thermal design power (TDP) capping that allows IT managers to set the parameters for power consumption and core power gating, said the director. This, he noted, would also help companies in saving energy consumption and costs.

If desktops are their 'bread and butter' and with a move to focus more on server end gear, wouldn't that have the side benefit of the tech trickling down to high end desktops?

[bowiee]

Power saving server chips are not my idea of a chip for gaming.

[.:Cyb3rGlitch:.]

No its not true ! /Hugs my gaming computer. Yes the sky is falling for lovers of "Maximum Power Computing"

Desktops were never maximum power computing anyway. :P

[sora3]

No its not true ! /Hugs my gaming computer. Yes the sky is falling for lovers of "Maximum Power Computing"

Desktops were never maximum power computing anyway. :P

So do you run a multi-server with at least 128 CPUs with at least 16TB worth of RAM? If not, you're not trying hard enough. :P

[SquallStrife]

If desktops are their 'bread and butter' and with a move to focus more on server end gear, wouldn't that have the side benefit of the tech trickling down to high end desktops?

Server chips usually lag behind the bleeding edge desktop parts by a generation.

Sandy Bridge is on the shelves now, but blades purchased today are probably still Nehalem based.

This post has been edited by SquallStrife: Dec 1 2011, 07:26 PM

[juggtron]

I haven't noticed any Core i3's in the ivy bridge slides - maybe Intel is 'letting' AMD have the budget segment to ward off any pesky anti-trust suits? Intel would get the best of both worlds - AMD's name gets dragged further through the mud and they get even more profit from the (higher volume & better margin) value midrange segment. I'm starting to wonder maybe it was a mistake to drop the ATI brand...