AMD CPUs, SoC Rumors and Speculations Temp. thread 2

[jimmysmitty]

Ok, we're not going to start beating the same old dead horses again. Any comments not directly related to the thread topic (AMD CPUs, SoC Rumors and Speculations) will be summarily deleted.

We can try but I can tell you from experience in the K10/BD/pretty much any AMD speculation thread it will always end up with the same debate "Intel screwed AMD, nothing AMD has done was bad".

 

[ex_bubblehead]

Well, there are some alternatives 😉

 

[de5_Roy]

i am all for "directly".
i am also for letting the old scabs heal and stuff.

however, let us leave this thread in idle mode for now until something new comes out.
right now amd is all about rolling out fiji and co. i am pretty sure we'll see some new leaks around the time windows 10 officially rolls out. amd won't waste the chance to show off cpu and gfx card benches on dx 12.

*pokes 8350rocks*

there's actually been some new "leaks" from benchlife - i'll address them all at once: those look way to fishy with socket fm3, way too similar to the beyond3d "leaks", some look so obviously 'shopped that their credibility is ruined (even if they turn out to be true). that is why i never posted them. not enough salt and stuff. also fool me once and stuff.

next month, intel will launch skylake platform - another chance for rumors and leaks.

 

[8350rocks]

I have honestly not been privy to anything official about benchmarks coming out...does not mean the leaks could not possibly have come from AMD. However, it surprises me that anything coming out now would be talking much about FM3 since it was essentially a hold over platform, and nothing more, also with the announcement of AM4 socket, not even sure we will still see FM3 at this point.

 

[gamerk316]

http://www.ign.com/articles/2015/07/02/rumors-suggest-nintendo-is-targeting-a-2016-nx-release-date-production-begins-in-october?utm_source=IGN%20hub%20page&utm_medium=Wii%20U&utm_content=5&utm_campaign=Blogroll

Might have 20 million NX systems next year all powered by Amd's architecture K12 or Zen

Or just turn down the quality settings to match NVidias drivers. They're playing IQ games again.

http://semiaccurate.com/forums/showthread.php?t=8749

I'd like to point out that every GPU manufacturer is going to create the same image a different way; there is always going to be some small differences between manufacturers as a result. But at least we're past the days of obvious EQ differences...

 

[jimmysmitty]

Or just turn down the quality settings to match NVidias drivers. They're playing IQ games again.

http://semiaccurate.com/forums/showthread.php?t=8749

Does it do this in every game? If not its a mistake one made by Amd several times
Edit original person said he had a recording error and it works fine also on page 5 i just about started cracking up, conspiracy theorist man.

Yea this seems a bit fishy. Seems like the guy recorded at lower image quality settings on the Titan X (12Mbps) vs the Fury X (30Mbps) so if you look at the video it looks worse but from what I can tell they look very similar. Also the "screenshots" are hard to use as they have soo much "movement" in them you can't get a clear shot.

That said, I have seen articles for both sides doing the AF cheat. Last time I saw NVidia do it was a long time ago. I think it was with the FX5000 series. AMD did it a few years back, was accidental or so it is said but it was corrected when it was found.

The real question is if someone with both GPUs (or even a 980Ti) could test it as well as test a static scene for image quality and not some shaky/blurry scene like the BF4 scene.

 

[gamerk316]

Chipmaker AMD cuts revenue estimate, citing weak PC sales:

http://www.reuters.com/article/2015/07/06/us-adv-micro-dev-estimates-idUSKCN0PG2BL20150706

And there went the past months stock gains.

 

[de5_Roy]

AMD Confirms 20nm Products Moved to FinFET, Warns on Q2’15 Earnings
http://www.anandtech.com/show/9418/amd-confirms-20nm-products-moved-to-finfet-warns-on-q215-earnings
actual insights to why this happened. the foundries' (mainly globalfoundries and tsmc) collective failure on 20nm has hit amd pretty badly. it says 33 mill for migration for the impact is much deeper than that.

 

[gamerk316]

Down 15% so far today, though it looks like it's stabilized.

 

[WarCrysis]

Great.. I just want to see new Zen Benchmarks. I am very interested in what they can come up with. i feel as though with a 2$ share price AMD is close to radical change as a company or a buy out.

I saw some talking about the Microsoft buy out rumor. Some think that its bad for microsoft. If Microsoft has 90 something billion dollars to spend then spending 2 billion on amd isnt a huge deal. As well as maybe 10 billion more in investment in to AMD. A few reasons i think this is a viable option is microsoft could reduce cost of everything making there products cost less and better then anyones. also they could put up real R&D money for AMD to maybe beat Intel.

Only conflict i see is Sony. Microsoft will make money off every PS4 sold. but we will leave that up to the courts. I dont think the Idea is as crazy as it sounds.

 

[de5_Roy]

zen benches won't come out for a while. may be... later this year, around october or later. or may be around intel's skylake lineup launches.

 

[jimmysmitty]

Great.. I just want to see new Zen Benchmarks. I am very interested in what they can come up with. i feel as though with a 2$ share price AMD is close to radical change as a company or a buy out.

I saw some talking about the Microsoft buy out rumor. Some think that its bad for microsoft. If Microsoft has 90 something billion dollars to spend then spending 2 billion on amd isnt a huge deal. As well as maybe 10 billion more in investment in to AMD. A few reasons i think this is a viable option is microsoft could reduce cost of everything making there products cost less and better then anyones. also they could put up real R&D money for AMD to maybe beat Intel.

Only conflict i see is Sony. Microsoft will make money off every PS4 sold. but we will leave that up to the courts. I dont think the Idea is as crazy as it sounds.

Microsoft makes money for every Android based device sold so making money on every PS4 wouldn't be a conflict.

The conflict I see for Microsoft is with Intel. They work very heavily with Intel as Intel tends to be very deeply rooted into newer technologies, such as the USB standard, memory standard etc. I doubt Microsoft would want to create issues with them as they have the majority of desktop and server market in terms of processors.

I highly doubt it would be Microsoft.

The other issue is the x86 license. It is non-transferable but I would think Intel would sign a new x86 license with the purchasing company to avoid a monopoly issue.

 

[gamerk316]

zen benches won't come out for a while. may be... later this year, around october or later. or may be around intel's skylake lineup launches.

Funny, since I've seen a few sites say "2016, probably later in the year" for Zen release. A year is a bit early on the benchmark front.

 

[de5_Roy]

Funny, since I've seen a few sites say "2016, probably later in the year" for Zen release. A year is a bit early on the benchmark front.

depends on when amd gets the test chips - which depends on how successful fabs (guessing glofo) are fabbing the cpus. 14nm has been in early production since q2...so amd should have early production chips and benchmark leaks later this year. this post has been a product of a lot of napkin math and a lot of napkins.

i'd watch out for leaks and rumors during intel's major launches.

 

[esrever]

There will always be the fake leaks and the AMD "internal benchmarks" that don't mean anything. Those will probably trickle out with skylake.

 

[de5_Roy]

There will always be the fake leaks and the AMD "internal benchmarks" that don't mean anything. Those will probably trickle out with skylake.

+1. leaks will get gradually reliable closer to launch.

 

[juanrga]

zen benches won't come out for a while. may be... later this year, around october or later. or may be around intel's skylake lineup launches.

There is no silicon to test still. The ES are expected for April 2016. But considering AMD finances, I would expect delays.

 

[jimmysmitty]

zen benches won't come out for a while. may be... later this year, around october or later. or may be around intel's skylake lineup launches.

There is no silicon to test still. The ES are expected for April 2016. But considering AMD finances, I would expect delays.

What is interesting me is the way Zen is being described. It is being described like I would see a Intel CPU. I do wonder if it is based on an older Intel design maybe Sandy Bridge?

Back when Intel and AMD settled, over the whatever AMD was up in arms about, they resigned a new agreement for cross patent agreements. Basically Intel and AMD had access to each others patents without having to pay or what not.

So I am wondering if AMD is using Intel ideas to create Zen. One thing that makes me thing this is SMT of course, that is an Intel design based on something DEC Alpha had (much like the IMC and the module approach for BD) but also the "new" high speed cache. Intel has always had a much faster cache design than AMD, even when their CPUs sucked their cache was faster.

It wouldn't surprise me. If you had access to ideas that worked, wouldn't be sued/fined for using them and your current ideas failed would you not utilize them to your advantage instead of investing billions of dollars you didn't have into R&D?

Just a thought.

 

[de5_Roy]

Desktop
http://support.amd.com/en-us/download/desktop?os=Linux+x86
very minor update about carrizo.

Vulkan API - Seekscale (interview)
http://home.seekscale.com/blog/the-future-of-graphics-programming-the-vulkan-api
the cpu sutff is pretty promising.

AMD A8-7670K "Godavari" APU is available for pre-order
http://www.cpu-world.com/news_2015/2015070801_AMD_A8-7670K_Godavari_APU_is_available_for_pre-order.html

 

[-Fran-]

What is interesting me is the way Zen is being described. It is being described like I would see a Intel CPU. I do wonder if it is based on an older Intel design maybe Sandy Bridge?

Back when Intel and AMD settled, over the whatever AMD was up in arms about, they resigned a new agreement for cross patent agreements. Basically Intel and AMD had access to each others patents without having to pay or what not.

So I am wondering if AMD is using Intel ideas to create Zen. One thing that makes me thing this is SMT of course, that is an Intel design based on something DEC Alpha had (much like the IMC and the module approach for BD) but also the "new" high speed cache. Intel has always had a much faster cache design than AMD, even when their CPUs sucked their cache was faster.

It wouldn't surprise me. If you had access to ideas that worked, wouldn't be sued/fined for using them and your current ideas failed would you not utilize them to your advantage instead of investing billions of dollars you didn't have into R&D?

Just a thought.

I don't think AMD has access to Intel's "how to's" to be honest.

Cheers!

EDIT: To add a little bit to it; they have access to the ISA and all, but they can't use the same techniques Intel has developed to actually process the ISA. That's how I understand it.

 

[gamerk316]

Vulkan API - Seekscale (interview)
http://home.seekscale.com/blog/the-future-of-graphics-programming-the-vulkan-api
the cpu sutff is pretty promising.

All I'm going to say: Do you REALLY trust the developer to handle manual memory and thread management correctly the first several months (years?) after the API hits? There's a reason APIs have been going the other way for the past two decades.

 

[gamerk316]

AMD/Intel can use eachothers extensions to the X86 ISA (AMD64, SSEx.y, etc), but the underlying architectures are still protected by various trademark agreements. EG: AMD simply can't rip off an existing Intel design.

 

[de5_Roy]

IBM Research successfully produces 7-nm test chips
http://techreport.com/news/28603/ibm-research-successfully-produces-7-nm-test-chips
how long till this trickles downs to other cpa members... if it pans out, tessie will fall behind sammy 😗
/edit2

All I'm going to say: Do you REALLY trust the developer to handle manual memory and thread management correctly the first several months (years?) after the API hits? There's a reason APIs have been going the other way for the past two decades.

i wouldn't write off every dev. and it's well understood that new paradigm will take from months to years to get going. the end results will be worth it.

@jimmysmitty: the cpu was rumored to be configured that way from the beyond 3d leaks which turned out to be fakes. those fakes have a real root though (which is what made those seem believable). it's amd's seattle soc. the seattle's cpu part is configured like a real desktop cpu. i assume the prankstar took that idea and projected on possible zen based cpu designs. even if the high level configuration as well as SMT looks the same, the underlying design architecture will likely be radically different.

 

[-Fran-]

IBM Research successfully produces 7-nm test chips
http://techreport.com/news/28603/ibm-research-successfully-produces-7-nm-test-chips
how long till this trickles downs to other cpa members... if it pans out, tessie will fall behind sammy 😗
/edit2

That is very interesting news. All I want to read next is "GloFo will get it next year after we exit the experimental phase", haha.

All I'm going to say: Do you REALLY trust the developer to handle manual memory and thread management correctly the first several months (years?) after the API hits? There's a reason APIs have been going the other way for the past two decades.

i wouldn't write off every dev. and it's well understood that new paradigm will take from months to years to get going. the end results will be worth it.

More to that, currently Devs DO work at that low level with Consoles. It's not something that no Dev, who works in the Game Industry, hasn't seen or worked on at all. Plus, you can always count on the graphical engines taking most of that burden going forward. Devs would only have to tweak (a lot) and that's about it.

@jimmysmitty: the cpu was rumored to be configured that way from the beyond 3d leaks which turned out to be fakes. those fakes have a real root though (which is what made those seem believable). it's amd's seattle soc. the seattle's cpu part is configured like a real desktop cpu. i assume the prankstar took that idea and projected on possible zen based cpu designs. even if the high level configuration as well as SMT looks the same, the underlying design architecture will likely be radically different.

Zen is supposed to be a "from scratch" approach, but that's never the case. They'll probably take the best of the Barcelona approach and adapt it somehow. Hell, I would not totally discard taking Bulldozer elements into the mix. There has to be something good about BD, right? Haha.

Cheers!

 

[de5_Roy]

Zen is supposed to be a "from scratch" approach, but that's never the case. They'll probably take the best of the Barcelona approach and adapt it somehow. Hell, I would not totally discard taking Bulldozer elements into the mix. There has to be something good about BD, right? Haha.

Cheers!

it's the same good thing that intel achieved in haswell-e cpus - scalability and modularity. these cpus will be optimized for non-customized processes unlike previous times. high clocks and long pipelines are likely to go away.