[Austin Statesman] AMD sees a way forward (with new Zen design)

[NTMBK]

Let's try to be somewhat realistic... Zen will be used to power the real life Starship Enterprise that AMD will be building in 2018.

Well it's not like PC sales are recovering any time soon on planet Earth, might as well try to flog those excess Trinity chips to the Klingons.

 

[AtenRa]

Qapla

 

[CHADBOGA]

Yeah like the blue brigade foretelling the end of times for AMD, ARM, Apple for the better part of this decade?

LOL

Find one quote on these forums by anyone being serious that ARM or APPLE were facing end times?

You know you won't be able to, yet still you speak absolute rot.

 

[R0H1T]

LOL

Find one quote on these forums by anyone being serious that ARM or APPLE were facing end times?

You know you won't be able to, yet still you speak absolute rot.

Does that include servers as well?

 

[myocardia]

Find one quote on these forums by anyone being serious that ARM or APPLE were facing end times?

You forgot to include the caveat of Apple not having problems after the first iPhone released. They were in fairly serious financial turmoil, until they invented the smartphone, IIRC.

edit: Losing 700M/quarter in 1996, for instance- http://www.nytimes.com/1996/03/28/business/apple-expects-it-will-lose-700-million.html

 

[CHADBOGA]

Does that include servers as well?

Whether ARM makes significant headway in Servers or not in the next few years, is a completely different matter to predicting the end of the company.

 

[dark zero]

ARM might get eventually bought by Intel. Is the only way to maintain the monopoly and it will be legal in the US due the help of Intel to the military forces.

AMD has 2 paths. Die or become Chinese. Only those ways are posible. Despite their efforts, I doubt that they will get something from their releases...

 

[Boze]

You forgot to include the caveat of Apple not having problems after the first iPhone released. They were in fairly serious financial turmoil, until they invented the smartphone, IIRC.

edit: Losing 700M/quarter in 1996, for instance- http://www.nytimes.com/1996/03/28/business/apple-expects-it-will-lose-700-million.html

No. They were in serious financial trouble until Steve Jobs returned. He quite literally did save the company, and did turn everything around. When he came back and pushed the new iMacs, Apple started gathering a lot of steam. When OS X was released on the desktop (server version had already existed for about a year IIRC), Apple really picked up the pace.

Of course, as you are no doubt aware, none of this really had anything to do with Intel. This was mostly EDIT: crappy (sorry langauge Nazis) management after Jobs was forced out of the company.

 

[Boze]

ARM might get eventually bought by Intel. Is the only way to maintain the monopoly and it will be legal in the US due the help of Intel to the military forces.

AMD has 2 paths. Die or become Chinese. Only those ways are posible. Despite their efforts, I doubt that they will get something from their releases...

If Microsoft's CEO really wanted to shake things up, he ought to buy AMD. Microsoft's Surface Pro 3 is a clear indication that Microsoft can make hardware products that are attractive to business & enterprise.

Microsoft had $20 billion in the past 12 months of profit. Revenue was nearly $100 billion. Dump some of that cash into hiring and expanding the R&D team of AMD and let's see some real competition to Apple, ARM, and Intel.

Plus, it would have the side effect of Microsoft having their own CPU / GPU hardware unit for future endeavors. If I were Microsoft, I'd negotiate a 5-8 year deal for hardware from Intel for my Surface Pro / Surface line, then buy AMD, dump a metric eff-ton into development of low-to-mid power parts, and take on both Apple and Intel at the same time.

But hey, I'm a crazy kind of guy.

 

[VirtualLarry]

Plus, it would have the side effect of Microsoft having their own CPU / GPU hardware unit for future endeavors. If I were Microsoft, I'd negotiate a 5-8 year deal for hardware from Intel for my Surface Pro / Surface line, then buy AMD, dump a metric eff-ton into development of low-to-mid power parts, and take on both Apple and Intel at the same time.

But hey, I'm a crazy kind of guy.

You really think MS has the kind of "Vision" to pull that off? I don't.

 

[ShintaiDK]

If Microsoft's CEO really wanted to shake things up, he ought to buy AMD. Microsoft's Surface Pro 3 is a clear indication that Microsoft can make hardware products that are attractive to business & enterprise.

Microsoft had $20 billion in the past 12 months of profit. Revenue was nearly $100 billion. Dump some of that cash into hiring and expanding the R&D team of AMD and let's see some real competition to Apple, ARM, and Intel.

