AMD CPU speculation... and expert conjecture

[juanrga]

Congratulations, you've lost the game. You showed that you are worse troll than hafijur.

8350rocks is not a troll.

hafijur is here for trolling, and you have been helping him with your anti-AMD rants. Up to now your tactic to fuel the troll was very astute and almost inappreciable. Now that you moved to insulting several posters here, everything is easier.

http://www.zdnet.com/blog/ou/is-amd-being-hypocritical-about-benchmark-ethics/438
http://www.zdnet.com/blog/ou/strike-3-for-amd-hypocrisy-on-benchmarking/463

AMD go even further they use benchmarks they get spanked on mainly but 1 out of a few they win they post to show they are better lol. Another thing is they use there apu or cpu performance and compare it to an intel cpu+igpu to make it sound like they are so superior at gaming lol instead of cpu tasks as they compare the processors, its quite funny them showing cpus so low end and trying to make it sound like amd are so much better. Haswell now stops amd from showing these benchmarks like they used to all the time and now intel focused on igpu amd fanboys call it hasfail as amds worst nightmares came true with haswell. Its common nature for both companies to show there cpus in the best frame but amd mainly nowadays just boast about there graphics performance in there apus lol and forget the apu cpu especially in laptops get beaten by core 2 duos as an all round chip from 7 years ago.

The irony is amds higher end graphics cards only work properly at full performance with an intel cpu so without intel there 8970m with an a10-5750m would get less then half the performance in games. The i7 4700mq one of intel entry level i7 mobile hasfail cpus you like to say beats out the fx8350 on multithreaded tasks.

http://www.notebookcheck.net/AMD-Radeon-HD-8970M.90656.0.html
http://www.notebookcheck.net/Intel-Core-i7-4700MQ-Notebook-Processor.92959.0.html

TROLLING ALGORITHM:

while ( 3 > 1 )
{
  post nonsense;
  ignore corrections;
}

 

[juanrga]

While I would love to see Kaveri with DDR5, I don't think its going to happen. AMD is going to want Kaveri to be cost competitive above all, and DDR4 or DDR5 would just make the total system builds too high for not that much gain (yet). Hopefully Excavator will have DDR5

I shared my hopes about GDDR5.

DDR5 doesn't exist.

Meant GDDR5 - got to typing too fast

Ok . Maybe carrizo (Kaveri sucessor) will include GDDR5.

I would like an Excavator APU with 1024 ALUs and 16GB GDDR5 in DIMM. My dream!

 

[hcl123]

ASUS announces two FM2+ motherboards: http://translate.google.se/translate?sl=sv&tl=en&js=n&prev=_t&hl=sv&ie=UTF-8&u=http%3A%2F%2Fwww.sweclockers.com%2Fnyhet%2F17328-asus-tillkannager-fm2-moderkort-med-stod-for-amd-a-serien-kaveri

• 
  Asus A88XM-A (Apparently named "Slurry" by Google )
  Asus A55BM-A/USB3

No DDR4!

There is a lot of "misconceptions" about DDR4... which has equivalent interface to GDDR5, as DDR3 was equivalent to GDDR4 (which never stuck, thanks to AMD and their GDDR5).

There will not be DDR4 above 2666Mhz before 2014.. perhaps in a "more realistic" vision only 2015 (if not 2016) (edt)...

JEDEC chart
http://pc.watch.impress.co.jp/img/pcw/docs/488/696/15.jpg

And there is a "misconception" about *power* to

MEMCON chart
http://pc.watch.impress.co.jp/img/pcw/docs/387/444/kaigai-10.jpg

So DDR4 is about "more bandwidth" and they do it at the expense of latency somehow, and does it "burning" more power in the process to. WORST, DDR4 is a point-to-point topology, that is, to have more than 1 DIMM per channel you need "buffer extenders" in those DIMMS... for "server world" that is no problem since all DRAM is already "registered", but for desktop it will make those DIMMS much more expensive...

So DDR4 will mostly good for "servers", more cores need more bandwidth...

Besides DDR3 is already 2400MHz, that is, above bandwidth to now DDR4 2133, and way above in latency... nothing prevents AMD to commercialize DDR3 2400Mhz sticks... it could do plenty the trick for kaveri...

Also i'm still convinced that "mobile" Kaveri will be pretty identical to Richland in the memory department(edt), and the Kaveri 2.0 ( the mainstream/desktop only in 2014)... could have some ESDRAM on die (like the XB one, Wuii )...

