AMD Piledriver rumours ... and expert conjecture

[Reynod]

We have had several requests for a sticky on AMD's yet to be released Piledriver architecture ... so here it is.

I want to make a few things clear though.

Post a question relevant to the topic, or information about the topic, or it will be deleted.

Post any negative personal comments about another user ... and they will be deleted.

Post flame baiting comments about the blue, red and green team and they will be deleted.

Enjoy ...

 

[noob2222]

Question: If Intel is a monopoly in the PC industry, then couldn't you argue ARM is an even greater monopoly in the mobile phone industry? You can't have it both ways.

Secondly, the "lesser of two evils" argument requires subjectivity, which automatically makes the argument prone to bias. Tell me, if AMD started to do the same thing, would you buy VIA, just because, performance be damned?

ARM doesn't even manufacture, in fact if you wanted to build a cpu based on ARM, you could (for the rough sum of 1.75M for the liscence). Why do you think there are soo many different ARM vendors, Qualcomm, Nvidia, TI, Samsung, ... Its estimated that each ARM chip sold costs roughly $0.11 each for the liscencing cost.

ARM is pretty much the very opposite of a monopoly as anybody can own a piece and use it however they want. For example, nvidia changed it soo much that its barely even called an ARM chip.

If AMD started breaking the law to the extent that Intel has, I would actually consider using Intel, but until (IF) that happens ...

 

[palladin9479]

Question: If Intel is a monopoly in the PC industry, then couldn't you argue ARM is an even greater monopoly in the mobile phone industry? You can't have it both ways.

Secondly, the "lesser of two evils" argument requires subjectivity, which automatically makes the argument prone to bias. Tell me, if AMD started to do the same thing, would you buy VIA, just because, performance be damned?

Intel maintains a near monopoly in the commodity server market and a crushing majority in the consumer market.

Don't even attempt to talk around that.

The fact that you even tried that tactic shows your not being honest.

Does anyone here seriously think their going to play WoW / CoD / BF3 / SCII / Skyrim on an ARM CPU?

Intel has a near monopoly for exact same reasons that Microsoft has a near monopoly, sheer size of install base.

The white knight defenders are down to historical revisionism as their last bastion of defense.

And yes there was a civil court case between the two, I even linked the documents. The Justice department has a separate section for anti-trust investigations which the FTC brought on via AMD, HP, DELL and others complaints. During this time the FTC launched it's own investigation into Intel's anti-trust practices (separate from the court case) and reported their findings. They issued a remediation to Intel, which is a piece of paper saying to do something or face a trial from the US Government. Intel followed the remediation actions, I linked what they were. You can still go to the Intel page where their offering compensation for having to recompile your code although it's now on a case by case basis.

No matter what FoZ / gamer / js say try to argue, this is history and the documents are freely available to anyone who wants to look. Both the court settlement between AMD - INTEL and the FTC's findings and remediation actions are freely available.

 

[fazers_on_stun]

felt the time for a change, my monitors keep looking back at me attempting to jedi mind trick me...
like I was having visions of a computerized puppeteer running amok via backdoor router scripting and web cams.....


:lol:
😛

Heh, well if your own avatar starts freaking you out, time to switch or lay off the herbs, Herb! 😀..

 

[amuffin]

hi

 

[fazers_on_stun]

I got news: X86 is not the dominant CPU architecture anymore, as ONLY the PC platform uses it. Hard to call Intel a monopoly on that basis alone. Nevermind that NOTHING is stopping AMD from moving to a different CPU architecture and competiting in some other market segment if they so choose.

If we are going to hold every company accountable for every "evil" deed they did, then we won't have anything left to purchase. How many people died because GM wanted to save a few bucks and stuff the gas tank under the back seat? Or how about Ford using a far to narrow wheel base on its SUV's for years on end, despite knowing this posed a rollover risk? Or taking your Coke analogy, guess what used to be a prime ingredient in coca-cola? Thats right: cocaine. Every company has skeletons in the closet, and I for the most part ignore them. I buy based on price/performance. Nothing more, nothing less.

LOL - +10! A little perspective on the nature of corporate evil is sorely needed.. Having your compiler check CPUID is somewhat less serious than puttling melamine in infant formula, lead paint in kids toys, or selling tobacco products with addictive ingredients, but you'd never know it judging from all the Intel-haters here.. Well OK maybe all 3 haters here.. 😛

I wonder if Intel's new ultrabook promotion & advertising campaign will mention specific products or manufacturers by name? If so, I'd bet there will be an exclusivity clause in any agreements signed with those manufacturers, as Intel wouldn't want to promote an untrabook model that is also available with a trinity CPU in it 😛..