Plus, it would have the side effect of Microsoft having their own CPU / GPU hardware unit for future endeavors. If I were Microsoft, I'd negotiate a 5-8 year deal for hardware from Intel for my Surface Pro / Surface line, then buy AMD, dump a metric eff-ton into development of low-to-mid power parts, and take on both Apple and Intel at the same time.

But hey, I'm a crazy kind of guy.

The result would first happen in ~2020 and forward. Assuming you essentially gave unlimited funds today.

Also ask yourself about the ROI. Just because someone got money doesn't mean they should outright waste it.

 

[DrMrLordX]

T
I even don't expect a big jump in Cinebench performance for Zen, as it already is rather cache friendly and raw FP throughput of XV is good. Of course a lower power consumption at iso clock frequency would allow for higher scores. Other FP heavy benchmarks might improve more than CB due to likely better data flow handling.

Two things to keep in mind when comparing Zen to XV:

1). XV has half the FP execution resources per module compared to Zen's cores:

http://fudzilla.com/news/processors/37639-new-leaked-amd-zen-slide

2). Zen will be moving to an inclusive cache architecture, presumably with much better cache performance than that of XV:

http://www.extremetech.com/gaming/204523-new-leak-hints-at-amd-zens-architecture-organization

Those two changes alone should account for a great deal of the projected 40% improvement in IPC moving from XV to Zen, at least for FP workloads anyway.

 

[Boze]

You really think MS has the kind of "Vision" to pull that off? I don't.

What I would "like to see" and what I get to see are a continual source of disappointment for me in daily life...

It might be that I hold people to too high a standard... or maybe its because most people don't strive for excellence in all things.

 

[Boze]

The result would first happen in ~2020 and forward. Assuming you essentially gave unlimited funds today.

Also ask yourself about the ROI. Just because someone got money doesn't mean they should outright waste it.

Then basically your argument is that AMD should just give up... after all, if $20 billion a year invested couldn't bring AMD back from the dead, then nothing could.

 

[ShintaiDK]

Then basically your argument is that AMD should just give up... after all, if $20 billion a year invested couldn't bring AMD back from the dead, then nothing could.

There are several parts you need to look at first.
What contracts, liabilities and obligations does AMD contain. (The dreaded WSA for one).
How would MS get the money back in a reasonable timeframe.
What guarantee is there for a competitive product to actually materialize.
Being fabless would make MS have to directly compete with Apple. A company that is willing to put 100-150$ chips into phones and pay 15000$ per wafer.

In short, how is the ROI. You may end up with a better deal for MS to buy bean cans.

The best thing AMD can do is to dump its disasterous CPU division and focus on graphics design and IP licensing. But that part may be ruined too by their Zen dreams.

 

[Fjodor2001]

Just like they got ruined by their dreams of there ever being an APU in the PS4 and XBONE. We all knew that would never happen either, right?

 

[frozentundra123456]

Just like they got ruined by their dreams of there ever being an APU in the PS4 and XBONE.

Yea, those console APUs are a financial bonanza. Just look at the huge profits AMD is making and the way the stock is skyrocketing.

 

[Phynaz]

Just like they got ruined by their dreams of there ever being an APU in the PS4 and XBONE. We all knew that would never happen either, right?

Which you're posting in a Zen thread because?

 

[Tsavo]

If Microsoft's CEO really wanted to shake things up, he ought to buy AMD. Microsoft's Surface Pro 3 is a clear indication that Microsoft can make hardware products that are attractive to business & enterprise.

Microsoft had $20 billion in the past 12 months of profit. Revenue was nearly $100 billion. Dump some of that cash into hiring and expanding the R&D team of AMD and let's see some real competition to Apple, ARM, and Intel.

Plus, it would have the side effect of Microsoft having their own CPU / GPU hardware unit for future endeavors. If I were Microsoft, I'd negotiate a 5-8 year deal for hardware from Intel for my Surface Pro / Surface line, then buy AMD, dump a metric eff-ton into development of low-to-mid power parts, and take on both Apple and Intel at the same time.

But hey, I'm a crazy kind of guy.

All rather naive points.

MS might be able to design a tablet that no one really buys, but that's nothing compared to making CPU's and GPU's.

Don't ignore the 700-lb rotting ape in the room: AMD's x86 license agreement with Intel evaporates the moment they sell the company, so it only has value to AMD as they are today. The roll of toilet paper in your aunt's bathroom is worth more on the open market than AMD's CPU division.

The only way Intel will let someone get that agreement is if Uncle Sam forces them to...fat chance on that. t

So, poof, there goes your dream of taking on ARM and Intel.

