AMD CPU speculation... and expert conjecture

[juanrga]

You know, the funny thing is, if you want to support green companies, AMD is actually a Gold rated green company in terms of environmental footprint. NVidia and Intel are not even silver rated.

Keep buying efficiency though...

There is nothing more green and efficient than AMD 220W CPUs. Any market where efficiency is key (servers, HPC, mobile...) is dominated by the green company with Golden rate.

If Amd makes something competitive in performance per watt in the 200 watt metric i'll buy it if its priced around 350$ or so.

The 300 series will be 'refreshes' (aka relabeled 200 cards) except by a pair of new models. Since AMD cannot compete on efficiency with Nvidia, AMD will target 'enthusiasts' with ridiculously power hungry top cards. AMD likes to corner itself and with the 390X will target a niche piece of a niche market.

Until then they can keep their re-branded slightly improved design(mostly the same since 2011). How much money will the 390X be anyways probably priced at the 980 both cards of which are to much money for most in PC gaming which the sweet spot is 200-350$.

Some people on forums insists on that the 390x will be cheap. When I ask them about price leaks they gave no response except they expect AMD will give the cards for free. HBM is not cheap and if use 20nm node again this add to the cost, as does the hybrid cooler.

If one day Amd makes a 100 watt CPU that is competitive in single core and multicore performance of the comparable priced Intel i'd switch.

Lisa Su is spinning AMD outside the PC space just to avoid the direct competition with Intel.

I'm not tied to one brand and i'm happy i'm not if i was i would either have horrible single core performance with lots of TDP and a GPU with lots of TDP for the same performance as the competition that has half the TDP.
Or worse i'd have great single core performance but horrible graphics with the IGPU(Intel).
Or i'd just have a Arm CPU with a few cuda cores.

Amd, Intel, Nvidia are all out to get money that's the point just don't let them get to ya over a brand sticker.

But that is not how fanboys think. They will try to convince you one of those brands is a nice non-profit ONG that cares about customers whereas others two are evil companies.

 

[Reepca]

You know, the funny thing is, if you want to support green companies, AMD is actually a Gold rated green company in terms of environmental footprint. NVidia and Intel are not even silver rated.

Keep buying efficiency though...

I am quite certain NVIDIA is a very green company.

*ba-da chaaa*.

 

[jdwii]

"half the TDP"

http://tpucdn.com/reviews/MSI/GTX_970_Gaming/images/power_maximum.gif

290X 328 watts
970 169 watts
328/169= 1.941

http://tpucdn.com/reviews/MSI/GTX_970_Gaming/images/power_peak.gif

During gaming 181 watts(more used during gaming hinting at Nvidia detecting furmark(lying trolls)
290X 294 watts
970 181 watts
294/181 = 1.624

So not half but still WAY WAY enough

How cute though to try and pick that out. When you actually do the math its sadder then when i just pull the half statement out of thin air. Again lets wait for Amd to actually improve their design instead of adding HBM or putting something on a smaller footprint and calling it their own work. One day it might be half if its already close in some cases.

 

[jdwii]

AT has a nice update. They did research why the 290X did run slower under Mantle than under DX12 in the dual-core test. They traced the cause to the higher batch submission time of the Mantle implementation. Oxide explains that the Mantle implementation includes an optimization routine for small batches that can increase the CPU bottleneck on the lowest CPUs. The interesting part is that the routine was giving Mantle and artificial performance benefit over DX12 when using more than 2-cores.

Thus my former words "DX12 was neck-to-neck with Mantle" must be replaced by DX12 is faster than Mantle.

If we turn off the small batch optimization feature, what we find is that Mantle' s batch submission time drops nearly in half, to an average of 4.4ms. With the second pass removed, Mantle and DirectX 12 take roughly the same amount of time to submit batches in a single pass. However as Oxide noted, there is a performance hit; the Mantle rendering path's performance goes from being ahead of DirectX 12 to trailing it.

This feature is enabled by default in our build, and by combining those small batches this is the likely reason that the Mantle path holds a slight performance edge over the DX12 path on our AMD cards. The tradeoff is that in a 2 core configuration, the extra CPU workload from the optimization pass is just enough to cause Star Swarm to start bottlenecking at the CPU again.

Now Windows game developers will have to choose between an API (Mantle) that works only on AMD hardware or a faster API (DX12) that works on AMD, Nvidia, and Intel.

LOL @ your doom and gloom.

Want to know the great thing about software? You can release a new driver update and optimize code to fix it.

Want to know the bad thing about hardware like the GTX 970? You cannot fix it with software, only lessen the impact...

Dude that is another fallacy and its classified as a red herring. You brought something up that had nothing to do with mantle or 12 at all.

The point about improving mantle support from a driver level is a moot point since the same can be said for 12. Plus we can probably make an argument that Mantle had its share of optimizations over the year or so since its been out.

 

[Embra]

AMD’S REAL CHINA PLAY IS STRATEGIC INVESTMENT:
http://www.vrworld.com/2015/02/09/amds-real-china-play-strategic-investment/

 

[-Fran-]

AT has a nice update. They did research why the 290X did run slower under Mantle than under DX12 in the dual-core test. They traced the cause to the higher batch submission time of the Mantle implementation. Oxide explains that the Mantle implementation includes an optimization routine for small batches that can increase the CPU bottleneck on the lowest CPUs. The interesting part is that the routine was giving Mantle and artificial performance benefit over DX12 when using more than 2-cores.

Thus my former words "DX12 was neck-to-neck with Mantle" must be replaced by DX12 is faster than Mantle.

If we turn off the small batch optimization feature, what we find is that Mantle' s batch submission time drops nearly in half, to an average of 4.4ms. With the second pass removed, Mantle and DirectX 12 take roughly the same amount of time to submit batches in a single pass. However as Oxide noted, there is a performance hit; the Mantle rendering path's performance goes from being ahead of DirectX 12 to trailing it.

This feature is enabled by default in our build, and by combining those small batches this is the likely reason that the Mantle path holds a slight performance edge over the DX12 path on our AMD cards. The tradeoff is that in a 2 core configuration, the extra CPU workload from the optimization pass is just enough to cause Star Swarm to start bottlenecking at the CPU again.

Now Windows game developers will have to choose between an API (Mantle) that works only on AMD hardware or a faster API (DX12) that works on AMD, Nvidia, and Intel.

I already mentioned the AT update, Juan.

And it's nice to have that in hand to enable it or disable it for an extra performance squeeze. Why be so negative about it? Your usual "let's twist every single word to favor my agenda" is strong lately. Did AMD touch you un-properly during 2014 or something?

Plus, DX12 will be TIED to Windows 10. MANTLE is not tied to Windows 10; hence MS is preempting OGLN and MANTLE so they're giving away Win10 for peanuts. There's a world of discussion about what's better, but I'll keep my allegiance to OGL(N) and MANTLE over DX(12) because of freedom of OS. I hope AMD fully supports OGLN, since it should be close to MANTLE as well. And nVidia will be the green troll as usual.

In any case, nice to read AMD will get some fresh greens.

Cheers!

 

[jdwii]

Spoiler

AT has a nice update. They did research why the 290X did run slower under Mantle than under DX12 in the dual-core test. They traced the cause to the higher batch submission time of the Mantle implementation. Oxide explains that the Mantle implementation includes an optimization routine for small batches that can increase the CPU bottleneck on the lowest CPUs. The interesting part is that the routine was giving Mantle and artificial performance benefit over DX12 when using more than 2-cores.

Thus my former words "DX12 was neck-to-neck with Mantle" must be replaced by DX12 is faster than Mantle.

If we turn off the small batch optimization feature, what we find is that Mantle' s batch submission time drops nearly in half, to an average of 4.4ms. With the second pass removed, Mantle and DirectX 12 take roughly the same amount of time to submit batches in a single pass. However as Oxide noted, there is a performance hit; the Mantle rendering path's performance goes from being ahead of DirectX 12 to trailing it.

This feature is enabled by default in our build, and by combining those small batches this is the likely reason that the Mantle path holds a slight performance edge over the DX12 path on our AMD cards. The tradeoff is that in a 2 core configuration, the extra CPU workload from the optimization pass is just enough to cause Star Swarm to start bottlenecking at the CPU again.

Now Windows game developers will have to choose between an API (Mantle) that works only on AMD hardware or a faster API (DX12) that works on AMD, Nvidia, and Intel.

LOL @ your doom and gloom.

Want to know the great thing about software? You can release a new driver update and optimize code to fix it.

Want to know the bad thing about hardware like the GTX 970? You cannot fix it with software, only lessen the impact...

Dude that is another fallacy and its classified as a red herring. You brought something up that had nothing to do with mantle or 12 at all.

The point about improving mantle support from a driver level is a moot point since the same can be said for 12. Plus we can probably make an argument that Mantle had its share of optimizations over the year or so since its been out.

do you really want to compare amd driver updates to dx driver updates ?

Well even Amd and Nvidia and Intel can improve directx performance so its not just Microsoft. Oh and for the 100th time 12 being tied to 10 isn't as bad as mantle being tied to just GCN video cards. Plus Microsoft giving 10 away for free really gets rid of all the burden of upgrading since businesses don't care if BF7 will work on their machines but a consumer might.

 

[palladin9479]

As palladin said, some of you either have very short term memories or are being deceptive. It just takes a few simple searches of the internet to see that Mantle came first.

To be fair, NVIDIA/MSFT claimed DX12 had been in development for over a year before Mantle was released. Whether that's political BS or not, I can't say, but I will say that I doubt MSFT could have come up with a brand new API, WDDM model, integrated it with its next version of it's OS, and get a working demo out running on old hardware using a subset of the new API, in just a few months.

I'm pretty sure MS had WDDM 2.0 and DX12 under development for awhile and was keeping it on the back burner in case Windows 8 wasn't successfully adopted. What I believe happened is that they observed AMD's Mantle being successful amongst developers and decided to incorporate features from it into their own API. As I said it wasn't a "copy" or "theft", just typical vender competition. AMD's method of accelerating and optimizing the graphics pipeline was successfully demonstrated to provide a large increase in efficiency, these optimizations aren't themselves vender specific and could be incorporated into any hardware provided the API supported them.

So yes we are seeing the API's start to merge, where the best features are being implemented into each one and eventually we'll have a single API. I predict that OpenGL will also implement these features and thus solve the open-vender cross-platform problem.

 

[cdrkf]

As palladin said, some of you either have very short term memories or are being deceptive. It just takes a few simple searches of the internet to see that Mantle came first.

To be fair, NVIDIA/MSFT claimed DX12 had been in development for over a year before Mantle was released. Whether that's political BS or not, I can't say, but I will say that I doubt MSFT could have come up with a brand new API, WDDM model, integrated it with its next version of it's OS, and get a working demo out running on old hardware using a subset of the new API, in just a few months.

I'm pretty sure MS had WDDM 2.0 and DX12 under development for awhile and was keeping it on the back burner in case Windows 8 wasn't successfully adopted. What I believe happened is that they observed AMD's Mantle being successful amongst developers and decided to incorporate features from it into their own API. As I said it wasn't a "copy" or "theft", just typical vender competition. AMD's method of accelerating and optimizing the graphics pipeline was successfully demonstrated to provide a large increase in efficiency, these optimizations aren't themselves vender specific and could be incorporated into any hardware provided the API supported them.

So yes we are seeing the API's start to merge, where the best features are being implemented into each one and eventually we'll have a single API. I predict that OpenGL will also implement these features and thus solve the open-vender cross-platform problem.

Well said. How anyone thinks this is cause to bash on AMD for Mantle is beyond me. @juan has a point- Mantle isn't going to stay around much longer imo, as once cross vendor API's are available what is the point? Still I do think it had a large part to play in *pushing* dx12 and OglN a bit further up the priority list, and it's not a big stretch to image both groups have taken a good look at what Mantle does well and made tweaks accordingly

The end result? Faster more efficient API's across the board and better support for mult core cpu's with weaker per core performance. It should help to mitigate future AMD cpu's from getting quite so badly spanked in games benchmarks.

 

[jdwii]

tbh i hope mantle does stay around to push 12 and push Microsoft. I know i might come off as pro this or that but i want some freaking competition.

 

[juanrga]

And it's nice to have that in hand to enable it or disable it for an extra performance squeeze. Why be so negative about it?

Do you mean realist.

Plus, DX12 will be TIED to Windows 10. MANTLE is not tied to Windows 10; hence MS is preempting OGLN and MANTLE so they're giving away Win10 for peanuts. There's a world of discussion about what's better, but I'll keep my allegiance to OGL(N) and MANTLE over DX(12) because of freedom of OS. I hope AMD fully supports OGLN, since it should be close to MANTLE as well. And nVidia will be the green troll as usual.

We know that DX12 will be tied to Windows 10 whereas Mantle is tied to GCN hardware. Why do you repeat the obvious whereas ignore the main point?

Also Microsoft decision to give W10 for peanuts is not related to anything Mantle/AMD. Microsoft is competing against Google and others.

@juan has a point- Mantle isn't going to stay around much longer imo, as once cross vendor API's are available what is the point? Still I do think it had a large part to play in *pushing* dx12 and OglN a bit further up the priority list, and it's not a big stretch to image both groups have taken a good look at what Mantle does well and made tweaks accordingly

Mantle would have a bright future if it did match the points I mentioned before:

• 
  ■ Not limited to GCN hardware.
  ■ Multiplatform. During Mantle presentation Mac and linux ports were promised. Apple rejected it. And AMD no longer talks about linux.
  ■ Superior performance thanks to lower-level access. AMD promised Mantle would be faster than DX12 [1] First tests show Mantle is neck-to-neck thanks to tricks and behind DX12 without those.
  ■ Broad acceptation among developers. AMD promised lots of new games before the end of 2014. Still waiting for them.
  ■ Earlier development. AMD promised final spec before the end of 2014. It is February 2015 and Mantle continues on perennial beta stage.

In short most if not all what AMD promised has vanished in the air... as usual.

[1] The quote from AMD on the differences between Mantle and DX12:

Mantle is all about trying to give tools to developers who want to extract more performance out of a game at any given situation, whereas DirectX is primarily about getting a broad set of compatibility so you can run your code once for DirectX and have it run on all of the different PC hardware that is out there.

 

[8350rocks]

As palladin said, some of you either have very short term memories or are being deceptive. It just takes a few simple searches of the internet to see that Mantle came first.

To be fair, NVIDIA/MSFT claimed DX12 had been in development for over a year before Mantle was released. Whether that's political BS or not, I can't say, but I will say that I doubt MSFT could have come up with a brand new API, WDDM model, integrated it with its next version of it's OS, and get a working demo out running on old hardware using a subset of the new API, in just a few months.

I'm pretty sure MS had WDDM 2.0 and DX12 under development for awhile and was keeping it on the back burner in case Windows 8 wasn't successfully adopted. What I believe happened is that they observed AMD's Mantle being successful amongst developers and decided to incorporate features from it into their own API. As I said it wasn't a "copy" or "theft", just typical vender competition. AMD's method of accelerating and optimizing the graphics pipeline was successfully demonstrated to provide a large increase in efficiency, these optimizations aren't themselves vender specific and could be incorporated into any hardware provided the API supported them.

So yes we are seeing the API's start to merge, where the best features are being implemented into each one and eventually we'll have a single API. I predict that OpenGL will also implement these features and thus solve the open-vender cross-platform problem.

AMD's OGL team is actually working with Khronos on the API to improve the new iteration.

 

[truegenius]

neither windows 8.1/8 or 7 will ever have dx12, ther isn't even the smallest bit of chance that would ever happen
Quatin from anandtech dx12 preview "Because DirectX 12 and WDDM 2.0 are tied at the hip, and by extension tied to Windows 10, DirectX 12 will only be available on Windows 10. Windows 8/8.1 and Windows 7 will not be receiving DirectX 12 support.
Backporting DirectX 12 to earlier OSes would require backporting WDDM 2.0 as well, which brings with it several issues due to the fact that WDDM 2.0 is a kernel component. Microsoft would either have to compromise on WDDM 2.0 features in order to make it work on these older kernels, or alternatively would have to more radically overhaul these kernels to accommodate the full WDDM 2.0 feature set, the latter of which is a significant engineering task and carries a significant risk of breaking earlier Windows installations. Microsoft has already tried this once before in backporting parts of Direct3D 11.1 and WDDM 1.2 to Windows 7, only to discover that even that smaller-scale project had compatibility problems. A backport of DirectX 12 would in turn be even more problematic."

oh fish , if your vision are real, does that mean end of world (w7) ? ( dialogue of psycho from crysis 3 )

and what sony will do now ? support mantle ?

 

[juanrga]

and what sony will do now ? support mantle ?

Mantle is Windows exclusive. Moreover why would Sony support something that is poor than what it has implemented on PS4?

 

[juanrga]

AMD's OGL team is actually working with Khronos on the API to improve the new iteration.

As are other teams from other companies.

 

[8350rocks]

and what sony will do now ? support mantle ?

Mantle is Windows exclusive. Moreover why would Sony support something that is poor than what it has implemented on PS4?

Mantle is not windows exclusive...

Juan, now you are outright lying. Time to stop the shenanigans.

 

[8350rocks]

AMD's OGL team is actually working with Khronos on the API to improve the new iteration.

As are other teams from other companies including Qualcomm. The official presentation of glNext will be made by a number of developers including a pair of former Nvidia guys.

I am sure they are, however, I will only point out that Khronos came *specifically* to AMD to help them engineer the improvements in pipelines and reduced overhead.

 

[-Fran-]

and what sony will do now ? support mantle ?

Mantle is Windows exclusive. Moreover why would Sony support something that is poor than what it has implemented on PS4?

Mantle is not windows exclusive...

Juan, now you are outright lying. Time to stop the shenanigans.

Just like MANTLE has no other target platform on the horizon, he's talking from what's already there. And just like we say that MANTLE "sparkled" MS to develop DX12 "closer to metal" due to "how the facts are laid down".

Juan has done a complete 180° with AMD and it's showing too much. From lover to hater (they do say there's a thin line, lol), so I just read his comments from a hater's perspective nowadays. Every twist of information is "for doom and gloom" instead of having a neutral analysis on the information.

Cheers!

 

[logainofhades]

tbh i hope mantle does stay around to push 12 and push Microsoft. I know i might come off as pro this or that but i want some freaking competition.

Competition would be nice. Not really been much, if any, CPU wise, and I am starting to not feel so good about the GPU side either. The refresh game, from both Nv and AMD, has gotten old.

 

[-Fran-]

http://www.anandtech.com/show/8718/the-samsung-galaxy-note-4-exynos-review

It seems ARMv8 's first implementation (A53 + A57) from Samsung is a power hungry, slightly more efficient (at times) SoC. It's still a step up, but by no means better all around. Interesting to read with *lots* of information that will fill your head like a toilet at a music fest (hahaha).

Cheers!

EDIT: "Among the first companies shipping a server CPU will be AMD and their A1100, so it will be interesting to see how A57 plays out in the server market over the coming year." <- from the article.

 

[juanrga]

and what sony will do now ? support mantle ?

Mantle is Windows exclusive. Moreover why would Sony support something that is poor than what it has implemented on PS4?

Mantle is not windows exclusive...

Juan, now you are outright lying. Time to stop the shenanigans.

I know you see Mantle everywhere:

Ubuntu 14.10 will have full mantle support via catalyst drivers...which means the drivers will be available in the debian tree at that time or sooner depending on what repositories you use.

14.2 beta is already out for linux OS...you might take a look at it. It is not full mantle support yet...but many of the features are already in the driver, it is just not polished enough to activate all of it.

Not only your words were untrue but Ubuntu 15 is being released without Mantle... because Mantle "is intended specifically for Windows on PCs.​" (AMD's own words)

AMD's OGL team is actually working with Khronos on the API to improve the new iteration.

As are other teams from other companies including Qualcomm.

I am sure they are, however, I will only point out that Khronos came *specifically* to AMD to help them engineer the improvements in pipelines and reduced overhead.

The history is different. Kronos group first announced development of next OpenGL iteration and latter AMD offered to the Kronos Group full access to the Mantle API. Kronos is using knowledge from different parties involved in the development of glNext. It is worth mentioning this old quote from John Carmack:

Nvidia's OpenGL extensions rival AMD Mantle

It is also relevant remark AMD's own words on the relationship between Mantle and DX12:

What is the relationship between Mantle and DirectX® 12?

​There is no direct relationship, but we support and celebrate a direction for game development that is aligned with AMD’s vision of lower-level, "closer to the metal" graphics APIs for PC gaming. While industry experts expect this to take some time, developers can immediately leverage efficient API design using Mantle.​

This is a missile to all those that claim that DX12 is Mantle in disguise.

 

[juanrga]

http://www.anandtech.com/show/8718/the-samsung-galaxy-note-4-exynos-review

It seems ARMv8 's first implementation (A53 + A57) from Samsung is a power hungry, slightly more efficient (at times) SoC. It's still a step up, but by no means better all around. Interesting to read with *lots* of information that will fill your head like a toilet at a music fest (hahaha).

Cheers!

EDIT: "Among the first companies shipping a server CPU will be AMD and their A1100, so it will be interesting to see how A57 plays out in the server market over the coming year." <- from the article.

They did test AArch32 mode, most of the improvements of the new cores require the native AArch64 mode. Their discussion of power measurement is weird.

Regarding the A1100, we already know is being crushed by both sides x86 and ARM: Intel Avoton and APM XGene give better performance and efficiency. AMD could try to compete on cheap.

 

[blackkstar]

So much wrong in this thread about Mantle. I see the sour grapes are out in full force :^)

http://community.amd.com/community/amd-blogs/amd-gaming/blog/2014/05/28/mantle-the-start-of-low-overhead-future

"Not directly related to DX12"
AMD is working with MS for DX12 like they always do in new versions of DX

Mantle still the easiest way to port to all platforms, including DX12

https://en.wikipedia.org/wiki/Mantle_%28API%29
Remember when all the anti-Mantle people were bashing on it for not being open? And now it's time for everyone to love DirectX when it's not open!

There's so much hypocrisy here. So many people who love Gameworks, Mantle, G-Sync, PhysX, etc suddenly turn on AMD Mantle for having vendor lock in. If you like any Nvidia proprietary tech, you have absolutely zero room to complain about AMD creating an API that they will open up that currently has vendor lock in. If anything, you should be upset at Nvidia for not embracing it.

You're probably all lining up defending Windows too. All the software is optimized for Intel there, give me my vendor lock in as long as I can keep my Windows! I don't get some of you. Windows 10 preview shipped with a keylogger, you know.
http://www.pcpro.co.uk/news/391114/microsoft-admits-windows-10-preview-has-a-keylogger

If you think that sort of vendor lock in is way better than vendor lock in for a completely optional API (I've yet to see a game solely support Mantle), you're beyond redemption. Even if Mantle has lock in, you can just not buy an AMD card and use DirectX or OpenGL. However, being locked into Windows means you take whatever Windows wants. If they want to monitor you on their discounted version of Windows, keylog you, and display ads, you shut up and take it, because it's that or nothing. The two types of lock in are completely not comparable.

OpenGLN and Mantle are our only chances of escaping Microsoft lock in. And they're both alternatives that have yet to block others from accessing things.

How proprietary things like Mantle and PhysX are not as harmful as being tied to Windows. If you don't have Mantle or PhysX support, you still get rendering and in game physics. If you don't have DirectX support, you get absolutely nothing. Zero, zilch, nada. No game at all.

The vendor lock in to Windows is more like games only support Mantle or they don't work. Or having games where you have an Nvidia card with PhysX or your game has no physics engine. The fact that some of you take the Windows option over another one is really, really confusing to me. Mantle is completely optional and supplements what you have if you have the right hardware. DX12 is assimilate or get nothing. Even things like G-Sync, PhysX, Gameworks, etc aren't as harmful to users as being locked into to software like you are with Windows. Because, like I said, the Nvidia and AMD technologies are supplemental and optional. But DX12 isn't. You either get DX11 on older versions of Windows, OGL/Mantle on other OSes, or you don't play games.

 

[juanrga]

http://community.amd.com/community/amd-blogs/amd-gaming/blog/2014/05/28/mantle-the-start-of-low-overhead-future

"Not directly related to DX12"
AMD is working with MS for DX12 like they always do in new versions of DX

Mantle still the easiest way to port to all platforms, including DX12

This image from Hallock's blog article is very interesting:

Because it shows that porting DX11-->DX12 requires only medium effort, whereas porting DX11-->Mantle-->DX12 requires medium+low. Therefore even AMD is giving another reason why game developers will prefer the direct native DX12 path.

 

[-Fran-]

Regarding the A1100, we already know is being crushed by both sides x86 and ARM: Intel Avoton and APM XGene give better performance and efficiency. AMD could try to compete on cheap.

Any information you have around that claim? From what the Article mentions, AMD will be the first ARMv8 on market, so that means APM XGene is still under work. How can you claim it will be better than AMD's implementation?

Cheers!