 

[sarinaide]

We cannot knitpick what company x or y wants to focus on as it is irrelevent, it is true DT is not a focal area as it used to be but irrelevent to that "us" being the majority on tech sites are not concerned one iota about how efficient a notebook is because our principle PC usage is on DT, what X or Y intends to focus on is just a matter of digression.

AMD are quite diverse now with high powered and low powered systems depending holistically on the end user, ie: if I want a day to day work notebook with some multimedia features I would buy a Tamash based system, being in the know I would not buy a Tamash if I was intending to use a notebook as say a gaming system, for that I would use the full A10 Elite Mobility setups with discrete solutions and in the same regard I would not be expecting the most efficient option nor would I give a crap about how long its battery lasts as I would be plugged into wall power. I won't game on a plane or other vehicle as I suffer from motion illness and get severe headaches doing that. Again its about the end user everything else matters not.

As for hearing over and over about how epic Intel is, why good for them my heart beats lumps of custard, maybe you should tell someone that cares. My A10 with its HD6970 is serving my needs just fine thanks.

 

[The Q6660 Inside]

Once and for all, can we keep the posts on topic? If you truly believe that Intel processors are 30 years ahead of AMD, fine that's your opinion. Go find an Intel fan page and brag about power usage and your vast supremacy for being better than everyone else and buying such a great Intel processor. You guys want a pat on the back for having Intel Inside? Well here it is:

Congratulations your processor is the greatest thing ever invented in the history of the universe and we should all start a religion to worship its grand and glorious design, and you must be a god yourself for having one..... Really I can't be anymore sarcastic than that, but I hope you get the point. Please go find an "Intel humpers" thread, this isn't the place to be posting about Intel's grand supremacy.

Now FOR THE LOVE OF ALL THAT IS HOLY does anyone have any Steamroller news? Any AMD news of any kind?

Some AMD news:

http://www.tomshardware.com/news/Rory-Reed-2Q-2013-Xbox-PLayStation-Temash,23597.html

http://www.tomshardware.com/news/AMD-Gaming-Evolved-Games-Logo-Partners-Titles,23560.html

I feel bad for amd now, I promise from now on I won't say anything bad within reason about them. I will for sure get the ps4 and see what it brings. I remember when the ps3 was close to release date and now the ps4 is coming out soon. Lets see what this next gen console architecture is made of, is it futureproof for next 6 years, unified 8gb ram check, new 8000 series architecture gpu check, 28nm gpu I wanted on next gen consoles check, cpu we don't know if it will work for future yet, lets see what developers say about it.

AMD steamroller will put amd back in the enthusiast market for sure, question is can they live to its hype this time and when will it be out. Early 2014 or even late 2013 and amd have a chance to sell well. I suppose without amd pc prices would still have been sky high like in the late 90's. Thats one good thing I have to say, you can only see this from intel targeting ultrabooks at low prices due to amd. In 1998 or was it 1999 we got a time machine computer package with p3 500mhz 128mb ram 20gb hdd really 16.1gb formatted, ATI rage 16mb graphics(top of the range recommended gpu back then for games), dvd rom, printer scanner digital camera 2.1mp fuji and epson stylus colour 300 printer 15" crt 1024x768 cost like just over a £1000 my first ever computer at home. At school they had slow acorn machines lol. I used supanet email system, my first ever email account back then and we had dial up. How times change. Now you get fast pc's dirt cheap.

Long live AMD and lets not forget intel can't be bothered to fix the heat spreader or whatever it is issue on ivy bridge or haswell. I mean at least amd cpus run cool, intel charge £250 or more for k series quad i7's and can't be bothered to fix the cooling!!! I mean there laptop cpus reach 90-100 degrees which is just unacceptable and its due to hyperthreading and some of it is trigate transistors and the main one cheap heat spreader.

I have to say I wanted to get the amd temash as a netbook type style computer as intel don't have this range of cpu and gpu working in tandem on low profile machines. AMD with there small budget and marketshare compete against intel and nvidia and still bring decent prices. I shall end it at that, I really want to see a leak of a proper mainstream steamroller engineering sample soon to see what it can do.

You know, that is the pure bitch of it. Intel needs AMD and AMD needs Intel for consumers to have a cost effective good product. If AMD would have never become a rival processor producer Intel chips would be much higher priced and we would probably be talking about that powerful new dual core as we wouldn't have quad cores. If Intel were to go out of business tomorrow AMD might keep its prices down for awhile, but with no competition they would jack their prices as well, there wouldn't be any incentive at R&D so advancements would stagnate. Both companies need each other to give us the best products at the best prices.

Congratulations, you've lost the game. You showed that you are worse troll than hafijur.

8350rocks is not a troll.

hafijur is here for trolling, and you have been helping him with your anti-AMD rants. Up to now your tactic to fuel the troll was very astute and almost inappreciable. Now that you moved to insulting several posters here, everything is easier.

http://www.zdnet.com/blog/ou/is-amd-being-hypocritical-about-benchmark-ethics/438
http://www.zdnet.com/blog/ou/strike-3-for-amd-hypocrisy-on-benchmarking/463

AMD go even further they use benchmarks they get spanked on mainly but 1 out of a few they win they post to show they are better lol. Another thing is they use there apu or cpu performance and compare it to an intel cpu+igpu to make it sound like they are so superior at gaming lol instead of cpu tasks as they compare the processors, its quite funny them showing cpus so low end and trying to make it sound like amd are so much better. Haswell now stops amd from showing these benchmarks like they used to all the time and now intel focused on igpu amd fanboys call it hasfail as amds worst nightmares came true with haswell. Its common nature for both companies to show there cpus in the best frame but amd mainly nowadays just boast about there graphics performance in there apus lol and forget the apu cpu especially in laptops get beaten by core 2 duos as an all round chip from 7 years ago.

The irony is amds higher end graphics cards only work properly at full performance with an intel cpu so without intel there 8970m with an a10-5750m would get less then half the performance in games. The i7 4700mq one of intel entry level i7 mobile hasfail cpus you like to say beats out the fx8350 on multithreaded tasks.

http://www.notebookcheck.net/AMD-Radeon-HD-8970M.90656.0.html
http://www.notebookcheck.net/Intel-Core-i7-4700MQ-Notebook-Processor.92959.0.html

TROLLING ALGORITHM:

while ( 3 > 1 )
{
  post nonsense;
  ignore corrections;
}

Calm down, he is dealing with it in a more constructive manner, he was just fed up by the Intel criticism. I must agree with him that AMD has to release a more competitive product against the Extreme Edition and 3930K processors that is not a binned 8350\Phenom II TWKR scheme and release something like a high-clocked, low-yield-ish 12-Core "12350".

 

[Krnt]

Some Richland Overclocking News:
http://www.techpowerup.com/187768/amd-a10-6800k-overclocked-to-8-2-ghz-breaks-and-sets-world-record.html

Seems like The Stilt got an slightly better processor.

 

[gamerk316]

GCC is the best compiler for Piledriver and Steamroller. BdverX was prepared by AMD (open source ).

No, ICC is about 50% faster then GCC for BD, and about 35% faster then MSVC:

http://www.behardware.com/articles/847-1/the-impact-of-compilers-on-x86-x64-cpu-architectures.html

GCC is actually the worst of ICC, MSVC, and GCC.

So it really comes down to this:

Can we blame the developers who succumb? This is the crux of the problem! If you offer developers the choice between using an equitable compiler against another that gives a 35% performance gain to AMD processor users and 54% to users of Intel processors, should we blame them for choosing the one that will improve the user experience in both cases (though in a partisan way)?

 

[Cazalan]

GCC is the best compiler for Piledriver and Steamroller. BdverX was prepared by AMD (open source ).

No, ICC is about 50% faster then GCC for BD, and about 35% faster then MSVC:

http://www.behardware.com/articles/847-1/the-impact-of-compilers-on-x86-x64-cpu-architectures.html

GCC is actually the worst of ICC, MSVC, and GCC.

A good read. People think optimization is a simple switch but it's incredibly complex. Unfortunately developers rarely get time to focus on optimization as just getting the software to run on the most platforms (Windows/Android/Linux/iOSx) is a far more important task.

Go to any game support forum and there will be hundreds/thousands of posts about not being able to run the game at all. Of course developers have to focus on those first.

If people want better AMD support out of the box then there needs to be some coalition/non-profit (Kickstarter) setup to fund more compiler support engineers. AMD has like 1/20th the budget of Intel when it comes to that sort of thing. It doesn't come for free.

 

[noob2222]

GCC is the best compiler for Piledriver and Steamroller. BdverX was prepared by AMD (open source ).

No, ICC is about 50% faster then GCC for BD, and about 35% faster then MSVC:

http://www.behardware.com/articles/847-1/the-impact-of-compilers-on-x86-x64-cpu-architectures.html

-cut img-

GCC is actually the worst of ICC, MSVC, and GCC.

So it really comes down to this:

Can we blame the developers who succumb? This is the crux of the problem! If you offer developers the choice between using an equitable compiler against another that gives a 35% performance gain to AMD processor users and 54% to users of Intel processors, should we blame them for choosing the one that will improve the user experience in both cases (though in a partisan way)?

The plot thickens when Intel then goes prospecting for software developers to offer them use of the Intel compiler, suggesting they use these Qax modes which, apart from being recommended by Intel, are detected as giving higher performance by those developers who take the time to check respective performance. In terms of transparency, one might find fault with Intel, particularly when developers of in vogue games or benchmarks are targeted as these are used in the specialised press to measure processor performance.

I thought no one used ICC.

Interesting article tho, still shows the disparity of ICC on an AMD cpu, in some cases speeding up intel cpus by +200% while at the same time speeding up AMD by 0%. Some tests on ICC actually cut amd performance up to 50% with certain dispatch versions via the infamous "adding of extra work for AMD cpus" (lbm being the most obvious)

 

[gamerk316]

GCC is the best compiler for Piledriver and Steamroller. BdverX was prepared by AMD (open source ).

No, ICC is about 50% faster then GCC for BD, and about 35% faster then MSVC:

http://www.behardware.com/articles/847-1/the-impact-of-compilers-on-x86-x64-cpu-architectures.html

-cut img-

GCC is actually the worst of ICC, MSVC, and GCC.

So it really comes down to this:

Can we blame the developers who succumb? This is the crux of the problem! If you offer developers the choice between using an equitable compiler against another that gives a 35% performance gain to AMD processor users and 54% to users of Intel processors, should we blame them for choosing the one that will improve the user experience in both cases (though in a partisan way)?

The plot thickens when Intel then goes prospecting for software developers to offer them use of the Intel compiler, suggesting they use these Qax modes which, apart from being recommended by Intel, are detected as giving higher performance by those developers who take the time to check respective performance. In terms of transparency, one might find fault with Intel, particularly when developers of in vogue games or benchmarks are targeted as these are used in the specialised press to measure processor performance.

I thought no one used ICC.

Interesting article tho, still shows the disparity of ICC on an AMD cpu, in some cases speeding up intel cpus by +200% while at the same time speeding up AMD by 0%. Some tests on ICC actually cut amd performance up to 50% with certain dispatch versions via the infamous "adding of extra work for AMD cpus" (lbm being the most obvious)

The performance drop for SSE4.1 is interesting due to BD officially supporting it. Clearly no bias in this case though, since the PIIX4 benefits. More likely a weird performance regression. [Its a known phenomina that sometimes compiling to a higher arch can reduce performance. Hence why SSE3 tends to be the default option for most compilers.]

Also

Note that ICC includes a third optimisation mode (arch:SSE2 for example) allowing it to create a build for a given level of functionality and not only a given Intel processor.

They used the arch: parameter for testing, so AMD chips should be getting the full optimization treatment in those tests. See the detailed test results for a more in depth explanation. They tested with and without the dispatcher. Interesting to note though, that AMD does benefit from SSE4 and above, even with the dispatcher in place, likely due to code gen differences.

And no, ICC typically isn't used due to cost. Performance makes it very attractive though.

 

[noob2222]

The performance drop for SSE4.1 is interesting due to BD officially supporting it. Clearly no bias in this case though, since the PIIX4 benefits. More likely a weird performance regression. [Its a known phenomina that sometimes compiling to a higher arch can reduce performance. Hence why SSE3 tends to be the default option for most compilers.]

PII x4 went from 258 sec to 304 with sse4.1 and as low as 185 with sse4.2 dispatch I wouldn't call that a benefit with no bias towards AMD.

PII was the same as BD, just not as obvious.

one thing is clear tho, with almost all these tests, the 8150 was faster than PII contrary to popular belief that PII is superior.

 

[jdwii]

Based on synthetic^ benchmarks

 

[noob2222]

lol, in other news, SA is not impressed with Intel's new marketig terminology.

http://semiaccurate.com/2013/07/25/intel-keeps-up-the-unethical-sdp-scam-with-new-4-5w-parts/

more importantly, richland hits 8.2 ghz.

http://www.techpowerup.com/187768/amd-a10-6800k-overclocked-to-8-2-ghz-breaks-and-sets-world-record.html

 

[The Q6660 Inside]

lol, in other news, SA is not impressed with Intel's new marketig terminology.

http://semiaccurate.com/2013/07/25/intel-keeps-up-the-unethical-sdp-scam-with-new-4-5w-parts/

Don't steer off topic, this is an AMD thread.

 

[jdwii]

lol, in other news, SA is not impressed with Intel's new marketig terminology.

http://semiaccurate.com/2013/07/25/intel-keeps-up-the-unethical-sdp-scam-with-new-4-5w-parts/

I'm still not sure this beats the best buy card thing however that was a classic.

 

[juanrga]

Besides DDR3 is already 2400MHz, that is, above bandwidth to now DDR4 2133, and way above in latency... nothing prevents AMD to commercialize DDR3 2400Mhz sticks... it could do plenty the trick for kaveri...

As said in a previous post DDR3 @ 2600MHz is already available. In fact my supplier sells DDR3 modules @ 2666MHz.

GCC is the best compiler for Piledriver and Steamroller. BdverX was prepared by AMD (open source ).

No, ICC is about 50% faster then GCC for BD, and about 35% faster then MSVC:

http://www.behardware.com/articles/847-1/the-impact-of-compilers-on-x86-x64-cpu-architectures.html

GCC is actually the worst of ICC, MSVC, and GCC.

For Intel CPUs it is more like

ICC >~ GCC >> MSVC

For AMD chips, ICC uses the Cripple_AMD function to reduce performance.

Benchmarks and further info:

http://www.linuxforge.net/docs/bm/bench-gcc-icc.php

http://stackoverflow.com/questions/1733627/anyone-here-has-benchmarked-intel-c-compiler-and-gcc

 

[mayankleoboy1]

GCC is the best compiler for Piledriver and Steamroller. BdverX was prepared by AMD (open source ).

No, ICC is about 50% faster then GCC for BD, and about 35% faster then MSVC:

http://www.behardware.com/articles/847-1/the-impact-of-compilers-on-x86-x64-cpu-architectures.html

GCC is actually the worst of ICC, MSVC, and GCC.

This test is using old compilers. Specially GCC has really picked up speed in the 4.7 version. MSVC has also some improvements in auto-vectorization in MSVC2012. So difficult to say how the current compilers compare. Plus the tests themselves use old code.

(BTW, this is the 3rd or 4th time this article is being cited here, to prove the same arguements. This thread is moving in circles. )

 

[noob2222]

^^ true, especially the worst test - c-ray.

30% from 4.6.3 to 4.7.0 (the behardware used 4.6.1)

current version is 4.8.0 bringing another 24%

 

[mlscrow]

Anyone have any idea when AMD will reveal their Desktop Roadmap for 2014? As skeptical as I am, I am hoping that they will reveal a 4m desktop CPU/APU with SR cores. If they do not, they are pushing some of their remaining loyalists like myself to the other side. I will pick up a 4770K if there is no 4m SR for 2014.

 

[griptwister]

Hmmm... cool little article I found not steamroller, but it's pretty impressive!

http://blogs.windows.com/windows/b/extremewindows/archive/2013/07/25/pushing-the-12k-pc-gaming-boundary-at-1-5-billion-pixels-per-second.aspx

 

[Cazalan]

Anyone have any idea when AMD will reveal their Desktop Roadmap for 2014? As skeptical as I am, I am hoping that they will reveal a 4m desktop CPU/APU with SR cores. If they do not, they are pushing some of their remaining loyalists like myself to the other side. I will pick up a 4770K if there is no 4m SR for 2014.

I'm not sure they have a Desktop Roadmap anymore. They have client (mobile/desktop) and server roadmaps.

The server roadmap leaves some open questions itself. The APUs already got a Richland update but for the servers they are still talking Piledriver "Warsaw" (12/16) at the high end. Also the 4/6/8 core server CPUs get replaced with 28nm Steamroller 4C APU.

AMD shed 15% of their employees last year. Adding a new ARM processor likely ate at their higher end desktop area.

 

[8350rocks]

I think they are still "tinkering" with the FX line for steamroller, hence no obvious news about SR powered Opterons in a format other than APUs. Plus, they may not go bigger than 8 core with SR, and the 12/16 core Opterons on PD architecture are holding their own against the current crop of Xeons.

Interesting bit about windows from S|A:

http://semiaccurate.com/2013/07/22/microsoft-drove-the-bus-off-the-cliff-now-it-tries-to-speed-up/

 

[mlscrow]

Anyone have any idea when AMD will reveal their Desktop Roadmap for 2014? As skeptical as I am, I am hoping that they will reveal a 4m desktop CPU/APU with SR cores. If they do not, they are pushing some of their remaining loyalists like myself to the other side. I will pick up a 4770K if there is no 4m SR for 2014.

I'm not sure they have a Desktop Roadmap anymore. They have client (mobile/desktop) and server roadmaps.

The server roadmap leaves some open questions itself. The APUs already got a Richland update but for the servers they are still talking Piledriver "Warsaw" (12/16) at the high end. Also the 4/6/8 core server CPUs get replaced with 28nm Steamroller 4C APU.

AMD shed 15% of their employees last year. Adding a new ARM processor likely ate at their higher end desktop area.

Yeah, I've been aware of the server roadmap since it was released, but thanks anyway. This server roadmap is why I say I'm skeptical that they will reveal a 4M desktop part, because usually their FX line is based on 1P server offerings and since what we see for 2014 for 1P is only a 2M part, it seems unlikely that we will have a 4M part.

I think they are still "tinkering" with the FX line for steamroller, hence no obvious news about SR powered Opterons in a format other than APUs. Plus, they may not go bigger than 8 core with SR, and the 12/16 core Opterons on PD architecture are holding their own against the current crop of Xeons.

Interesting bit about windows from S|A:

http://semiaccurate.com/2013/07/22/microsoft-drove-the-bus-off-the-cliff-now-it-tries-to-speed-up/

What info do you have that makes you think they are still tinkering with their FX line as well as thinking that they are even going to offer an 8 core Steamroller part? I've been looking all over for any information regarding an 8 core Steamroller part and I can't find anything, nor about AMD working on the next version of their FX, outside of the 9590 of course, but that's already been released. All I have is that they are offering a dual core Steamroller part in Kaveri, nothing more. It's upsetting to say the least.

 

[hcl123]

Some Richland Overclocking News:
http://www.techpowerup.com/187768/amd-a10-6800k-overclocked-to-8-2-ghz-breaks-and-sets-world-record.html

Seems like The Stilt got an slightly better processor.

The better "The Stilt" got 5.6 Ghz on water, and 6.3 Ghz on a "vapo" system... this is usable, LN2 is not.

also "average" power for stock clocks is below the 100W

http://www.xtremesystems.org/forums/showthread.php?286367-Richland-the-best-OC-APU-infos-experience-issues

 

[hcl123]

Besides DDR3 is already 2400MHz, that is, above bandwidth to now DDR4 2133, and way above in latency... nothing prevents AMD to commercialize DDR3 2400Mhz sticks... it could do plenty the trick for kaveri...

As said in a previous post DDR3 @ 2600MHz is already available. In fact my supplier sells DDR3 modules @ 2666MHz.

Matter of fact TeamGroup has DDR3 at 2800Mhz for "more than a year" now... and better was/is passively cooled and 1.65v.

http://www.teamgroup.com.tw/filterable_product/tabs_detail/data/en/9/600/zfKsFi.html

there are several others... G.Skill and Corsair i think... with DDR 3000Mhz now (at least Corsair is not passively cooled... i think)

this is the memory for those APUs, no real need for DDR4 for now.

The problem is "certification", none of those speeds are JEDEC (or other) certified. Max JEDEC is 2133Mhz (they are at sleep lol). I mentioned DDR3 2400Mhz because it has a good chance of being certified at 1.5v, the original voltage spec for DDR3... DDR3 2667Mhz i doubt it will ever be 1.5v.

For a vendor to sell certified products is half way to mass adoption, makes everything easier for other IDMs, mobo to components etc. DDR3 2400Mhz could be that new product, certified, fulfilling a hole while DDR4 does not push above 2667Mhz(theoretical max possible 4233Mhz) .