 

[fazers_on_stun]

Still wondering how soon we will see Samsung fabricating AMD chips, http://www.tomsitpro.com/articles/amd-samsung-server-opteron-x86,1-159.html

Dunno what that would gain them. According to http://semiaccurate.com/2012/04/03/netronome-is-a-22nm-intel-foundry-customer-too/

[/URL]

 

[jdwii]

Intel has done nothing to amd in like 5+ years lets drop it, It was Amd who was quiet about BD for over 2+ years, it was Amd who thought CMT was better then just making a die shrink of the phenom II and adding 2 more cores. It's Amd who charges to much for their newer Processors. The only thing Amd has right now worth anything is their graphics(why else would you buy a APU from them) which is keeping them alive, If they didn't take over Ati i don't know where they would be. I just know one thing i'm still happy with my Amd 6950 and my x6 clock at 3.9Ghz and i love my current Laptop with a A8 3820 in it.

 

[dogman_1234]

What AMD was trying to do with Bulldozer was to provide an innovative way to gain consumers...and kola.

What ended up happening was the design was not clean and did not perform as expected. AMD got it in the foot. Thanks to the AIT GPU technology and Llano, AMD still has green coming in.

 

[noob2222]

Dunno what that would gain them. According to http://semiaccurate.com/2012/04/03/netronome-is-a-22nm-intel-foundry-customer-too/

[URL=http://imageshack.us/photo/my-images/94/inteltransistornetronom.png/][url]http://img94.imageshack.us/img94/6613/inteltransistornetronom.png[/url][/URL][/quote]
must be an old intel slide because it shows 22nm in 2011. http://semimd.com/blog/2012/03/14/globalfoundries-samsung-tip-finfet-migration-at-14nm/

Main point : my belief is samsung will overtake GF in processes and does full nodes unlike TSMC's half node only. The comparison wasn't for Intel because they won't manufacture for competitors.

 

[noob2222]

Intel has done nothing to amd in like 5+ years lets drop it, It was Amd who was quiet about BD for over 2+ years, it was Amd who thought CMT was better then just making a die shrink of the phenom II and adding 2 more cores. It's Amd who charges to much for their newer Processors. The only thing Amd has right now worth anything is their graphics(why else would you buy a APU from them) which is keeping them alive, If they didn't take over Ati i don't know where they would be. I just know one thing i'm still happy with my Amd 6950 and my x6 clock at 3.9Ghz and i love my current Laptop with a A8 3820 in it.

Intel has never stopped trying, what do you think we have been talking about with the compilers not utilizing AMD chip functionality?

As for proof .. Thu Mar 16 2009 Intel tried to deny AMD the ability to avoid bankrupcy.

http://hothardware.com/News/Intel-And-AMD-Fight-Over-Patent-Rights/
http://www.theinquirer.net/inquirer/news/1051455/intel-amd-dead

That was only 3 years ago, take off the blue colored glasses, Intel has NO INTENTION of playing fair. Look at what they did to Nvidia ... motherboard division - gone overnight.

 

[Chad Boga]

If we are going to hold every company accountable for every "evil" deed they did, then we won't have anything left to purchase. How many people died because GM wanted to save a few bucks and stuff the gas tank under the back seat? Or how about Ford using a far to narrow wheel base on its SUV's for years on end, despite knowing this posed a rollover risk? Or taking your Coke analogy, guess what used to be a prime ingredient in coca-cola? Thats right: cocaine. Every company has skeletons in the closet, and I for the most part ignore them. I buy based on price/performance. Nothing more, nothing less.

What you fail to realise is that the greatest moral tests that exist for all of humanity, is in their choice of CPU purchase.

Intel has never stopped trying, what do you think we have been talking about with the compilers not utilizing AMD chip functionality?

As for proof .. Thu Mar 16 2009 Intel tried to deny AMD the ability to avoid bankrupcy.

http://hothardware.com/News/Intel-And-AMD-Fight-Over-Patent-Rights/
http://www.theinquirer.net/inquirer/news/1051455/intel-amd-dead

That was only 3 years ago, take off the blue colored glasses, Intel has NO INTENTION of playing fair. Look at what they did to Nvidia ... motherboard division - gone overnight.

No company has an intention of playing fair. They play to the letter of the law, or at least to what their legal department's interpretation of that law is.

This emphasis on "fairness" is just the outpourings of the emotionally immature who don't understand how businesses operate or why.

 

[dogman_1234]

Also, what seems 'fair' to one, is a cruel game to another.

 

[amuffin]

sometimes it just happens ?

http://www.passmark.com/forum/showthread.php?t=3656

Passmark does not justify anything. For all I care it is false information.

 

[jimmysmitty]

Intel has done nothing to amd in like 5+ years lets drop it, It was Amd who was quiet about BD for over 2+ years, it was Amd who thought CMT was better then just making a die shrink of the phenom II and adding 2 more cores. It's Amd who charges to much for their newer Processors. The only thing Amd has right now worth anything is their graphics(why else would you buy a APU from them) which is keeping them alive, If they didn't take over Ati i don't know where they would be. I just know one thing i'm still happy with my Amd 6950 and my x6 clock at 3.9Ghz and i love my current Laptop with a A8 3820 in it.

