AMD CPU speculation... and expert conjecture

[Cazalan]

The relevant bullet: Blazing fast CPU not necessary, a mid-range CPU will be sufficiently fast to drive the best GPU

I find it fascinating how they showed a concrete example where MANTLE improved performance by 3x (200%) and still gamerk pretends that the DX overhead is at best of 10% :ROFL:

Marketing demos are hardly concrete examples of anything. BF4 Mantle edition needs to come out and then we'll see.

 

[Cazalan]

The ultimate in VR.

Mantle + Oculus Rift + Novint Falcon + Tenga

Lol!

http://www.tomshardware.com/news/oculus-rift-sex-simulator,25082.html

 

[noob2222]

"PC gamers and developers deserve the benefits of this model as well, which is why devs like DICE approached us and requested a technology like Mantle

http://www.tomshardware.com/news/amd-mantle-api-xbox-one-directx-11.2,24691.html

Just in case anyone ever wondered why mantle is surrounded by bf4 promos?

 

[palladin9479]

The FX-8350 to Kaveri comparisons are getting tired. They're different products, made on different processes, with significant TDP differences, for different market segments.

I still don't understand why people are even humoring a comparison. The fx8350 destroys any APU by virtue of physics. Instantly dismiss any attempt at using FLOPS, scalar CPU's don't do FLOPS vector ones do. Any benchmark that shows some large FLOPS number is actually begin run on an iGPU. So what you do is run it against something comparable like 8350 + 7970. Scalar CPU's use dhrystone as an appropriate benchmark for integer computing power (the VAST majority of work done). Whetstone is the benchmark used to determine a CPU's (not GPU) ability to do SIMD / FPU work.

Very important to separate performance of the scalar and vector components as really big vector numbers don't translate into actual real world performance.

 

[juanrga]

The relevant bullet: Blazing fast CPU not necessary, a mid-range CPU will be sufficiently fast to drive the best GPU

I find it fascinating how they showed a concrete example where MANTLE improved performance by 3x (200%) and still gamerk pretends that the DX overhead is at best of 10% :ROFL:

Marketing demos are hardly concrete examples of anything. BF4 Mantle edition needs to come out and then we'll see.

But it is not a marketing demo to average Joe but a technical talk to developers. During the talk, it was shown with a concrete example (aka code) that the overhead can be reduced by one order of magnitude. The hardware with MANTLE was able to compute 10x more draw calls per second... whereas gamerk said that such numbers would be impossible.

About BF4 I already did my prediction here. I repeat it again: I expect the MANTLE version of BF4 to be 30--50% faster than the DX version.

 

[de5_Roy]

Spoiler

then he kept repeatedly bringing it up saying that benchmark is the reason amd cancelled higher core SR-FX cpus and apparently that's what amd 'told' oem representatives at that event (even thought he failed to provide an audio transcript after repeated asking).

This is pure and simply false (*). The reasons why AMD is not releasing SR FX line are multiple and I explained them here before: transition to APU, reorganization of server plans, and lack of demand. I gave details and further explanations for each one.

oh really? false, is it? here's what you said in your post, ad verbatim:

In their talk they were saying to OEMs that the new APU gives >90% of the FX-8350 performance at one fraction of the cost. That is what the slide #13 says. That doesn't look as promoting the FX line. The FX-4350 didn't even was mentioned by AMD.

I see two options either (i) AMD will be abandoning the FX-4000 and refreshing the FX-6000/8000/9000 a la Warsaw or (ii) AMD will drop the entire FX line completely and focus on an APU line only.

I am not assuming anything, I asked you a specific question.

but wait.. there is more, in case you have conveniently forgotten (along with the poorly executed deflection when i ask for audio version):

In their talk they were saying to OEMs that the new APU gives >90% of the FX-8350 performance at one fraction of the cost. That is what the slide #13 says. That doesn't look as promoting the FX line. The FX-4350 didn't even was mentioned by AMD.

talk? there's an audio version? i didn't listen to that. if amd said that the new apu(6790k) gives >90% of fx8350's performance, they were intentionally being vague (>90% of what? what tasks?) to pitch 6790k. benchmarketing at play - largely irrelevant, since independent reviews will reflect the real world perf/price.

^^ bolded the relevant part.

That part that you bolded is "That is what the slide #13 says."

I am saying which is the message given by the slide. Not that the slide is the reason for AMD plans. You got it backwards.

The reasons why AMD is abandoning FX line (and not releasing improvement or refresh) were given by my before:

The reasons why AMD is not releasing SR FX line are multiple and I explained them here before: transition to APU, reorganization of server plans, and lack of demand. I gave details and further explanations for each one.

There is another post here who is also explaining why AM3+ is a dead platform. I suppose he is not being read as well.

in your messed-up reply post, i cannot see anything you said except the first line. please re-edit and re-post properly. frm what i can dig out of that mess:
you claim some message is given by the slide #13, but never explain how you got it. the 'explanation' you gave do not relate to the slide at all. the slide only shows a gpu-bound game benchmark, nothing about your claims.

btw, you are wrong on both of your 'amd (i), (ii)options' in your claim: as long as vishera exists, amd will continue selling lower binned, dual module vishera as fx4000 and they are not abandoning fx entirely for apus. they're simply carrying on the vishera lineup for now, according to the latest desktop roadmap.
http://farm3.staticflickr.com/2836/10861225406_8dbbd0dc46_o.png

but here's the real kicker at the bottom of the slide:

amd roadmaps are subject to change without notice or obligations to notify of change. placement of boxes intended to represent first year of production shipments.

this last line completely contradicts your earlier claim about "side by side boxes on a roadmap mean product replacement". rofl. if amd meant replacement, they woulda said so. the boxes only mean when new products are supposed to start shipping, and even that is subject to change without notice i.e. nothing is set in stone. now i know why some sites crop part of promo slides.

 

[Cazalan]

But it is not a marketing demo to average Joe but a technical talk to developers. During the talk, it was shown with a concrete example (aka code) that the overhead can be reduced by one order of magnitude. The hardware with MANTLE was able to compute 10x more draw calls per second... whereas gamerk said that such numbers would be impossible.

Unless the exact code was released to the public and independently verified it means diddly squat. Any number of tricks or omissions of optimizations could be used behind the scenes to make one method appear faster than another. They're trying to create buzz and sell the idea of Mantle to the developer community. Of course they will have some nice canned demos.

I've written hundreds of thousands of lines of code. You can't tell much from a few code snippets they put up on a slide at a conference. It's just 1 piece of a giant puzzle.

 

[juanrga]

you claim some message is given by the slide #13, but never explain how you got it. the 'explanation' you gave do not relate to the slide at all. the slide only shows a gpu-bound game benchmark, nothing about your claims.

I got the message from reading the slide, also between lines.

btw, you are wrong on both of your 'amd (i), (ii)options' in your claim: as long as vishera exists, amd will continue selling lower binned, dual module vishera as fx4000 and they are not abandoning fx entirely for apus. they're simply carrying on the vishera lineup for now, according to the latest desktop roadmap.
http://farm3.staticflickr.com/2836/10861225406_8dbbd0dc46_o.png

Just in the last page I said that I already have a copy of the roadmap. This proves again that you don't read the thread.

Yes, my hypothesis (i) and (ii) were wrong, but infinitely close to reality that all those who speculated about SR FX, Phenom return, 10-core CPUs, AM4 sockets, FX replacement...

My hypothesis (ii) was that AMD abandons the FX-4000 series and release a refresh of the 6000/8000/9000 FX series a la Warsaw for 2014. But the sales of the FX chips must be so low that AMD is not even releasing such refresh. They must be selling existences.

And yes, they are abandoning FX by APUs. That is why the APU line receives Steamroller update and many more improvements, but the FX-line does not.

but here's the real kicker at the bottom of the slide:

amd roadmaps are subject to change without notice or obligations to notify of change. placement of boxes intended to represent first year of production shipments.

this last line completely contradicts your earlier claim about "side by side boxes on a roadmap mean product replacement". rofl. if amd meant replacement, they woulda said so. the boxes only mean when new products are supposed to start shipping, and even that is subject to change without notice i.e. nothing is set in stone. now i know why some sites crop part of promo slides.

Every roadmap is subject to change. You are not saying anything new. If tomorrow the foundry where Kaveri is being fabricated explodes and all the production is lost, that legal note saves AMD against investors and others, because the roadmap would be delayed. At contrary, if tomorrow Intel surprises everyone with an accelerated roadmap and releases Broadwell in two months, AMD would also change its roadmap to reply that challenge, probably by accelerating Carrizo arriving. And so on.

The other part of the note is not even related to what you say. You are completely misreading this one. The "first year of production shipments" means, for instance, that Kaveri will start to ship in late 2013, which means that you will have to wait up to early 2014 to purchase one.

Of course products in boxes at the right are replacing products in boxes to the left:

Radeon 200 series replace HD 8000 series
Kaveri replaces Richland
Beema replaces Kabini
Mullins replaces Temash
Warsaw replaces Opteron 6300/4300
Berlin replaces Opteron 3300
Seattle replaces Opteron-X
Bald Eagle and Hierofalcon replaces embedded R series
Steppe Eagle replaces embedded G series

And so on

But it is not a marketing demo to average Joe but a technical talk to developers. During the talk, it was shown with a concrete example (aka code) that the overhead can be reduced by one order of magnitude. The hardware with MANTLE was able to compute 10x more draw calls per second... whereas gamerk said that such numbers would be impossible.

Unless the exact code was released to the public and independently verified it means diddly squat. Any number of tricks or omissions of optimizations could be used behind the scenes to make one method appear faster than another. They're trying to create buzz and sell the idea of Mantle to the developer community. Of course they will have some nice canned demos.

I've written hundreds of thousands of lines of code. You can't tell much from a few code snippets they put up on a slide at a conference. It's just 1 piece of a giant puzzle.

Basically you are suggesting that the guy who gave the talk was lying. Care to repeat your accusation using your real name?

Also if you are a developer and you are familiar with this, you would know that every game developer knows that the overhead is 10x (or even higher) and many of them report this number. I recall Carmack complaining about the 10x overhead in DX several years ago, in public, before anyone mentioned a single word about MANTLE. Do you?

 

[-Fran-]

Thinking on what AMD is doing with HSA, it makes sense to me to delay the FX successor line for at least until they got things sorted out for the APU line. It is the most reasonable thing to do as a company: chase the market you're strong in, position yourself (secure the market) and then go into the rest of your weak points.

AMD doesn't have the cash to waste going onto markets they don't have anything competitive on right now. Some will argue about server landscape, but perf/power of 32nm and even 28nm Bulk is not competitive to Intel's no matter how we spin it. AMD is making a bet on instead of chasing Intel vertically, they'll do it horizontally (ironically, HSA is actually the MOAR COARS approach, haha). Making a good consumer chip and moving that expertise to Server landscape is very bright.

In particular, gamerk, Java is HUGE. You can't even start imagining how many corporations use it. Starting with IBM and Oracle packing it with ALL of their enterprise solutions. AIX + Crapsphere + IBM's JDK, RedHat + JBoss + Oracle's JDK and of course Oracle + Weblogic + Oracle's JDK. AMD pushing HSA standards into Java is going to be a GREAT cash in for server market if they pull it of.

I know from experience I can make JBoss thread up to a ridiculous amount of hard threads (up to 10 per client on a 100 client concurrent system) that each can eat up a CPU, plus the bloody garbage collector running. I can totally picture an APU taking advantage of certain workloads. It would require a little tunning, but nothing too big. Same with ETL's.

From all the HSA partners or initiatives, I think Java is the most important one.

Cheers!

 

[de5_Roy]

Intel forced AMD to adopt SDP metric
http://www.fudzilla.com/home/item/33149-intel-forced-amd-to-adopt-sdp-metric

not directly related to amd but..
asrock unveils 2 mobos for bitcoin mining
http://www.techpowerup.com/194548/asrock-unveils-two-motherboards-made-for-bitcoin-mining.html
they could easily repurpose an amd 990fx chipset mobo for that. and they're too late.

 

[de5_Roy]

I got the message from reading the slide, also between lines.

"reading between the lines" is not an explanation. it means you admit to using semantics instead of facts and then stretch it out.

Just in the last page I said that I already have a copy of the roadmap. This proves again that you don't read the thread.

lawl. this proves that i don't care if you had a copy or posted if you had a copy. i am using it for my references. it doesn't prove i don't read the thread.

Yes, my hypothesis (i) and (ii) were wrong,

i'll leave it here, reference it later.

but infinitely close to reality that all those who speculated about SR FX, Phenom return, 10-core CPUs, AM4 sockets, FX replacement...

amd doesn't comment on future products. you realize that you're arguing speculations, right? don't you? if you do, you're already contradicting the 'reality' part of 'infinitely close to reality'.

My hypothesis (ii) was that AMD abandons the FX-4000 series and release a refresh of the 6000/8000/9000 FX series a la Warsaw for 2014.But the sales of the FX chips must be so low that AMD is not even releasing such refresh. They must be selling existences.

amd (and others) abandons silicon only when it fails lowest bin. the dual modules that passes binning will continue to be sold.

And yes, they are abandoning FX by APUs. That is why the APU line receives Steamroller update and many more improvements, but the FX-line does not.

the current roadmap(something you keep swearing by) disagrees with you. here it is one more time:
http://farm3.staticflickr.com/2836/10861225406_8dbbd0dc46_o.png
looks like you can't read it even when you claimed in the last page that you have a copy.:lol:
and your reasons are mere speculations. not worth arguing.
oh wait, you already admitted your hypotheses(sure looked like fact-based claims when you argued... 😛) were wrong. my bad.

Every roadmap is subject to change. You are not saying anything new.

of course not. i am using it to argue your claims. did i say i was saying something new? don't put words in my posts.

Spoiler

If tomorrow the foundry where Kaveri is being fabricated explodes and all the production is lost, that legal note saves AMD against investors and others, because the roadmap would be delayed. At contrary, if tomorrow Intel surprises everyone with an accelerated roadmap and releases Broadwell in two months, AMD would also change its roadmap to reply that challenge, probably by accelerating Carrizo arriving. And so on.

unrelated

my original text:

Spoiler

but here's the real kicker at the bottom of the slide:

amd roadmaps are subject to change without notice or obligations to notify of change. placement of boxes intended to represent first year of production shipments.

this last line completely contradicts your earlier claim about "side by side boxes on a roadmap mean product replacement". rofl. if amd meant replacement, they woulda said so. the boxes only mean when new products are supposed to start shipping, and even that is subject to change without notice i.e. nothing is set in stone. now i know why some sites crop part of promo slides.

The other part of the note is not even related to what you say. You are completely misreading this one. The "first year of production shipments" means, for instance, that Kaveri will start to ship in late 2013, which means that you will have to wait up to early 2014 to purchase one.

unrelated? misreading? read my post And your own reply. you're not disagreeing with me. except the 'unrelated' and 'misreading' parts - those, you got wrong. what i said are related.

Of course products in boxes at the right are replacing products in boxes to the left:

Spoiler

Radeon 200 series replace HD 8000 series
Kaveri replaces Richland
Beema replaces Kabini
Mullins replaces Temash
Warsaw replaces Opteron 6300/4300
Berlin replaces Opteron 3300
Seattle replaces Opteron-X
Bald Eagle and Hierofalcon replaces embedded R series
Steppe Eagle replaces embedded G series

And so on

wow, so many wrongs (or lies? rhetorical. never mind.).
hd8k is oem (some of those were existing gcn 1.0 gpus iirc). majority current r200 are just existing asics with minor clockrate tweaks except bonair and hawaii based gpus. bonair didn't replace pitcairn nor cape verde, hawaii didn't replace tahiti.
kaveri is still not out. neither is beema, mullins, warsaw or other future products.(quick note: you also claimed that amd Is replacing temash in tablets with an arm soc (in reality, mullins, temash's Successor, has x86 puma cores. says on the mobile roadmap. i hope you had a copy of that.).)
as a comparison, both zambezi and vishera cpus are on sale, even though desktop roadmap showed the boxes side by side. same for socket fm1 and fm2 apus/cpus. replacement depends on the company and the product. "succeed" and "replace" are two different things.
here's a case of real replacement - radeon 6950, 6930, 6970(cayman) all were replaced by gcn gpus.

 

[juanrga]

Thinking on what AMD is doing with HSA, it makes sense to me to delay the FX successor line for at least until they got things sorted out for the APU line. It is the most reasonable thing to do as a company: chase the market you're strong in, position yourself (secure the market) and then go into the rest of your weak points.

AMD doesn't have the cash to waste going onto markets they don't have anything competitive on right now. Some will argue about server landscape, but perf/power of 32nm and even 28nm Bulk is not competitive to Intel's no matter how we spin it. AMD is making a bet on instead of chasing Intel vertically, they'll do it horizontally (ironically, HSA is actually the MOAR COARS approach, haha). Making a good consumer chip and moving that expertise to Server landscape is very bright.

In particular, gamerk, Java is HUGE. You can't even start imagining how many corporations use it. Starting with IBM and Oracle packing it with ALL of their enterprise solutions. AIX + Crapsphere + IBM's JDK, RedHat + JBoss + Oracle's JDK and of course Oracle + Weblogic + Oracle's JDK. AMD pushing HSA standards into Java is going to be a GREAT cash in for server market if they pull it of.

I know from experience I can make JBoss thread up to a ridiculous amount of hard threads (up to 10 per client on a 100 client concurrent system) that each can eat up a CPU, plus the bloody garbage collector running. I can totally picture an APU taking advantage of certain workloads. It would require a little tunning, but nothing too big. Same with ETL's.

From all the HSA partners or initiatives, I think Java is the most important one.

Cheers!

Another excellent post.

 

[tracker45]

There are no demonstrations of Mantle ??? What are you referring to ?

 

[juanrga]

I got the message from reading the slide, also between lines.

"reading between the lines" is not an explanation. it means you admit to using semantics instead of facts and then stretch it out.

And still my reading of that infamous slide #13 was accurate, with APU 13 confirming several of my claims about APU competing with FX-8 and i7 in gaming performance

Just in the last page I said that I already have a copy of the roadmap. This proves again that you don't read the thread.

lawl. this proves that i don't care if you had a copy or posted if you had a copy. i am using it for my references. it doesn't prove i don't read the thread.

When you are offering me something that I said that I have, you are proving that you are not reading.

Yes, my hypothesis (i) and (ii) were wrong,

i'll leave it here, reference it later.

but infinitely close to reality that all those who speculated about SR FX, Phenom return, 10-core CPUs, AM4 sockets, FX replacement...

amd doesn't comment on future products. you realize that you're arguing speculations, right? don't you? if you do, you're already contradicting the 'reality' part of 'infinitely close to reality'.

Nope. we are all discussing what AMD would release in 2014 and all people here who was waiting SR FX, Phenom return, 10-core CPUs, AM4 sockets, FX replacement... was disproved by the 2014 roadmap. This is reality 1.0

I recall when I said "there is no 10-core FX, no phenom return and no SR FX for 2014" and someone said me "wait for the roadmap". Well the roadmap is here and he is now silent.

My hypothesis (ii) was that AMD abandons the FX-4000 series and release a refresh of the 6000/8000/9000 FX series a la Warsaw for 2014.But the sales of the FX chips must be so low that AMD is not even releasing such refresh. They must be selling existences.

amd (and others) abandons silicon only when it fails lowest bin. the dual modules that passes binning will continue to be sold.

In servers, Warsaw only comes as 12 and 16 cores. There is no 2/4/6/8 core versions. This means that dual-module and single-module dies (passing binding) are not reused for lower core Warsaw versions, but trashed due to lack of demand.

And yes, they are abandoning FX by APUs. That is why the APU line receives Steamroller update and many more improvements, but the FX-line does not.

the current roadmap(something you keep swearing by) disagrees with you. here it is one more time:
http://farm3.staticflickr.com/2836/10861225406_8dbbd0dc46_o.png
looks like you can't read it even when you claimed in the last page that you have a copy.:lol:

At contrary, the roadmap reflects accurately what I am saying. It was also confirmed by Lisa Su (AMD Vicepresident) during opening keynotes this week. APU are the center of universe for AMD. AMD plans to dominate the market from phones to servers using APUs. FX and Opteron are EoL products.

Of course products in boxes at the right are replacing products in boxes to the left:

Spoiler

Radeon 200 series replace HD 8000 series
Kaveri replaces Richland
Beema replaces Kabini
Mullins replaces Temash
Warsaw replaces Opteron 6300/4300
Berlin replaces Opteron 3300
Seattle replaces Opteron-X
Bald Eagle and Hierofalcon replaces embedded R series
Steppe Eagle replaces embedded G series

And so on

wow, so many wrongs (or lies? rhetorical. never mind.).
hd8k is oem (some of those were existing gcn 1.0 gpus iirc). majority current r200 are just existing asics with minor clockrate tweaks except bonair and hawaii based gpus. bonair didn't replace pitcairn nor cape verde, hawaii didn't replace tahiti.
kaveri is still not out. neither is beema, mullins, warsaw or other future products.(quick note: you also claimed that amd Is replacing temash in tablets with an arm soc (in reality, mullins, temash's Successor, has x86 puma cores. says on the mobile roadmap. i hope you had a copy of that.).)
as a comparison, both zambezi and vishera cpus are on sale, even though desktop roadmap showed the boxes side by side. same for socket fm1 and fm2 apus/cpus. replacement depends on the company and the product. "succeed" and "replace" are two different things.
here's a case of real replacement - radeon 6950, 6930, 6970(cayman) all were replaced by gcn gpus.

😆 That is the 2013-2014 roadmap. Ok? The new products: Kaveri, Beema, Mullins, Warsaw, Seattle... are 2014 products. They will replace the older products the next year, no today when the roadmap has been presented.

I was not alone on saying that AMD will be releasing ARM tablets. It has been reproduced in several sites like this

http://techreport.com/news/25461/report-amd-to-introduce-arm-based-tablet-chip-this-year

You finding both zambezi and vishera cpus on sale, doesn't mean what you pretend. AMD is no more fabricating Zambezi, because it was replaced by Vishera. What happens is that some stores have inventory of the old processors. And continue selling them, up to finishing it.

The same happens with any other AMD product. Kaveri replacing Richland doesn't imply you cannot find Richland APus in your favourite store during some time, before the inventory finish.

 

[de5_Roy]

And still my reading of that infamous slide was accurate, with APU 13 confirming several of my claims about APU/CPU performance

no it was not. you admitting of using semantics instead of facts simply confirmed my accusations, whether you realized it or not. i might not require your calculations and results after all.

what claims?

When you are offering me something that I said that I have, you are proving that you are not reading.

let me make this perfectly clear: a lot of forum posters here at toms post, asking for help with getting into their laptop/phones/pcs/whatever device since they lost their password. they use forum ids, email address, cell numbers etc. that look quite legit. i will tell you the basically same thing i tell them: "there is no way to legitimately verify your claim of having the roadmap." woulda been far easier to verify if you posted it right away (claiming nda won't hold water). i know you didn't post anything, i read the thread 😆. since i know now that you misinterpret, misread, mislead, lie(your own admission), i just don't believe you.

Spoiler

Nope. we are all discussing what AMD would release in 2014 and all people here who was waiting SR FX, Phenom return, 10-core CPUs, AM4 sockets, FX replacement... was disproved by the 2014 roadmap. This is reality 1.0
I recall when I said "there is no 10-core FX, no phenom return and no SR FX for 2014" and someone said me "wait for the roadmap". Well the roadmap is here and he is now silent.

ah reality... here's a reality check for you: those people were speculating, unlike you. you are arguing speculations and conjecture about future products that may or may not exist.

In servers, Warsaw only comes as 12 and 16 cores. There is no 2/4/6/8 core versions. This means that dual-module and single-module dies (passing binding) are not reused for lower core Warsaw versions, but trashed due to lack of demand.

now there're servers too? i wonder what else did the slide#13 didn't address...
you seem to have a really hard time understanding the binning process. as long as amd makes a 'big' die with 4 modules, 2/4/6/8 core cpus will exist. 12 and 16 cores are based on 'bigger' mcm [strike]die[/strike] package, 12 core variant being lower binned. technically, amd can make a 10 core fx/opteron out of those by binning 4+1 modules or 3+2 modules. they don't do it for economical reasons and for a 10 core cpu like that isn't worth using the big mcm package.

did amd say warsaw will be a monolithic 16 core cpu? if they did, i want to see an official proof. afaik, their 'bigger' cpu is still 2x 4M mcm package.

there.is.no.dual.or.single.module.die.for.cpus. those are all binned from the 'big' die. only apus have dual module but those dies also contain igpu, u.n.b. etc.

lack of demand? of nonexistent silicon? yeah....

At contrary, the roadmap reflects accurately what I am saying. It was also confirmed by Lisa Su (AMD Vicepresident) during opening keynotes this week. APU are the center of universe for AMD. AMD plans to dominate the market from phones to servers using APUs. FX and Opteron are EoL products.

while it is true that amd is focussing on apus, your claim stays unverified until you show where lisa su says amd is abandoning fx, in "during opening keynotes this week".
i also want to see an official confirmation of fx and opterons being e.o.l. although i know for a fact that you are lying about opterons being e.o.l. kyoto has opteron branding, remember? warsaw will have opteron brand as well.

😆 That is the 2013-2014 roadmap. Ok? The new products: Kaveri, Beema, Mullins, Warsaw, Seattle... are 2014 products. They will replace the older products the next year, no today when the roadmap has been presented.

oh, you mean the 2014 roadmap? y'know, the one that says(like very other roadmap) that roadmaps only indicate first year of product shipment and Nothing about replacement?

It was not alone on saying that AMD will be replacing Temash by ARM SoC. It has been reproduced in several sites like this
http://techreport.com/news/25461/report-amd-to-introduce-arm-based-tablet-chip-this-year

somehow i knew you'd resort to that article. please, read the article carefully and show me where amd says they will Replace Temash with ARM soc. the techreport article(source:sweclockers) says nothing on replacing temash. even the sweclockers source says nothing about replacing temash. i checked that article using google translate. there's always a chance that i missed something, so you will have to provide proof of your own claim.

Spoiler

You finding both zambezi and vishera cpus on sale, doesn't mean what you pretend. AMD is no more fabricating Zambezi, because it was replaced by Vishera. What happens is that some stores have inventory of the old processors. And continue selling them, up to finishing it.
The same happens with any other AMD product. Kaveri replacing Richland doesn't imply you cannot find Richland APus in your favourite store during some time, before the inventory finish.

now you're switching to fabricating, eh? your entire tirade about replacement was based on roadmaps that only indicate intended shipment timeline, not fabrication. fabrication begins earlier. and it does not apply to seattle and kyoto which address the same market yet use different uarches.
i was writing from customer's p.o.v.(that's why i mention sales). yes, amd might not be fabricating zambezi and soon richland anymore but in the market you can choose from both.

 

[8350rocks]

The FX-8350 to Kaveri comparisons are getting tired. They're different products, made on different processes, with significant TDP differences, for different market segments.

I still don't understand why people are even humoring a comparison. The fx8350 destroys any APU by virtue of physics. Instantly dismiss any attempt at using FLOPS, scalar CPU's don't do FLOPS vector ones do. Any benchmark that shows some large FLOPS number is actually begin run on an iGPU. So what you do is run it against something comparable like 8350 + 7970. Scalar CPU's use dhrystone as an appropriate benchmark for integer computing power (the VAST majority of work done). Whetstone is the benchmark used to determine a CPU's (not GPU) ability to do SIMD / FPU work.

Very important to separate performance of the scalar and vector components as really big vector numbers don't translate into actual real world performance.

+1 Perhaps with you telling him this, instead of me...he will listen.

Performance AIO != HEDT

 

[8350rocks]

Spoiler

you claim some message is given by the slide #13, but never explain how you got it. the 'explanation' you gave do not relate to the slide at all. the slide only shows a gpu-bound game benchmark, nothing about your claims.

I got the message from reading the slide, also between lines.

btw, you are wrong on both of your 'amd (i), (ii)options' in your claim: as long as vishera exists, amd will continue selling lower binned, dual module vishera as fx4000 and they are not abandoning fx entirely for apus. they're simply carrying on the vishera lineup for now, according to the latest desktop roadmap.
http://farm3.staticflickr.com/2836/10861225406_8dbbd0dc46_o.png

Just in the last page I said that I already have a copy of the roadmap. This proves again that you don't read the thread.

Yes, my hypothesis (i) and (ii) were wrong, but infinitely close to reality that all those who speculated about SR FX, Phenom return, 10-core CPUs, AM4 sockets, FX replacement...

My hypothesis (ii) was that AMD abandons the FX-4000 series and release a refresh of the 6000/8000/9000 FX series a la Warsaw for 2014. But the sales of the FX chips must be so low that AMD is not even releasing such refresh. They must be selling existences.

And yes, they are abandoning FX by APUs. That is why the APU line receives Steamroller update and many more improvements, but the FX-line does not.

but here's the real kicker at the bottom of the slide:

amd roadmaps are subject to change without notice or obligations to notify of change. placement of boxes intended to represent first year of production shipments.

this last line completely contradicts your earlier claim about "side by side boxes on a roadmap mean product replacement". rofl. if amd meant replacement, they woulda said so. the boxes only mean when new products are supposed to start shipping, and even that is subject to change without notice i.e. nothing is set in stone. now i know why some sites crop part of promo slides.

Every roadmap is subject to change. You are not saying anything new. If tomorrow the foundry where Kaveri is being fabricated explodes and all the production is lost, that legal note saves AMD against investors and others, because the roadmap would be delayed. At contrary, if tomorrow Intel surprises everyone with an accelerated roadmap and releases Broadwell in two months, AMD would also change its roadmap to reply that challenge, probably by accelerating Carrizo arriving. And so on.

The other part of the note is not even related to what you say. You are completely misreading this one. The "first year of production shipments" means, for instance, that Kaveri will start to ship in late 2013, which means that you will have to wait up to early 2014 to purchase one.

Of course products in boxes at the right are replacing products in boxes to the left:

Radeon 200 series replace HD 8000 series
Kaveri replaces Richland
Beema replaces Kabini
Mullins replaces Temash
Warsaw replaces Opteron 6300/4300
Berlin replaces Opteron 3300
Seattle replaces Opteron-X
Bald Eagle and Hierofalcon replaces embedded R series
Steppe Eagle replaces embedded G series

And so on

But it is not a marketing demo to average Joe but a technical talk to developers. During the talk, it was shown with a concrete example (aka code) that the overhead can be reduced by one order of magnitude. The hardware with MANTLE was able to compute 10x more draw calls per second... whereas gamerk said that such numbers would be impossible.

Unless the exact code was released to the public and independently verified it means diddly squat. Any number of tricks or omissions of optimizations could be used behind the scenes to make one method appear faster than another. They're trying to create buzz and sell the idea of Mantle to the developer community. Of course they will have some nice canned demos.

I've written hundreds of thousands of lines of code. You can't tell much from a few code snippets they put up on a slide at a conference. It's just 1 piece of a giant puzzle.

Basically you are suggesting that the guy who gave the talk was lying. Care to repeat your accusation using your real name?

Also if you are a developer and you are familiar with this, you would know that every game developer knows that the overhead is 10x (or even higher) and many of them report this number. I recall Carmack complaining about the 10x overhead in DX several years ago, in public, before anyone mentioned a single word about MANTLE. Do you?

You did not read what I got in terms of word from AMD on the last page did you?

It seems many people missed the update from my contact.

Your entire post is refuted by word from someone working at AMD.

Good day.

 

[de5_Roy]

^^ oi. post something like that before i spend my precious time researching and looking for information, hit reply, type, edit, edit out sarcasm, proofread, revise and hit answer. 😆
i missed the carrie diaries for this !@#567!!!

 

[gamerk316]

MS now has to buy APUs from AMD when AMD has created a direct competitor to DirectX.

More like MSFT will allow some more low-level access into the API. Not like they can do too much more graphically until Ray Tracing comes along.

Graphically, we're tapped. Hence all the focus on new AA modes these days. We can't do much more via rasterization.

I don't know what else you want. At APU13 AMD announced Mantle will make it possible to shift:

1. Rendering across multiple GPUs and CPU cores
2. Physics to GPU
3. AI to GPU
4. Global Illumination across GPUs

You can ALREADY do this, and its already been done.

I find it fascinating how they showed a concrete example where MANTLE improved performance by 3x (200%) and still gamerk pretends that the DX overhead is at best of 10%

They used an example with an excessive amount of draw calls, which no developer in their right mind would ever do due to the performance hit you would take. So they tested an artificially created "worst case" example.

Secondly, I said the overhead of Draw Calls is about 10% in DX11. You can effect other speedups via lower-level processing.

Yes I know. I have the roadmap, do you remember? And it shows AMD will continue selling old FXs.

Do you see it now? All your speculations about SR FX, Phenom return, AM4, FX replacement... vanished in the air.

The word "performance" also appears in the Kaveri row. And don't be confused but Kaveri is much more powerful than a FX-8350

FX-8350: 256 GFLOP
Kaveri: 118+738 = 856 GFLOP

It will be funny to do the first HSA benchmarks and show how Kaveri destroy a FX-8350 in a number of workloads.

Only in HSA/OpenCL/Compute workloads that both use the GPU, and is reasonably parallel, and scales enough to use the GPU resources. CPU wise, Kaveri is about 60% slower then PD absolute, so its about HALF as fast as an Intel chip. In CPU workloads, it will perform like crap. Likewise, when paired with a discrete GPU, it will perform like crap.

When you talk about Kaveri, you MUST do it remembering it has an integrated GPU. Give the FX-8350 a GTX 630, and guess what? Kaveri gets slaughtered.

 

[de5_Roy]

from amd's official page, i got some numbers amd expects from new products:
http://www.amd.com/us/press-releases/Pages/amd-2014-mobile-apu-2013nov13.aspx

Spoiler

mullins platform is expected to deliver up to 139% better productivity performance per watt when compared to the previous generation “Temash” platform. PCMark 8 - Home score divided by TDP (W) is used to simulate productivity performance per watt; the Mullins platform (4.5W) scored 1809 while the Temash platform (8W) scored 1343. AMD "Larne" reference platform system used for both APUs. Temash-based AMD A6-1450 quad-core APU with AMD Radeon™ HD 8250 Graphics, 2x2GB of DDR3-1333MHz RAM (running at 1066MHz), Windows 8.1, 13.200.11.0 - 03-Sep-2013 driver. Pre-production engineering sample of “Mullins” quad-core APU with next generation AMD Radeon graphics (model number TBD), 2x2GB DDR3-1333MHz RAM, Windows 8.1, and unreleased reference driver.

The new 2014 AMD A-Series mainstream APU platform, codenamed “Beema,” is expected to deliver up to 104 percent better productivity performance per watt when compared to the previous generation “Kabini” platform. Testing conducted by AMD Performance Labs on optimized AMD reference systems. PC manufacturers may vary configuration yielding different results. PCMark 8 - Home score divided by TDP (W) is used to simulate productivity performance per watt; the Beema platform (15W) scored 2312 while the Kabini platform (25W) scored 1861. AMD "Larne" reference platform system used for both APUs. Kabini-based AMD A6-5200 quad-core APU with AMD Radeon™ HD 8400 Graphics, 2x2GB of DDR3-1600MHz RAM, Windows 8.1, 13.200.11.0 - 03-Sep-2013 driver. Pre-production engineering sample of “Beema” quad-core APU with next generation AMD Radeon graphics (model number TBD), 2x2GB DDR3-1600MHz RAM, Windows 8.1, and unreleased reference driver.

ifixit tears down PS4
http://www.tomshardware.com/news/playstation-4-repair-guide-ps4-teardown-ifixit,25093.html
ps4 has these two components i did not notice before:
• Samsung K4B2G1646E-BCK0 2 Gb DDR3 SDRAM
• Macronix MX25L25635FMI 256 Mb Serial Flash Memory

 

[noob2222]

I was not alone on saying that AMD will be releasing ARM tablets. It has been reproduced in several sites like this

http://techreport.com/news/25461/report-amd-to-introduce-arm-based-tablet-chip-this-year

You finding both zambezi and vishera cpus on sale, doesn't mean what you pretend. AMD is no more fabricating Zambezi, because it was replaced by Vishera. What happens is that some stores have inventory of the old processors. And continue selling them, up to finishing it.

The same happens with any other AMD product. Kaveri replacing Richland doesn't imply you cannot find Richland APus in your favourite store during some time, before the inventory finish.

Looks like your starting to get caught up in your own lies now. you originally said replace and was completely adamant at the fact that beema and mullins would be canceled since they "had no design wins" and "are not competative"

Even de5 caught the pre-edited version with "replace" in it.

I told you when you first brought up this article that the word release != replace. Considering beema and mullins are socket compatible, I wonder what the chances are of seeing opteron x label stamped on a few as well.

 

[gamerk316]

And in other news: NVIDIA announces CUDA 6, with 100% more Unified Address Space:

http://www.fudzilla.com/home/item/33144-nvidia-announces-new-cuda-6

Well, that advantage lasted long. I'm predicting Unified Memory will now be the key feature for DX12.

 

[juanrga]

And still my reading of that infamous slide was accurate, with APU 13 confirming several of my claims about APU/CPU performance

no it was not. you admitting of using semantics instead of facts simply confirmed my accusations, whether you realized it or not. i might not require your calculations and results after all.

what claims?

No semantics, but merely reading the slide, including the FPs for each processor. I recall you pretending that the slide didn't gave that information, but it was untrue.

I mean the claims made about how the weigth of CPU performance is minimized with MANTLE. The claims about how a medium-performance CPU can feed a high-end dGPU.

When you are offering me something that I said that I have, you are proving that you are not reading.

let me make this perfectly clear: a lot of forum posters here at toms post, asking for help with getting into their laptop/phones/pcs/whatever device since they lost their password. they use forum ids, email address, cell numbers etc. that look quite legit. i will tell you the basically same thing i tell them: "there is no way to legitimately verify your claim of having the roadmap." woulda been far easier to verify if you posted it right away (claiming nda won't hold water). i know you didn't post anything, i read the thread 😆. since i know now that you misinterpret, misread, mislead, lie(your own admission), i just don't believe you.

And you show again that you don't read, because someone suggested that I was looking at some APU roadmap and I gave a hint about it.

If you don't read how do you pretend to analize what others write? It is amazing.

Spoiler

Nope. we are all discussing what AMD would release in 2014 and all people here who was waiting SR FX, Phenom return, 10-core CPUs, AM4 sockets, FX replacement... was disproved by the 2014 roadmap. This is reality 1.0
I recall when I said "there is no 10-core FX, no phenom return and no SR FX for 2014" and someone said me "wait for the roadmap". Well the roadmap is here and he is now silent.

ah reality... here's a reality check for you: those people were speculating, unlike you. you are arguing speculations and conjecture about future products that may or may not exist.

Most were speculating, but some few were convinced that a "FX replacement" was coming soon. I knew that nothing of that was real.

In servers, Warsaw only comes as 12 and 16 cores. There is no 2/4/6/8 core versions. This means that dual-module and single-module dies (passing binding) are not reused for lower core Warsaw versions, but trashed due to lack of demand.

now there're servers too? i wonder what else did the slide#13 didn't address...

Evidently the slide #13 is not about servers, but the server roadmap gave a hint to what expect for desktops. You and others didn't see the signals, but you cannot be hungry by that.

you seem to have a really hard time understanding the binning process. as long as amd makes a 'big' die with 4 modules, 2/4/6/8 core cpus will exist. 12 and 16 cores are based on 'bigger' mcm [strike]die[/strike] package, 12 core variant being lower binned. technically, amd can make a 10 core fx/opteron out of those by binning 4+1 modules or 3+2 modules. they don't do it for economical reasons and for a 10 core cpu like that isn't worth using the big mcm package.

did amd say warsaw will be a monolithic 16 core cpu? if they did, i want to see an official proof. afaik, their 'bigger' cpu is still 2x 4M mcm package.

there.is.no.dual.or.single.module.die.for.cpus. those are all binned from the 'big' die. only apus have dual module but those dies also contain igpu, u.n.b. etc.

lack of demand? of nonexistent silicon? yeah....

And another misunderstanding-fest.

AMD is making a big die with 4 modules based in a PD refresh. They will be using those for the new Warsaw 16 core (8x2 MCM). Dies with one defective module will be used for the 12 core version of Warsaw. Dies with two defective modules will be trashed. There is no 2/4/6/8 core version of Warsaw.

The previous 2/4/6/8/12/16 core Opteron 6300/4300 series get replaced by Warsaw (12 and 16 core). There is no demand for AMD fabricating 2/4/6/8 core versions of Warsaw. Warsaw is only 12 and 16 core.

Do you get it now or still don't?

At contrary, the roadmap reflects accurately what I am saying. It was also confirmed by Lisa Su (AMD Vicepresident) during opening keynotes this week. APU are the center of universe for AMD. AMD plans to dominate the market from phones to servers using APUs. FX and Opteron are EoL products.

while it is true that amd is focussing on apus, your claim stays unverified until you show where lisa su says amd is abandoning fx, in "during opening keynotes this week".

Of course, when she said that APUs are how they want to lead the market from phones to servers, She did really mean "we are going to release a 10-core SR FX @ 5GHz tomorrow Stay tunned" :ROFL:

End sarcasm.

i also want to see an official confirmation of fx and opterons being e.o.l. although i know for a fact that you are lying about opterons being e.o.l. kyoto has opteron branding, remember? warsaw will have opteron brand as well.

It is evident that FX are eol. They don't even receive a refresh a la Warsaw and all the important stuff is being made in FM2+ platform for APUs.

A simple look to the server roadmap reveals that are gone:

Opteron-X replaced by Seattle.
Opteron 6300/4300 replaced by Warsaw.
Opteron 3000 replaced by Berlin.

And AMD also stated that Warsaw is a only a legacy product for customers who will be slow on migrating to Seattle.

😆 That is the 2013-2014 roadmap. Ok? The new products: Kaveri, Beema, Mullins, Warsaw, Seattle... are 2014 products. They will replace the older products the next year, no today when the roadmap has been presented.

oh, you mean the 2014 roadmap? y'know, the one that says(like very other roadmap) that roadmaps only indicate first year of product shipment and Nothing about replacement?

Nope the one (and only one) that show what products are being replaced and the names and basic properties of the new products.

I gave you a list of replaced products and the new products. I think you miss that also, but you don't read, remember?

It was not alone on saying that AMD will be replacing Temash by ARM SoC. It has been reproduced in several sites like this
http://techreport.com/news/25461/report-amd-to-introduce-arm-based-tablet-chip-this-year

somehow i knew you'd resort to that article. please, read the article carefully and show me where amd says they will Replace Temash with ARM soc. the techreport article(source:sweclockers) says nothing on replacing temash. even the sweclockers source says nothing about replacing temash. i checked that article using google translate. there's always a chance that i missed something, so you will have to provide proof of your own claim.

The article clearly says that Temash has been a failure in the tablet market and says that the ARM design could be competitive and help AMD to gain markest share in tablets. Replacement is when you take an old product or one that is not competitive and substitute by one that is competitive. The 2014 roadmap shows that AMD is retiring Temash from its products (because was not competitive), it doesn't mention the ARM SoC for tablet (the HieroFalcon Soc mentioned in the techreport article however appears in AMD 2014 roadmap).