Take on Apple? Get real. Nearly everything MS has done in mobile has been a disaster, and this will not change. Their only bright spot is the Surface Pro, but it's a niche product. Also, good luck running one of those with an AMD CPU in it; you'd need to carry it around in a hand cart.

Why would you want to compete in the low-mid end of the compute device market? No one makes any money there.

 

[Fjodor2001]

Which you're posting in a Zen thread because?

It's just one similar example of ShintaiDK's previous dead certain doom and gloom predictions about AMD that proved to be incorrect.

 

[Dresdenboy]

Two things to keep in mind when comparing Zen to XV:

1). XV has half the FP execution resources per module compared to Zen's cores:

http://fudzilla.com/news/processors/37639-new-leaked-amd-zen-slide

2). Zen will be moving to an inclusive cache architecture, presumably with much better cache performance than that of XV:

http://www.extremetech.com/gaming/204523-new-leak-hints-at-amd-zens-architecture-organization

Those two changes alone should account for a great deal of the projected 40% improvement in IPC moving from XV to Zen, at least for FP workloads anyway.

These slides are fake and made their rounds earlier this year.

We still don't know much about Zen's FPU.

Compared to SR the XV FPU has faster division and square root (1/2 latency) and also other nice improvements like move elimination. As AMD likely continues to use a separate FPU, and given the complexity and effort which went into it, they might widen and update the current one and put the human ressources into improving power efficiency, adaptivity, and further optimizing common use cases. As long as it's not shared between cores but just between 2 SMT threads, the utilization will be lower than in a CMT module, where the SMT FPU kept up with the pace except for not enough 256b execution ressources.

Edit: Some examples of the per clock performance of AMD's SMT FPU:

AIDA FPU VP8:

http://www.planet3dnow.de/cms/18564...vergleich-der-architekturen/subpage-aida-fpu/

CB 11.5 ST:

CB 15 ST:

http://www.planet3dnow.de/cms/18564...er-architekturen/subpage-rendering-cinebench/

 

[dark zero]

Just like they got ruined by their dreams of there ever being an APU in the PS4 and XBONE. We all knew that would never happen either, right?

That is a mistake that never might happen again. (/post)

AMD is so awful that games are subpar and are a dissaster. Even Intel Skylake GPU are way stronger that those crappy consoles GPU according to Intel users.

Also... MS current generation was a dissaster thanks to AMD.stupidity. Maybe they will leave permanently the console division after the failure of Xbox One.

Sony wants to reduce.costs and they might move to 12 ARM core or a 8 Atom core chip in order to prepare the next generation. Despite AMD did well here (selling millions to get some millions dollars), still their efforts weren't enough. Sony will move from AMD to find better prices. Intel can easily sells their Atoms at subsided prices.

And if AMD dies sooner than expecting, Nintendo will have an horrible time since they won't have a real sucessor to.counter Ps4 and Xbox One. If not, Nintendo must know that AMD will die sooner than expecting, so is better to prepare the replacement of that short lived console.

In few words, AMD is dooming the gaming market hard.

 

[RussianSensation]

AMD is so awful that games are subpar and are a dissaster. Even Intel Skylake GPU are way stronger that those crappy consoles GPU according to Intel users.

Facepalm. You have no clue what you are talking about. Literally ever since you've joined the forum, what you need to do is learn, read, learn, research, learn more, reflect, read, learn. Instead, you post the most unsubstantiated drivel with the confidence of a senior technical member, yet fail to substantiate any of your opinions. How old are you 15?

Let's actually look at how ludicrous your statements are:

Technical analysis
The GPU in the Xbox One is about as fast as an HD7790, and that performance level is very close to the GTX750. The GPU in the PS4 is at least as fast as an R7 265, which is faster than the GTX750Ti by 16%. Since the PS4's GPU also has custom features such as 8 ACE units and 176GB/sec memory bandwidth, along with an extremely fast inter-connect between the CPU and GPU, the actual maximum GPU performance lies somewhere between an HD7850 and the HD7870. That level of GPU performance is miles away from any reasonably affordable Skylake iGPU.

Let's talk about 3 specific custom areas of PS4's GPU that ensure that overall package is even better than the HD7850:

1) There is a separate bus from the CPU to the GPU that allows it to read directly from system memory or write directly to system memory, bypassing its own L1 and L2 caches. As a result, if the data that's being passed back and forth between CPU and GPU is small, you don't have issues with synchronization between them anymore. PS4 can pass almost 20 gigabytes (20GB/sec) a second down that bus. That's faster than PCIe 3.0 x16.

2) Next, to support the case where you want to use the GPU L2 cache simultaneously for both graphics processing and asynchronous compute, PS4's custom GPU has tags of the cache lines, aka 'volatile' bit. You can then selectively mark all accesses by compute as 'volatile,' and when it's time for compute to read from system memory, it can invalidate, selectively, the lines it uses in the L2. When it comes time to write back the results, it can write back selectively the lines that it uses. This innovation allows compute to use the GPU L2 cache and perform the required operations without significantly impacting the graphics operations going on at the same time -- in other words, it radically reduces the overhead of running compute and graphics together on the GPU.

Considering NV's and Intel GPUs of 2013 were vastly inferior to GCN 1.0/1.1 in compute, this means as far as long-term longevity goes, they weren't even on the map in 2013 when it comes to using Asynchronous Compute + graphics.

3) "The original AMD GCN architecture allowed for one source of graphics commands, and two sources of compute commands. For PS4, Sony has worked with AMD to increase the limit to 64 sources of compute commands (8 ACE engines) -- the idea is if you have some asynchronous compute you want to perform, you put commands in one of these 64 queues, and then there are multiple levels of arbitration in the hardware to determine what runs, how it runs, and when it runs, alongside the graphics that's in the system."

http://gamasutra.com/view/feature/191007/inside_the_playstation_4_with_mark_.php?page=2

In other words, when it comes to pure graphics throughout, PS4's GPU with 32 ROPs, 72 TMUs, 1024 stream processors over 176GB/sec memory bandwidth bus, 8ACE engines, volatile compute, and direct 20GB/sec bus to the CPU would wipe the floor with any Skylake iGPU in gaming performance.

Considering neither ARM, nor Intel, nor NV could provide an affordable quad-core or 8-core CPU that could match 1.6-1.75Ghz Jaguar APU cores, even if the Jaguar APU is slow relative to an i5 2500K or similar, it was hands down THE best option given the budgets MS/Sony had to work with.

Financial situation of MS and Sony before PS4/XB1's launch

Report: Microsoft's Xbox division has lost nearly $3 billion in 10 years

July 2015, IGN reported that MS posted a $2.1B loss.

Sony's Game business has been nothing short of abysmal prior to PS4.

I am not going to do more research on 2011-2013 years but in October 2014, IGN reported that Sony overall posted $1.2B loss.

In other words, Xbox 360 and PS3 were utter failures from a business perspective. That means neither Sony nor MS could afford or could sustain a strategy of selling a console which bill of materials substantially exceeded its sale price.

Tying Xbox 1/PS4 BOM to launch MSRP

"The bill of materials (BOM) for the PlayStation 4 amounts to $372. When the manufacturing expense is added in, the cost increases to $381. This comes in $18 lower than the $399 retail price of the console.

When other expenses are tallied, Sony initially will still take a loss on each console sold. But the relatively low BOM of the PlayStation 4 will allow the company to break even or attain profitability in the future as the hardware costs undergo normal declines."
http://press.ihs.com/press-release/...-breakeven-point-playstation-4-hardware-costs

"Microsoft may be spending slightly more to build and distribute the Xbox One than it's taking in, according to a new report from IHS iSuppli.
The research firm said Wednesday that its preliminary teardown of the new game console and calculation of its bill of materials (BOM) put the cost of components in the Xbox One at $457. Tacking on an estimated $14 for manufacturing puts the system's total estimated cost to build at $471, according to IHS, just $28 less than the $499 Microsoft and its partners are charging for the next-gen console at retail."

Source

Summary

Taking all of that information into account, to this date not a single person online, with vastly superior technical knowledge and all the extensive research done compared to you could still come up with a superior PS4 console than Sony did. Knowing the BOM of PS4, it's literally impossible to have made a more powerful console. Not surprising, considering hundreds of engineers have worked on the design with the top firms in the semi-conductor field for years prior to PS4's launch. You think you are smarter than even 1 of those people?

MS could have made a more powerful console than Xbox One but they decided to priorities on other features.

Technically speaking, there were very few choices of a superior GPU for Sony/MS around mid-2013 when they would have needed to finalize their design at the latest. HD7970M/8970M was actually faster than Nvidia's 680MX or 770M. Sony/MS would have needed to get a deal of a lifetime on the 780M to be able to fit that much horsepower in a TDP limited/constrained console. Do you care to take a guess how much a 780M style binned GPU cost in the summer of 2013?

