If IPC can rival Intel, I don't care what the clocked rates are. It's core efficiency I care about, numbers don't always mean performance.

Polaris is also clocking considerably less than Pascal, so it's indeed quite possible architectural decisions aren't the only explanation and GlobalFoundries simply can't get it up at the moment. Although that being said, nothing explains why TSMC would have it working better since they are no more experienced with the smaller tech than GloFo

but of course they are. TSMC 16FF is used in many more designs by much more clients than Glofo 14FF

This clocks are on 95w and under, not max OC clocks Correction, it says AMD "HT" hyper threading is a Intel propetary SMT tecnology, must say AMD SMT implementation. about the new, its well that new batches has higher clocks, i see people being negative with the 3.60GHz or less clock but remenber... this CPU integrates much controllers, is hard to get higher base clocks @95w they don´t want get higher TDP this time, if they get 85% Skylake IPC and $350 price or less its ok for me. Other thing, this is base and turbo clocks @95w is much posible that Motherboards will support 125--140w then the OC could be +4.50GHz if the SoC and Chipset it´s well designed...

Cryengine 3 with Crysis 3 is capable of using 8 core processors As i am looking forward to Star Citizen, i like you guys. To tell me if it would be okey to go for 8 core cpu. When i want to upgrade, i am only going to play SC, ( SC is using a moddet Cryengine 3.6 ) Thanks for any reply's 🙂

So it's still close to intels 8 core cpus in ghz with tad lower usage atm. Max turbo of 3.7 on 6900k is also single core turbo. 3.6 vs 3.7 not that bad. Now if they get base frequency up a bit while still delivering less wattage.

So it's still close to intels 8 core cpus in ghz with tad lower usage atm. Max turbo of 3.7 on 6900k is also single core turbo. 3.6 vs 3.7 not that bad. Now if they get base frequency up a bit while still delivering less wattage.

The wattage difference is more than 40% in favor of Zen. I bet the clock differences aren't even near that. I'm sure that Zen clocks lower in general. But if it has similar/a bit faster IPC, it's clearly the better CPU at that point. Everything will depend on the price. If AMD can deliver this eight core in the same ballpark as a 7700k/6700k, it's a winner no matter how you see it.

I'm in 100% for a new 8core 16thread AM4 set up but how long do you give it for maturity. My current system is still pumping out great fps in all games i play but i still feel the need to go back to an AMD build but i think i'll give Zen a whole year to iron out any bump in the road. 4Ghz+ on a Zen 8c 16t will surely be more than enough for future gaming.

Maybe I'm being pessimistic, but I feel like people are so desperate for AMD to bring Zen out and that its something so big and earthshattering, that they are letting their excitement and buying into the hype cloud their judgement a little bit. Some of the stuff I've been reading about from people hyped up is truly fascinating.. acting as if this will be the best CPU to get based on numbers alone. As we should all know by now, this is fallacy, and until we get some real world benches on things that actually matter, all of this hype is just hot air. I sure hope they can bring a competitive chip to market.. by now AMD better have it absolutely nailed once its released. They imo have a short leash and people will not be patient with them if this release doesn't knock everyone's socks off.

unlike Intel who just have to let out a little boost to there cpu's over time, this time at the very least AMD fans will get an very noticeable upgrade this time around.

Its not Intel's fault that AMD hasn't released a cpu (in tech years) in an enternity. I would certainly hope that AMD users get a huge upgrade lol Hardly a fair comment to make imo

I'm not that fussed about the chip since an AM3+ 990FX FX8xxx/9xxx is matching Intel systems now in modern games FPS wise there's hardly any difference already as HH has shown in recent tests. The main thing most AMD fans will want is to finally get all the latest protocols like DDR4 and PCI-e 3.0, M.2 and a new chip-set finally since AM3 is ancient. I'm looking forward to a nice shiny AM4 mobo UD5 or similar.

Maybe I'm being pessimistic, but I feel like people are so desperate for AMD to bring Zen out and that its something so big and earthshattering, that they are letting their excitement and buying into the hype cloud their judgement a little bit. Some of the stuff I've been reading about from people hyped up is truly fascinating.. acting as if this will be the best CPU to get based on numbers alone. As we should all know by now, this is fallacy, and until we get some real world benches on things that actually matter, all of this hype is just hot air. I sure hope they can bring a competitive chip to market.. by now AMD better have it absolutely nailed once its released. They imo have a short leash and people will not be patient with them if this release doesn't knock everyone's socks off.

It's all a performance/cost game. Zen at 95W seems to be very close to a similar entry from Intel in the 140W range. If it sells at something like 2/3 of its price, then it's game over, and it's a very exciting thing, yes. We have had stagnation for almost a decade in the CPU space. Keep in mind also that Intel has basically been refining the Sandy Bridge design for almost six years now. The Core architecture is very very mature, and I'm not sure that large perfomance gains can be had from it, unless Intel presents something entirely new (like they did with Core back then). Seeing how both Zen and the Core CPUs are so similar to each other (micro op caches etc), it seems unlikely to me. Interesting times ahead.

I'd be delighted to swap the CPU in my sig out for a Zen 8 core that's in the same ballpark as a i7 6900k and the price is in the $300-350 range. I wouldn't be surprised to see the flagship higher then that though.

Do we have confirmation how many PCIe 3.0 lanes the consumer CPU and Chipset will provide? How many lanes on the DMI between the CPU and Chipset? Hoping to run a graphics card at x16 and at least two m.2 drives in x4 on my next build. Currently impossible on Skylake and presumably Kabylake as the chipset DMI is only x4 so the m.2 drives will be fighting each other for bandwidth.

Its always fun reading about the demise of AMD, yet they still exist. They are a small company compared to Intel, yet they continue to innovative in both their CPU/GPU's with limited funds. Always been an AMD fan, building my setups from the early days of Cyrix, before switching over to AMD. Never once have I had an issue running a game. Always paired my CPU with the right GPU and other components. Also knew that I didn't have to have all the eye candy set to max to enjoy my gaming experience. Maybe when you have a still screen you can see the difference from Min to Max settings, but in todays fast paced games, i'll bet most people can't. My latest build was a FX9590/Dual Fury X's. Played all games with ease at max settings, yet cost me hundreds less than a comparably Intel/NVIDIA setup. For some reason I decided that I needed an I7-5960k to feed my Fury's while draining my bank account to get a new CPU/MB/RAM. Turned out I didn't unless I was benching. I know when Zen arrives it will be see ya Intel hello AMD again for me.

Do we have confirmation how many PCIe 3.0 lanes the consumer CPU and Chipset will provide? How many lanes on the DMI between the CPU and Chipset? Hoping to run a graphics card at x16 and at least two m.2 drives in x4 on my next build. Currently impossible on Skylake and presumably Kabylake as the chipset DMI is only x4 so the m.2 drives will be fighting each other for bandwidth.

Won't it integrated the I/O? I thought this was one of the innovations.

Not again with the "clocks matter" BS. what a CPU is clocked at has never mattered people, give up on that whole idea. The only thing that matters is how it performs.

Correction, it says AMD "HT" hyper threading is a Intel propetary SMT tecnology, must say AMD SMT implementation.

Where did you found that info?! As far as I know, HT is not invented by intel

@ghostxl/cryohellinc Not sure what "100y" os/games you guys running ;-) but on my 3770k (4gh on all cores), load is around 30-60% on all 8 "cores" for games from the last couple of years till 2015 (mainly FPS). Even that i dont see any need for heavy oc (i do remember testing 4.8 at 1.3V), i still wanna upgrade to zen to get M2 and more usb/sata ports (mAtx), but since all the "news" after that "we're as fast as intel.." bench are not thaaat promising (even if its an e-sample, since they cant do magic over night and double perf), i might go for a 8C xeon. and i dont blame intel for higher prices. its not their fault, that there isnt anything coming from amd. if its your business, would give stuff away, just because there's no competition?? @ht_addict low detail/texture/shadow etc: http://www.bilder-hochladen.net/show/20df484b41 not even maxed settings: http://www.bilder-hochladen.net/show/d1e97c420b the fact i had a 960 when the game came out last year, i was only able to do 1440p with low settings, after upgrading to 1070 (and most settings on high, texture/LOD on ultra), i could instantly tell the difference, even when running. not even talking about things like HBOA (ambient occlusion) when its dark(er), and stuff like shadows/water reflections etc. (so, if you cant see a difference, please see an eye care professional ;-) the main reason to play on a pc, is the higher IQ settings. otherwise i could have spend 400 on a ps4/Xbone instead of a 1070 (if i was interested on low/med settings)..

Where did you found that info?! As far as I know, HT is not invented by intel

They have a cross-patent agreement. Hyper threading is just the marketing term for Intel's implementation of SMT (simultaneous multithreading). There is nothing unique to Intel about it, and even if it was AMD could use it, like Intel is using x86-64 and FMA.

Not again with the "clocks matter" BS. what a CPU is clocked at has never mattered people, give up on that whole idea. The only thing that matters is how it performs.

Even if you have 5960X 3Ghz, despite how it performs you overclock it or the 8core CPU will get rekt by a higher clocked 4core 6700k. So yes, it still does matter unfortunately.

Even if you have 5960X 3Ghz, despite how it performs you overclock it or the 8core CPU will get rekt by a higher clocked 4core 6700k. So yes, it still does matter unfortunately.

That used to be true with a lot of older games. The only reason for fast single threaded performance is games, yet modern engines are tweaked around eight cores/threads because of the consoles. Even old Bulldozers are doing well with most properly made engines. In everything else the multicore chip is miles ahead anyways. A 3.2/3.3GHz chip with Broadwell/Skylake IPC would wreck most quads in most modern situations, and if a miracle happens and it's priced close to them, then the choice is a no-brainer. I have to say here that after reading Anandtech's dive in to the Zen architecture, I'm very curious to see the extent of I/O integration into the CPU itself, and the effect it will have in overall system latencies.