Possibly putting that $5ish billion towards R&D instead of over paying for ATI?

I loved ATI but they did over pay.

Maybe BD would have been released earlier, maybe that FAB in NY would have been done and opened way earlier increasing their FAB capacity.

Whos to say. Or maybe they would be bought up.

must be an old intel slide because it shows 22nm in 2011. http://semimd.com/blog/2012/03/14/globalfoundries-samsung-tip-finfet-migration-at-14nm/

Main point : my belief is samsung will overtake GF in processes and does full nodes unlike TSMC's half node only. The comparison wasn't for Intel because they won't manufacture for competitors.

It shows 22/20nm end of 2011 till beginning of 2013. It seems pretty spot on otherwise since it did take 3.5years to push out a HK/MG from anyone else.

Passmark is crap.

Its not utter crap. Its a good test to see if your overclock is stable, better than Prime95 since it does more active tasks and not just full blown CPU stress.

 

[jimmysmitty]

IBT is better..

Not always. I do like IBT as it takes way less time than Prime95 to see if a CPU is bad and as well seems to always make the CPU run hotter overall than P95, but I had my 2500K stable at 1.28v for 4.5GHz running 10 passes of IBT using VH settings (4GB of RAM). But after a couple of hours of just even web browsing, I would get a 101 BSoD which is CPU related, in my case the VCore was too low. Thats how I found my perfect number at 1.3v and then got it stable with Offset Mode voltage so it idles at lower voltage.

I wonder what AMD uses to test their CPUs before shipping......

 

[amuffin]

Its not utter crap. Its a good test to see if your overclock is stable, better than Prime95 since it does more active tasks and not just full blown CPU stress.

We mean that passmark charts, are crap. Look at this: http://www.cpubenchmark.net/high_end_cpus.html An fx-6200 outperforming all of the older i7-9xx cpus and even the i5-2400???

 

[palladin9479]

We mean that passmark charts, are crap. Look at this: http://www.cpubenchmark.net/high_end_cpus.html An fx-6200 outperforming all of the older i7-9xx cpus and even the i5-2400???

I've mentioned before that bench's only show what their programmed to show. So the question becomes, what instructions are they possibly using that could have the FX-6200 outperform the older 4-core CPUs. I agree that is' mostly a crap benchmark as no desktop program will act that way.

 

[truegenius]

chíkmàa

 

[r0ck3tm@n]

Not always. I do like IBT as it takes way less time than Prime95 to see if a CPU is bad and as well seems to always make the CPU run hotter overall than P95, but I had my 2500K stable at 1.28v for 4.5GHz running 10 passes of IBT using VH settings (4GB of RAM). But after a couple of hours of just even web browsing, I would get a 101 BSoD which is CPU related, in my case the VCore was too low. Thats how I found my perfect number at 1.3v and then got it stable with Offset Mode voltage so it idles at lower voltage.

I wonder what AMD uses to test their CPUs before shipping......

Well, they don't use anything built with the Intel compiler.

 

[jimmysmitty]

Well, they don't use anything built with the Intel compiler.

Didn't mean performance, I meant more for stress testing and certifying they are good CPUs and stable at their designated clock speeds. Intel uses Intel Burnin Test which uses Linpack to stress the CPU.

I just wonder what AMD uses, be it Prime95 or maybe their own in house program which I wouldn;t mind messing with.

 

[amuffin]

 

[Chad Boga]

Well, they don't use anything built with the Intel compiler.

They get better Spec scores with the Intel compiler than they do with GCC or Microsoft.

 

[r0ck3tm@n]

They get better Spec scores with the Intel compiler than they do with GCC or Microsoft.

That's good to hear. I suppose that all the proper machine code is being generated then?

 

[palladin9479]

That's good to hear. I suppose that all the proper machine code is being generated then?

He gave you a very loaded statement.

GCC and MSC compilers can't do code dispatching (yet), meaning their only compiled to one level of code. They don't bias their code on the CPU their running because they only have one set of code and no generic "slow code" to bias to.

That being said, Intel Compiler after 2010 will still use SSE2 on an AMD CPU. If an application is compiled with GCC / MSC and set to SSE2 level, then the Intel compiler will be faster regardless. Intel's compiler puts out the fastest code for Windows x86 period, so when comparing code of the same level then Intel's compiler wins even when it's gimping the non-Intel CPUs. Also SPEC will compiler their testing suite with whatever compiler and optimization levels the customer asks for. Intel's compiler doesn't make Sparcv9 code yet several of the Spec benchmarks can run on Sparc platforms. They will also provide you with the source code to the testing loops, provided you sign an NDA concerning them.

 

[Chad Boga]

He gave you a very loaded statement.

I just gave the TRUTH and then you fleshed it out some more.

BTW, With Larry Ellison now admitting he lied about Itanium's roadmap, does that mean you have sworn off SPARC, what with you being the great moralist and all? :lol: