AMD Zen Features Double the Per-core Number Crunching Machinery to Predecessor

[GorbazTheDragon]

Title is misleading... It only doubles floating point, not integer performance.

 

[nem]

intel fanboys after read this new.

 

[theeldest]

Just hope AMD isn't going to try to charge a premium for it. Course if they'll finally have CPUs that will go toe to toe with Intel then I'm sure they will.

If it goes toe-to-toe with Intel they had BETTER charge a premium for it. Otherwise how will they regain financial stability?

 

[Solidstate89]

Just hope AMD isn't going to try to charge a premium for it. Course if they'll finally have CPUs that will go toe to toe with Intel then I'm sure they will.

If it actually does somehow manage to go toe-to-toe with Intel (or at least close to it) why shouldn't they charge more for it then their current CPU offerings?

AMD is a company out to make a profit, not be your best friend.

 

[cdawall]

AMD has to recreate their name. Even if this competes with Intel's offering (outside of the server market) AMD still comes with the negative connotation of being, hot power hungry underperforming processors. This will force AMD to sell their CPU's at a lower price to regain marketshare.

 

[ZeDestructor]

Considering though that Intel brought nothing to the table since Sandy Bridge, they might have a chance. (lower lithography gives better power, and very slightly better performance which will be null, when Zen will come, cpu graphics is irrelevant for performance machines, and the rest of the performance increase over sandy is mostly due to higher stock clocks)

Not true. Intel has made steady 5-10% IPC improvements from SNB to SKL (I expand on it in fair detail here and here, complete with sources!)

AMD has to recreate their name. Even if this competes with Intel's offering (outside of the server market) AMD still comes with the negative connotation of being, hot power hungry underperforming processors. This will force AMD to sell their CPU's at a lower price to regain marketshare.

They have to compete at the server level if they want to actually earn real money, and with the success that Intel has had since SNB-EP/IVB-EX and newer chips, it's very clear that you need lots of cores, lots of bandwidth, lots of IPC and a decent amount of vector processing on the CPU, even on very GPU-centric boxes, and that's where AMD is reverting to after the utter failure of their HSA "bet" (Bulldozer cores with strong Integer perf, offload FP to GPU(s)).

 

[FordGT90Concept]

Title is misleading... It only doubles floating point, not integer performance.

FPU is where AMD is weakest because, presently, it is shared across two cores. AMD assumed most FPU work was going to move to the GPU which is false. They took the Achilles' Heel of the CPU and broke it.

 

[Casecutter]

Did anyone notice the picture actually says (with little speculation) from Mark Waldhauer. Well it seems “Dresdenboy” aka Matthias Waldhauer took inspiration from this patch and hypotheses more here as to what the patch might revel.

I think it's all speculation and a just really to early to start beating any Zen drum.

While it's fun to think what Mark Keller might have come up with, I think the premise of...

"what if a Steamroller module of two cores was just one big core, and supported SMT instead."

Does anyone believe it was that straight forward of a revamp.

I say tapper the enthusiasm as this is "speculation", and nothing that has fact in what AMD/Keller went about laying out.

 

[RejZoR]

While it's nice to keep us in the loop, I kinda wish AMD would just drop Zen out of the blue and shock Intel a bit. Giving them headsup like this only means Intel will stop sleeping on laurels and start doing shit right now. Which is bad for AMD considering the resources (in)balance...

 

[FordGT90Concept]

If AMD comes out swinging, they'll have a window to reclaim marketshare before Intel can push out processors with more cores. I'm not entirely sure LGA 15## can even handle a mainstream AMD monster. Intel could get caught with its pants down with only LGA 2011 being able to respond. My hope is that AMD will force Intel to move LGA 2011 to mainstream/enthusiast and LGA 15## to budget.

 

[librin.so.1]

Giving them headsup like this only means Intel will stop sleeping on laurels and start doing shit right now. Which is bad for AMD considering the resources (in)balance...

Yeah, because corporate espionage doesn't exist and intel finds this stuff out the same time we normal people get to know it.

 

[ZeDestructor]

Did anyone notice the picture actually says (with little speculation) from Mark Waldhauer. Well it seems “Dresdenboy” aka Matthias Waldhauer took inspiration from this patch and hypothesis more here as to what the patch might revel.

I think it's all speculation and a just really to early to start beating any Zen drum.

While it's fun to think what Mark Keller might have come up with, I think the premise of...
Does anyone believe it was that straight forward of a revamp.

I say tapper the enthusiasm as this is "speculation", and nothing that has fact in what AMD/Keller went about laying out.

If the source data is real, then the diagram is pretty logical and straightforward as far as overall speculation goes. And from where I'm sitting, it looks very much like an clone of the SNB-HSW lineup as far as the overall design goes (EU counts, port counts, number of each EU type)

If AMD comes out swinging, they'll have a window to reclaim marketshare before Intel can push out processors with more cores. I'm not entirely sure LGA 15## can even handle a mainstream AMD monster. Intel could get caught with its pants down with only LGA 2011 being able to respond. My hope is that AMD will force Intel to move LGA 2011 to mainstream/enthusiast and LGA 15## to budget.

LGA1155 can easily handle 8, maybe even 12 cores. The reason it can lies in the PCIe and memory controllers being on the same die as the CPU core, and as a result, the socket just acts as a unified PCIe/RAM/DMI socket. Hell, if they wanted to, they could probably fit their 18-core monster Xeon on LGA1155, but at that point, dual-channel RAM would be an actual limiting factor.

EDIT: I talked about it in this old thread a bit, and my expetation hasn't changed: if Zen outperforms SKL/KBL, Intel will just change core counts at the different price points and match AMD again, the obvious one being enabling HT on i5 and moving i7 to 6core HT.

 

[FordGT90Concept]

It's the power/VRM that's the problem. As far as we know LGA 1151 tops out at 95 watt which i7-6700K uses. If they add two more cores, something is going to suffer be it cache, removal of the GPU, or a significant drop in clockspeed.

 

[ZeDestructor]

It's the power/VRM that's the problem. As far as we know LGA 1151 tops out at 95 watt which i7-6700K uses. If they add two more cores, something is going to suffer be it cache, removal of the GPU, or a significant drop in clockspeed.

Minor drop in clockspeed (heat/power scales linearly with clock, quadratically with voltage, and voltage scales with clockspeed.. effectively something between linear and cubic scaling overall), and require the beefier end of mobos (the average Z170 should be good, with the average 25+% OC considered bare minimum). Possibly more restrained Turbo as well, for good measure.

 

[librin.so.1]

removal of the GPU

When was the last time Intel did this to a non-Xeon CPU?

 

[Scrizz]

AMD has to recreate their name. Even if this competes with Intel's offering (outside of the server market) AMD still comes with the negative connotation of being, hot power hungry underperforming processors. This will force AMD to sell their CPU's at a lower price to regain marketshare.

That sounds like Intel in the P4-Pentium D era.

 

[FordGT90Concept]

When was the last time Intel did this to a non-Xeon CPU?

It would have made sense on 6700K. Even so, we don't see 6 core Xeons on LGA 15## either.

 

[Corey]

To little to late?

Nah they ain't on their death bed yet, they are close. I think a lot of people have been waiting for this chip for a long time. I have personally been waiting to upgrade my FX 8350 and my Sandy Bridge i7. It might be better than the FX but I don't think its going to do much over the Sandy (should be a lil faster but prob not enough to upgrade). What is going to matter the most is if they bring Zen out at the cheaper price point of the i5's hoping it sits between the i5 and i7 in performance.

 

[RejZoR]

It's the power/VRM that's the problem. As far as we know LGA 1151 tops out at 95 watt which i7-6700K uses. If they add two more cores, something is going to suffer be it cache, removal of the GPU, or a significant drop in clockspeed.

No it's not. They already have 5820K and boards can deliver way over 95W for LGA1151. Thy just need to sack the stupid GPU part in 6700K and replace it with extra cores. But they're not going to do that because 5820K already exists and they don't want to spit in their own enthusiast bowl...

 

[FordGT90Concept]

5820K is on LGA 2011. See this post for context.

 

[Steevo]

I look at it all this way.


We are close to the end of the line with IPC improvements for Intel or AMD, the rest will come through process, cache, and instruction set/hardware support. We are close to the end of the line with Silicon in the high performance categories. We can slap more cores in, more sockets, more memory. But the next big thing is either going to be quantum computing, or photon based.


Either way until I see hard numbers from a source I trust, AMD is a zombie.

 

[AsRock]

While it's nice to keep us in the loop, I kinda wish AMD would just drop Zen out of the blue and shock Intel a bit. Giving them headsup like this only means Intel will stop sleeping on laurels and start doing shit right now. Which is bad for AMD considering the resources (in)balance...

I am sure intel have so many ide;as by now that even i f AMD did release a totally awesome CPU intel would have some thing better as we all know intels been holding back it's had a hell long time to plan shit.

 

[RejZoR]

5820K is on LGA 2011. See this post for context.

You know, considering I own one, I'd most likely know that, don't you think? What I was saying is that they have the tech in retail form. They can easily produce something similar for the LGA1151 in no time.

Only thing that AMD will mess up is the way how mainstream and enthusiast is now separated. Because if Intel will have to bump mainstream up quickly, it means they'll have to bump up enthusiast as well, otherwise they'll make them equal which means they'll sell less of the more expensive enthusiast platforms and CPU's. But that's good as far as consumers go, assuming the prices stay low...

 

[FordGT90Concept]

It takes at least a year to produce a new processor including design from an existing microarchitecture to prototyping to mass production. If AMD put out a processor that beats 6700K (in today's terms) for $300, Intel couldn't respond with a competitive product for a year. Even assuming they already did prototyping and have a design ready for mass production the shelves, AMD could still get a several month long window to grab market share.

Intel's immediate response would probably be cutting prices across the board which likely moves LGA 2011 into mainstream prices. As I said in the post, a shocker from AMD could turn LGA 1151 into a budget socket overnight and LGA 2011 into mainstream.

 

[geon2k2]

Not true. Intel has made steady 5-10% IPC improvements from SNB to SKL (I expand on it in fair detail here and here, complete with sources!)

You're veteran here and for sure you know what you are saying.
Obviously there were improvement and new instruction sets but they were underwhelming in current applications to say the least.

Of course Intel will still have the performance crown for a long time from now on as they have strong Enthusiast parts, but that is not the mainstream market. The mainstream market is formed by the i3, i5 and i7 non E series.

If anyone, can produce a competitive product here then they are in business, and intel in this area was pretty much sleeping over the years. You said it yourself in those long posts, ivy was die shrink, haswell brought very good power consumption, broadwell its a different beast but I'd exclude as it has expensive edram, and its mostly a very expensive GPU with CPU cores, good for laptops ... but not very viable in my opinion from cost perspective for desktop. Lake mostly die shrink ... again a bit underwhelming.

I found an article which compares different generation performance at same clock also.
It doesn't include the latest architecture though, but it is focused on gaming, which is basically the reason for which most of the people buy these processors, otherwise for browsing or office even the atom is good. If you think is too extreme, fine go with an i3.

See here: http://wccftech.com/intel-sandy-bridge-ivy-bridge-haswell-graphics-compared-10-difference-average/

I put them in a table also, and compared similar products:

Lowest details i7 2600K @4.5 i7 4770K @4.5 % increase
Crysis 3 91 93 2%
Black Ops 2 355.4 382.8 8%
Bioshock 243.2 265.9 9%
Battlefield 3 199.8 200 0%
Unigine Heaven 4243 4280 1%
Firestrike 7292 7466 2%

Average: 4%

This increase you will probably get by just shrinking the original sandy without any architecture change. Maybe things will change once the new instructions sets will start to be used ... but if they are not supported by most of the computers in use, they will mostly be scattered optimizations in one app or another.

So yes, if anybody can get sandy level performance, good clocks and reasonable power consumption they are back in business.