AMD CPU speculation... and expert conjecture

[de5_Roy]

earlier i said am1 platform is the most interesting thing amd brought out this year...or something like that. but this type of motherboard is awesome...almost:
ECS Unveils the A78F2-TI Thin Mini-ITX Motherboard
http://www.techpowerup.com/199223/ecs-unveils-the-a78f2-ti-thin-mini-itx-motherboard.html

socket fm2+, no pcie x16 slot, dc power input, so-dimm slots instead of full size ones \o/. all it needs is a bit more vrm (4+1, with hs), 2-4 more sata ports, no mpcie, more usb 3.0 ports and 2 sata power input...? since carrizo is rumored to have lower tdp and compatible with fm2+ socket, it should be a drop-in upgrade.

AMD Athlon 5150 and 5350 APUs are available in US
http://www.cpu-world.com/news_2014/2014032401_AMD_Athlon_5150_and_5350_APUs_are_available_in_US.html
seems a little overpriced.

 

[sarinaide]

APU's taking over the worlds. Sure not a sentiment for high end junkies but I am still wholly convinced a Athlon FM2 with a R7 260/265X represents really good value for money and a sound gaming system experience, even a game like BF4 should be very playable with some candy's

 

[Cazalan]

Man how much more stuff is Facebook going to ruin. The one piece of hardware I was actually really looking forward to, will now be spying on my every click.

http://www.brightsideofnews.com/news/2014/3/25/facebook-to-acquire-rift-maker2c-oculus-vr-for-242-billion.aspx

 

[blackkstar]

@gamerk,

I find it far more likely that DX12 is a knee-jerk reaction to:

1. SteamOS and Valve doing what they can to pro-actively reduce reliance on Windows
2. AMD releasing Mantle, which has potential to go cross platform.
3. EA is a major part of Mantle. Origin is developed in QT4 which is cross-platform, which to me hints that EA is keeping that door open
4. MS has lost power. Do you not find it odd that game developers went to someone besides MS to come up with an answer to a lousy graphics API when DX is the standard?
5. Game developers are turned off by Metro and Windows 8 and feel the need to no longer depend on MS

DX12 is a reaction to "Mantel" as you love to call it. It is also a reaction to trying to maintain vendor lock in for gaming PCs. It is a reaction to keep game developers in Windows.

One thing you more than likely love, because you seem to be so resistant to any form of change that you'd go as far as to derive insane conspiracy theories with absolutely zero evidence to continue justifying your beliefs.

You are in the wrong field if you are so opposed to changes. Technology is rapidly evolving. You remind me of a crotchety old man who whines about every new change. "OMG BLACKS CAN VOTE!!! THE WORLD IS OVER", "OMG GAYS CAN MARRY SOCIETY IS FINISHED!!!"

I just imagine you on some Pentium G running DX12 on an Nvidia card yelling at kids with fast CPUs playing games with Mantle on Linux going "YOU RUINED EVERYTHING I HATE YOU DAMN KIDS"

No, its not. DX12 has been kicking for YEARS. That's how versioned development works: You have one team working LTS, and another starting up the next version as soon as the previous one is finished. Hence why NVIDIA confirmed they were working on the specifications with MSFT four years ago.

As for your individual points: Tell me when Linux market share actually gets past 2% on Desktops.

http://www.pcper.com/reviews/Graphics-Cards/NVIDIA-Talks-DX12-DX11-Efficiency-Improvements

Since Juanrga likes AMD sponsored charts:

Um, you are aware that Star Swarm dynamically controls how many objects are on a screen based on performance, right? Did you not watch the presentation where Oxide is talking about how Mantle allows them to throw a ton of stuff on the screen? The first marketing slide is basically the equivalent of AMD posting a benchmark where GTX 780 is running PhysX on the rendering GPU and then going "WE R SO FAST XDDD" The two don't provide the same game experience, and it's comparing apples to oranges.

The Mantle experience differs completely from the DirectX experience. Oxide is using Mantle to provide a more unique gaming experience from DX, while EA/DICE are using Mantle to do nothing but increase performance. Which makes sense when you consider EA is money grubbing and making their games available to more people means more money. But it seems like there's a big perception that Mantle's only purpose is to make games more available to people with lower end systems. It is a problem Mantle helps solve, but it's not the only thing Mantle is capable of. It can definitely be used to just throw more things at the CPU to provide a better experience, and that's exactly what Star Swarm is doing. But it's pretty difficult to put "we played Star Swarm with Mantle and there were a ton more units on screen" into a bar graph, and it's pretty easy to leave that out.

The second graph is a 3 fps difference. In a marketing slide. Meaning that Nvidia is doing everything it can to grasp at straws and seem relevant.

Also, we have no idea what CPUs were used for those market slides. If Nvidia used a super high end Intel CPU, of course Mantle won't matter.

But it's a marketing slide. Doing something like benchmarking with 4960X, not telling anyone, and then going "look the new Nvidia driver is just as good as Mantle!" and implying that's with every CPU is the goal of this market slide.

And look, it's working like a charm. I don't think I've ever seen people get so excited about 3fps.

Critical thinking can take you far and it can really save you from embarrassment. :^)

However, I have a good feeling about your response, and I personally (and I'm quite sure others agree with me) trust Nvidia marketing the least out of AMD, Intel, and Nvidia. All marketing slides tend to exaggerate, but Nvidia does have the whole "our GPU is in great shape! Look at it! Oh ignore the wood screws it really works!" thing. And Intel has the whole "look at how great this game plays on the new Intel GPU!! (cuts to VLC video of gameplay)".

AFAIK, AMD is the only company of the big three that hasn't blatantly been caught pulling a massive marketing deception. There have been issues like JF-AMD making claims that aren't real, but it wasn't official mention endorsed by the company.

 

[gamerk316]

Mantle performance benchmarks:

http://www.techspot.com/review/793-thief-battlefield-4-mantle-performance/page2.html

Some points:

Little improvement in BF. Kinda expected, given the game is already GPU bound in most situations.

Thief shows massive improvements on AMD, though its worth noting how unacceptably poor the game runs on AMD without Mantle. There is NO reason an i3 should be 12 FPS faster then an FX-8350.

In all cases, the same trend repeats: the weaker the GPU and higher the settings, the less gains you see. Big winners are weaker CPUs such as Trinity/Kaveri and the i3. Which again begs the question why AMD isn't shoving APU's in laptops...

In other news:

http://www.techspot.com/news/56129-meet-pascal-nvidias-next-generation-gpu-that-could-render-pcie-obsolete.html
http://devblogs.nvidia.com/parallelforall/nvlink-pascal-stacked-memory-feeding-appetite-big-data/#more-3097

Pascal won’t ship anytime soon but there’s plenty to get excited about in the meantime. Instead of the tried and true PCIe interface, the GPU will utilize a new socket called NVLink that was co-developed with IBM. It’ll offer speeds up to 12 times faster than what is currently possible over PCIe.

Of course, getting mobo makers to support this is going to be the challenge here...Then of course, the pesky "up to". But it looks like NVIDIA, since they can't make a CPU/GPU combo [lacking an X86 license] is doing the next best thing: Creating as much bandwidth as possible to hide the disadvantage of having a dGPU across a different bus.

Things are getting very interesting on the hardware front, since for the first time in a while, it looks like we are going to have several standards wars at the same time. Its like the 90's all over again...As it is, I'd be very hesitant to build a long term rig at this juncture, till we know what the future standards are going to be. Guess my 2600k has a few more years in it before I upgrade...

 

[palladin9479]

Which again begs the question why AMD isn't shoving APU's in laptops...

You answered your own question, in the same post.

Of course, getting mobo makers to support this is going to be the challenge here...

Notebook makers do use the APU, they just like to cripple it and "up sell" you on the pricer i5/i7 + dPGU model. Low priced notebooks don't have a large profit margin even though they sell in mass quantities. You do see a TON of A4 sub $300 notebooks all over store shelves.

 

[sarinaide]

The market is Intel and Nvidia centric, and AMD gets the left overs. Unfortunately there may come a day in the future that there will no longer be a choice and you will end up being raped out of pocket for a machine.

 

[vmN]

The market is Intel and Nvidia centric, and AMD gets the left overs. Unfortunately there may come a day in the future that there will no longer be a choice and you will end up being raped out of pocket for a machine.

There will always be a choice.
Intel cannot have monopoly over the CPU marked.
Hardware actually got cheaper, back in the days hardware was crazy expensive.

 

[sarinaide]

The market is Intel and Nvidia centric, and AMD gets the left overs. Unfortunately there may come a day in the future that there will no longer be a choice and you will end up being raped out of pocket for a machine.

There will always be a choice.
Intel cannot have monopoly over the CPU marked.
Hardware actually got cheaper, back in the days hardware was crazy expensive.

If AMD cannot survive then there will be no choice. Cheaper hardware is because prices got regulated by laws these are also global.

It is quite simple really, in the current state of technology, money lets you waste more in persuit of excellence, AMD doesn't have much to waste this is why Bulldozer was not scrapped and will run its course from Zambezi to whatever Excavator is called only then will AMD go back to the drawing board, fact is it will bankrupt AMD scrapping the plan and re-developing as opposed to rolling out the best product it can on said arch design and try sell it in a marketcentric environment.

So we are in a situation where AMD has very little profit on sales, probably enough to float but without any hope of raising stocks and capital. AMD is teetering on the edge and if they keep losing the graphics marketshare the inevitable fall may come. Right now AMD's losing marketshare on things as trivial as heat and slightly more power and suggestions that performance is not good when I haven't heard of any issues. Firstly for a gamer power and heat are minor issues, I have a 680w gold PSU and hardly concerned about marginal power differences, and I have good enough airflow that heat is not an issue, similarly I find the idea of super HTPC ludicrous at best with people slamming highest end graphics into small form factor with high end CPU's. The long and the short is AMD and you have seen in that thread "Dispelling myths" is seen in bad light and almost at will Intel and Nvidia are getting away with smear tactics, all having a fundamental effect. If it continues there will be no AMD going forward.

 

[vmN]

No, I do believe IBM made some deals with Intel and AMD that there would not be an monopoly over CPUs. (If someone could elaborate I would be more than happy to read it).

I do believe AMD are actually getting some market-share back in this current situation. AMD are not aiming for the market they used to anymore.

 

[sarinaide]

No, I do believe IBM made some deals with Intel and AMD that there would not be an monopoly over CPUs. (If someone could elaborate I would be more than happy to read it).

I do believe AMD are actually getting some market-share back in this current situation. AMD are not aiming for the market they used to anymore.

They are targeting diverse markets (whatever that means), emerging markets. In short low income markets that have not been pillaged by Intel and Nvidia because they don't see money in that segment.

AMD's market is system on chip APU's, and handheld or portable devices not laptops/ultra-thins, the other market is graphics cards. As for performance chips, OBR stated mid 2012 that Piledriver was going to be AMD's last performance CPU for the foreseeable future, many including myself were sceptical on this but as it turns out, there is no FX Steamroller and so it all was correct. The APU's occupy a safe market, Intel is not inclined to challenge budget systems again there is very little profit to be had and AMD's on chip graphics is far ahead but in a market where integrated graphics is still just short of entry level gaming won't see Intel throw $700 iris SoC's into the budget end. That said APU's are evolving well but we already know what needs to happen for AMD to really thrust APU's as legitimate entry level gaming chips (gaming level graphics on chip). Bandwidth is key and either a eDram is needed like with Iris or significantly more bandwidth through DDR4 and/or quad channel or some kind of GDDR5 soldered onto motherboards all of which are expensive.

 

[8350rocks]

All I know is this, my AMD contact has been telling me to "keep the faith, it is coming"....and that says to me that they are not done with HEDT. Conversationally, it seems fab constraints are not what they want at this time for another large die CPU.

The consensus is, keeping their own fab would have made them more competitive, however, it would have drowned them in debt through the process. It is a double edged sword. If they had the level of control over R&D for fabs that Intel had, even with a smaller budget, they would be farther along admittedly. Though in this circumstance, things are not going the way they would like. So, rather than spend money developing a low yield large die CPU on processes that are not yet mature...they are holding off until they are satisfied they can get the yields they are looking for. Good from a business perspective, though not great news from a consumer perspective.

 

[CooLWoLF]

But any new HEDT cpu from AMD is probably at least 2 years away, wouldn't you say?

 

[jdwii]

The market is Intel and Nvidia centric, and AMD gets the left overs. Unfortunately there may come a day in the future that there will no longer be a choice and you will end up being raped out of pocket for a machine.

Until they leave their module design behind i expect nothing more.

 

[gamerk316]

All I know is this, my AMD contact has been telling me to "keep the faith, it is coming"....and that says to me that they are not done with HEDT. Conversationally, it seems fab constraints are not what they want at this time for another large die CPU.

The consensus is, keeping their own fab would have made them more competitive, however, it would have drowned them in debt through the process. It is a double edged sword. If they had the level of control over R&D for fabs that Intel had, even with a smaller budget, they would be farther along admittedly. Though in this circumstance, things are not going the way they would like. So, rather than spend money developing a low yield large die CPU on processes that are not yet mature...they are holding off until they are satisfied they can get the yields they are looking for. Good from a business perspective, though not great news from a consumer perspective.

Hence my long standing belief AMD selling its fabs was a mistake. You do not outsource core functions.

 

[vmN]

It was not an mistake.
It is crazy expensive having your own fab, which AMD marketshare is too low for currently.

 

[truegenius]

All I know is this, my AMD contact has been telling me to "keep the faith, it is coming"....and that says to me that they are not done with HEDT. Conversationally, it seems fab constraints are not what they want at this time for another large die CPU.

then what they are waiting for ?

their best effort was fx9590 which was another crap specially because of high tdp and anyone with 8350 can achieve that type of clock rate at lower tdp

and i think that now intel is going to release 8 core for desktop, amd won't be able to touch this type of performance in near future, so i think that they are done with hedt only thing is that they are too stubborn to accept it

 

[logainofhades]

I think it was a mistake as well. It was a fix to a short term problem, that hurt them in the long run. Having to depend on other fabs, and basically having no control of their product as a result, does them 0 good. They should have kept at least 1 fab for quality control purposes and R&D. Depending on other companies, and you have no control of a process, is always riddled with issues. I say this from experience working in automotive parts supply. If a supplier f's up a component, that you need to make a particular part, it can royally screw you. Especially when one of the most important components to the product are made on the other side of the world. GF's process issues have slowed down AMD's progress. AMD would much rather be on a smaller node right now, but are unable to because the outsourced fabs can't seem to get the problems figured out.

 

[vmN]

For one: Money

Second: I do believe AMD want to make sure their next HEDT will actually gain a performance boost over the usual, so they actually can scrap in some market share.

Third: The entire 9xx0 series was a rather bad marketing strategy, and kinda ruined the FX line for further products. SR should have been x5x0, but since 9590 is piledriver it is rather messed up. But AMDs marketing team is alot stronger now, and are messing with Nvidias marketing team, which are doing terrible at this point.

 

[8350rocks]

According to what I have heard...evidently they have designs drawn up already and have done some testing, but the yields were not in line with what they were looking for. Additionally, as I understand it, Jim Keller is currently working on "something big"...I cannot get details but it does give me hope...

 

[de5_Roy]

and i think that now intel is going to release 8 core for desktop,

that will sell at a prohibitively expensive prices. x99 platform will have plenty of bugs if past launches are of any indication. intel will launch quite a few of new technologies and hardware, unlike with haswell and sb-e. intel might even charge higher than $1000.
amd doesn't need to chase that performance because of teh moniez needed to design such chips. afaik, chip design is very expensive business.
if amd admits they're not chasing hedt and amd's paid shills stop hyping "future(!) products" (e.g. bulldozer the core i7-killer) - their fx lineup will die overnight. intel will gladly help. it might not draw any legal implication because apus will still be in "competition". amd won't admit anything until carrizo comes out, at the least. by then they should be able to clear out their stock.
imo, if they can get a good share in b.r.i.c. countries and get enough feedback for an hedt platform, they might comeback to that. then again, those are not high margin markets and an hedt platform will likely be aimed at american and european markets.

 

[Cazalan]

For one: Money

Second: I do believe AMD want to make sure their next HEDT will actually gain a performance boost over the usual, so they actually can scrap in some market share.

Third: The entire 9xx0 series was a rather bad marketing strategy, and kinda ruined the FX line for further products. SR should have been x5x0, but since 9590 is piledriver it is rather messed up. But AMDs marketing team is alot stronger now, and are messing with Nvidias marketing team, which are doing terrible at this point.

If it's so bad then why is Intel copying them? Devil's Canyon will be Intel's line of pre-overclocked parts.

 

[jimmysmitty]

Man how much more stuff is Facebook going to ruin. The one piece of hardware I was actually really looking forward to, will now be spying on my every click.

http://www.brightsideofnews.com/news/2014/3/25/facebook-to-acquire-rift-maker2c-oculus-vr-for-242-billion.aspx

Damn. I wish VALVe kept them or bought them instead. Now they will become nothing in the end.

I can't wait to play Candy Crush on a Oculus Rift VR.

The market is Intel and Nvidia centric, and AMD gets the left overs. Unfortunately there may come a day in the future that there will no longer be a choice and you will end up being raped out of pocket for a machine.

There will always be a choice.
Intel cannot have monopoly over the CPU marked.
Hardware actually got cheaper, back in the days hardware was crazy expensive.

Funny thing, I tend to collect older CPUs. I one time found a Pentium 75MHz that still had the price tag on it. The price was $999.99.

People forget those days. I would gladly pay $300 for a quad core CPU that makes that 75MHz Pentium look like a steam machine.

It was not an mistake.
It is crazy expensive having your own fab, which AMD marketshare is too low for currently.

Yet sadly, when AMD had a better CPU they sold off one of their FABs instead of keeping the second one.

For one: Money

Second: I do believe AMD want to make sure their next HEDT will actually gain a performance boost over the usual, so they actually can scrap in some market share.

Third: The entire 9xx0 series was a rather bad marketing strategy, and kinda ruined the FX line for further products. SR should have been x5x0, but since 9590 is piledriver it is rather messed up. But AMDs marketing team is alot stronger now, and are messing with Nvidias marketing team, which are doing terrible at this point.

Marketing is always a strong point. While AMD didn't have the commercial marketing they have always had the "look at this number vs whatever".

Of course the newer marketing is interesting but it is the same cherry picking that every company does.

 

[jimmysmitty]

For one: Money

Second: I do believe AMD want to make sure their next HEDT will actually gain a performance boost over the usual, so they actually can scrap in some market share.

Third: The entire 9xx0 series was a rather bad marketing strategy, and kinda ruined the FX line for further products. SR should have been x5x0, but since 9590 is piledriver it is rather messed up. But AMDs marketing team is alot stronger now, and are messing with Nvidias marketing team, which are doing terrible at this point.

If it's so bad then why is Intel copying them? Devil's Canyon will be Intel's line of pre-overclocked parts.

How are they copying AMD? Devils Canyon only pushes the speed up 100MHz, much like with the i7 2600K to i7 2700K.

If they were to copy AMD they would just release a 4GHz stock overclocked part which I don't see listed in the parts at all.

 

[de5_Roy]

Man how much more stuff is Facebook going to ruin. The one piece of hardware I was actually really looking forward to, will now be spying on my every click.

http://www.brightsideofnews.com/news/2014/3/25/facebook-to-acquire-rift-maker2c-oculus-vr-for-242-billion.aspx

Damn. I wish VALVe kept them or bought them instead. Now they will become nothing in the end.

I can't wait to play Candy Crush on a Oculus Rift VR.

te possibilities are simply... endless. for example:
new gesture-based "likes" -
"wink" - by winking one eye. each eye will have different meaning.
"tch" - the vr headset's integrated mic will capture verbal commands of disapproval.
"raised (one) eyebrow", "looking down on .." - apple fanboy.
and so on.

there will be new kinds of data leaks. for example -
cybercriminals will find it ridiculously easy to steal retinal scanner data and other biometric data and ways to mimic them.
heartrate monitor and other fitness monitors will get far, far easier to hack.
during vr-enhanced candy crush gameplay, suddenly jason voorhees will barge in wielding a chainsaw thanks to a recent security breach or 1000.
two-factor authentication system will be a piece of cake to bypass.
and so on.

htc will make another facebook phone... this time with input port for vr headset so that people can use the rift while walking in public... like in train stations, bridges, crossings etc.