Spoiler

You finding both zambezi and vishera cpus on sale, doesn't mean what you pretend. AMD is no more fabricating Zambezi, because it was replaced by Vishera. What happens is that some stores have inventory of the old processors. And continue selling them, up to finishing it.
The same happens with any other AMD product. Kaveri replacing Richland doesn't imply you cannot find Richland APus in your favourite store during some time, before the inventory finish.

now you're switching to fabricating, eh? your entire tirade about replacement was based on roadmaps that only indicate intended shipment timeline, not fabrication. fabrication begins earlier. and it does not apply to seattle and kyoto which address the same market yet use different uarches.
i was writing from customer's p.o.v.(that's why i mention sales). yes, amd might not be fabricating zambezi and soon richland anymore but in the market you can choose from both.

Of course I am mentioning fabrication to explain you why a product being replaced by other in 2014 doesn't mean you cannot find the old product in some store with unsold old products in their inventory.

:ROFL:

 

[juanrga]

Spoiler

you claim some message is given by the slide #13, but never explain how you got it. the 'explanation' you gave do not relate to the slide at all. the slide only shows a gpu-bound game benchmark, nothing about your claims.

I got the message from reading the slide, also between lines.

btw, you are wrong on both of your 'amd (i), (ii)options' in your claim: as long as vishera exists, amd will continue selling lower binned, dual module vishera as fx4000 and they are not abandoning fx entirely for apus. they're simply carrying on the vishera lineup for now, according to the latest desktop roadmap.
http://farm3.staticflickr.com/2836/10861225406_8dbbd0dc46_o.png

Just in the last page I said that I already have a copy of the roadmap. This proves again that you don't read the thread.

Yes, my hypothesis (i) and (ii) were wrong, but infinitely close to reality that all those who speculated about SR FX, Phenom return, 10-core CPUs, AM4 sockets, FX replacement...

My hypothesis (ii) was that AMD abandons the FX-4000 series and release a refresh of the 6000/8000/9000 FX series a la Warsaw for 2014. But the sales of the FX chips must be so low that AMD is not even releasing such refresh. They must be selling existences.

And yes, they are abandoning FX by APUs. That is why the APU line receives Steamroller update and many more improvements, but the FX-line does not.

but here's the real kicker at the bottom of the slide:

amd roadmaps are subject to change without notice or obligations to notify of change. placement of boxes intended to represent first year of production shipments.

this last line completely contradicts your earlier claim about "side by side boxes on a roadmap mean product replacement". rofl. if amd meant replacement, they woulda said so. the boxes only mean when new products are supposed to start shipping, and even that is subject to change without notice i.e. nothing is set in stone. now i know why some sites crop part of promo slides.

Every roadmap is subject to change. You are not saying anything new. If tomorrow the foundry where Kaveri is being fabricated explodes and all the production is lost, that legal note saves AMD against investors and others, because the roadmap would be delayed. At contrary, if tomorrow Intel surprises everyone with an accelerated roadmap and releases Broadwell in two months, AMD would also change its roadmap to reply that challenge, probably by accelerating Carrizo arriving. And so on.

The other part of the note is not even related to what you say. You are completely misreading this one. The "first year of production shipments" means, for instance, that Kaveri will start to ship in late 2013, which means that you will have to wait up to early 2014 to purchase one.

Of course products in boxes at the right are replacing products in boxes to the left:

Radeon 200 series replace HD 8000 series
Kaveri replaces Richland
Beema replaces Kabini
Mullins replaces Temash
Warsaw replaces Opteron 6300/4300
Berlin replaces Opteron 3300
Seattle replaces Opteron-X
Bald Eagle and Hierofalcon replaces embedded R series
Steppe Eagle replaces embedded G series

And so on

But it is not a marketing demo to average Joe but a technical talk to developers. During the talk, it was shown with a concrete example (aka code) that the overhead can be reduced by one order of magnitude. The hardware with MANTLE was able to compute 10x more draw calls per second... whereas gamerk said that such numbers would be impossible.

Unless the exact code was released to the public and independently verified it means diddly squat. Any number of tricks or omissions of optimizations could be used behind the scenes to make one method appear faster than another. They're trying to create buzz and sell the idea of Mantle to the developer community. Of course they will have some nice canned demos.

I've written hundreds of thousands of lines of code. You can't tell much from a few code snippets they put up on a slide at a conference. It's just 1 piece of a giant puzzle.

Basically you are suggesting that the guy who gave the talk was lying. Care to repeat your accusation using your real name?

Also if you are a developer and you are familiar with this, you would know that every game developer knows that the overhead is 10x (or even higher) and many of them report this number. I recall Carmack complaining about the 10x overhead in DX several years ago, in public, before anyone mentioned a single word about MANTLE. Do you?

You did not read what I got in terms of word from AMD on the last page did you?

It seems many people missed the update from my contact.

Your entire post is refuted by word from someone working at AMD.

Good day.

It is evident that you didn't even read the post.

I have one doubt, now that AMD has released the official roadmap for 2014 and there is no SR FX, no FX replacement, no Phenom reborn, no AM4... (as I said to you), will you start to claim in forums that a SR FX / AM4 platform is coming for 2015?

 

[griptwister]

Ok, here's what I know...

AMD's 2014 product roadmap will release shortly if not today...

Piledriver FX is going to be the HEDT line through 2014 until further notice. AMD concedes Kaveri is not going to replace FX in HEDT, and they know the segment is crucial for gaming, and it's the only segment growing at this time in DT PC.

AMD expects to see the PD arch gain performance boosts in several facets and remain mostly competitive at the price points they have established through software optimization and other means.

The next HEDT platform already has an internal codename, though he could/would not tell me what it was. There were no promises that we would see the HEDT successor in SR arch, it may not be until excavator that they move forward. The issue is simply that they will not do HEDT on bulk given the way things went with Kaveri. If they had been able to completely meet/exceed the design goals on bulk with Kaveri they would have moved forward with HEDT. However, he expressed the sentiment that they regretted being forced to downgrade their process to bulk and that they know they lost performance to release the APUs in a timely manner.

Next HEDT line may come on 28nm or 20nm, he wasn't clear there, but he was certain it would not be bulk, and felt like they would move back toward FD-SOI if GF could get their "issues" sorted out for 28/20nm SHP. He could not tell me if the next iteration of their HEDT line would be a FM2+/FM3 product, or if they would in fact move forward with AM4 when DDR4 arrives, though he expects that the timing of DDR4 and the new HEDT platform would be pretty close. Supposedly, there is a new IMC in the works for next gen HEDT hardware (to coincide with DDR4).

That's all I could get...

I REALLY hope for AM3+ support...