Also... MS current generation was a dissaster thanks to AMD.stupidity. Maybe they will leave permanently the console division after the failure of Xbox One.

AMD had nothing to do with MS's decisions such as prioritizing Kinect over GDDR5 and a faster GPU which forced MS to compromise on the GPU and go with a weaker embedded ram and DDR3. MS could have easily incorporated an APU at least as good as PS4's had they ditched Kinect from the start.

AMD had nothing to do with MS's decision of always on connection to play games, non-upgradeable internal HDD, blocking used games, no rechargeable lithium batteries in the controller, external PSU, VCR styling, $499 MSRP with useless Kinect, etc.

MS's failure with Xbox 1 is 100% on MS, no one else. AMD simply provides what they were contracted to provide but it's ultimately MS that determines what their priorities are and what they are willing to pay for a given level of CPU+GPU performance.

Sony wants to reduce.costs and they might move to 12 ARM core or a 8 Atom core chip in order to prepare the next generation.

Once again showing that you have no idea what you are talking about. Switching to a new CPU architecture mid-cycle in the console generation breaks BC on existing consoles and it breaks future x86 BC on PS4/XB2. The fact that you keep posting this nonsense either tells me you are 15 or less or are just trolling.

Despite AMD did well here (selling millions to get some millions dollars), still their efforts weren't enough. Sony will move from AMD to find better prices. Intel can easily sells their Atoms at subsided prices.

Right, like Intel offered superior processors to MS/Sony in 2013....oh wait they didn't. Why is that? Because Intel aims for >60% gross margins. They have 0 interest in consoles and their budget i3s aren't a good fit a for a console that allocates 1-2 cores for the OS and background functions alone.

And if AMD dies sooner than expecting, Nintendo will have an horrible time since they won't have a real sucessor to.counter Ps4 and Xbox One. If not, Nintendo must know that AMD will die sooner than expecting, so is better to prepare the replacement of that short lived console.

Last quarter AMD had 829M of Cash & Equivalents, that's higher than in the previous 4 consecutive quarters preceding Q2 2015. In fact, as of Q2 2014, AMD only had 503M of C&E.
https://www.google.com/finance?q=NASDAQ:AMD&fstype=ii&ei=YdUJVsCWN4XHmAGh4rOABQ

Chances are, AMD is not what's going to make or break the NX, but Nintendo's strategies and priorities of the new NX eco-system.

Again, even if AMD were to fail (something I've heard for 10 years+ :sneaky, there are certainly contracts in place that will allow for Sony, MS and Nintendo to fabricate the chips directly at TSMC, GloFo or Samsung, etc. If AMD were to fail, it means it has breached its own contract by being unable to provide the product per the Supply Agreement.

You honestly think the lawyers at Sony, MS and Nintendo are 15 year old kids that don't plan for contingencies in their contracts? :hmm:

In few words, AMD is dooming the gaming market hard.

More clueless drivel. Your reputation on AT is already 0, but keep digging into the negatives with 0 facts.

"In an interview with Eurogamer, co-founder of CD Projekt Red, Marcin Iwinski, explained to them...

If the consoles are not involved there is no Witcher 3 as it is, we can lay it out that simply. We just cannot afford it, because consoles allow us to go higher in terms of the possible or achievable sales; have a higher budget for the game, and invest it all into developing this huge, gigantic world."
http://www.cinemablend.com/games/Witcher-3-Wouldn-t-Exist-Consoles-72064.html

My recommendation for you is to start learning to back up your arguments with data/facts/some support or otherwise, it's just baseless opinions that sound like childish vitriol against AMD (or insert any other firm you decide to bash like your ludicrous arguments of how Intel's APUs will soon crush AMD/NV discrete graphics...ya sure!).

 

[CHADBOGA]

Facepalm. You have no clue what you are talking about. Literally ever since you've joined the forum, what you need to do is learn, read, learn, research, learn more, reflect, read, learn. Instead, you post the most unsubstantiated drivel with the confidence of a senior technical member, yet fail to substantiate any of your opinions. How old are you 15?

You have spent a lot of time arguing with someone who at best must be conducting some kind of experiment to see what reactions he gets when he bags Company A, compared to Company B or Company C.

Now the above is what I reckon is the best that could be said for your jousting partner, what do you think the worst is? :hmm:

 

[Dresdenboy]

@RussianSensation: Kudos for that lengthy answer. Make it a seeking alpha article to earn some money with it.

@CHADBOGA: This is what I assume about many guys here.