AMD Zen processors...Are you interested?

[Jimmy Thick]

The upcoming preview will happen next month and there should be a lot of pomp and fanfare about the new AMD Zen processors but do we as gamers even care about AMD anymore?

Example.

Take the I5 6600k which is the gaming standard dollar for power right now. It has for the most part a $210 price point more or less.

In my opinion, AMD's first Zen release has to be as good performance wise, at the same or cheaper price point and not draw as much power. If they cannot accomplish all those goals, and it has to be all those goal you can't skimp because then by default the Intel option is superior, then Zen is going to fail.

I think if AMD wants to play, they have to deliver big time with Zen, personally I don't like their chances.

Discuss.

[Puupi]

Halt the "discussion" until they have actually published the processors.

[Droodid]

AMD has always been the processor that is cheap and gets the job done if you're on a tight budget (until recently, where they've just fallen behind). I think Zen will up that to continue the trend of a decent performing, cheap CPU, but I don't think they're going to aim for the price/performance of Intel. That's not their goal.

[Crillam]

Tbh, it is AMD so I don't really care.

[Lathais]

I am pretty sure the Zen chips will come close to Haswell performance. If they can get there, at half to three-quarters the cost of a Skylake chip, it will be a consideration again for sure.

[Artorius]

The main 8C/16T SKU won't be as cheap as 210 USD no matter how cheap it is. I'd expect something between 350~500 USD.

Also, "we gamers" care about Zen. Zen is a whole new uarch with emphasis on closing the IPC gap and giving AMD its profits back. If we assume that all SKUs will have SMT, the SR5 SKU is likely to be a 6C/12T part for less than 300 USD and this would be pretty appealing if the L3 cache doesn't suck with 2 cut-down cores.

Jim Keller worked on Zen, Intel has said that it'll make a huge impact in 2017 so you should be pretty hyped about it.

Power consumption depends more on Samsung/GloFo than AMD itself at this point though, I'd expect it to work very well under moderately high clocks like ~3.4 GHz, but it'll turn into a fireball if you OC it just like Intel's 8 cores do. Zen cores are also wider than Skylake cores so if they managed to make the whole thing consume less (it's rated at 95W), that's actually pretty amazing.

Also, AMD is apparently licensing some Radeon stuff to Intel. It was apparently signed this week. Intel wanted to make their iGPUs better and doing this is honestly the smartest thing they could've done. Hopefully this means that AMD will have access to Intel foundries if they ever decide to open it to other companies, which would be amazing for both sides.

[ComputerNerd]

From a competition point of view, there is no reason not to be interested.
Everyone who puts any thought into what they buy should care.

AMD has always been the processor that is cheap and gets the job done if you're on a tight budget (until recently, where they've just fallen behind). I think Zen will up that to continue the trend of a decent performing, cheap CPU, but I don't think they're going to aim for the price/performance of Intel. That's not their goal.

I think part of that is the multi-threading ideal ahead of their time, that the software is barely taking advantage of.

[Fascinate]

I have an odd feeling about zen, i think its gonna do great in synthetics but it wont translate into CPU intensive games. My guess is single threaded synthetics it will do pretty good, but the first WoW benchmark we see will put it at ~ivy bridge levels or slightly above.

[Maduk]

The CPU-market won't change.
Zen will only be another option to take, in which performance sector, that we will see next year.

[chazus]

Am I interested? Yes.

Am I expecting anything special? No really.

After 3-4 instances of "We're definitely bringing new tech to the table" and being utterly disappointed, I'm not holding my breath until I see retail unit benchmarks. Not keynotes. Not leaks. Not technical samples.

[TyloBedo]

Yeah, I'm pretty interested. Mostly because I have a fair amount of money invested into AMD though so if zen surprises I've got a lot to gain.

[AztechZero]

I'm interested but, I'm not expecting Zen to set the world ablaze. I'm hoping for at least Haswell or Broadwell IPC though.

[Lillyth]

Its AMD, cannot say I care. been going with Intel since forever and AMD has nothing that would convince me to change from this tradition.

[Sturmbringe]

I believe Zen will be a disappointment/fail.

A strong and competitive AMD that manufacture fast and competitive products is good for the consumer. However, I don't think that Intel and NVIDIA will allow this to happen ever again.

In addition, I am very unimpressed by Lisa Su.

So, IMO this discussion is a waste of time.

[Blackmist]

I'd like to see what they can do, but I'm not exactly waiting for them. If they come out before I get angry enough at my ageing PC, I'll take a look, but chances are I'm going with whatever i5 K series CPU is current when I get one.

AMD showed the Zen "beating" an underclocked i7, to show it can do more IPC, but how fast will the Zen go? If it maxes out at 3GHz and needs 200W, how is that any better than the shit they have now? I quite like the AMD GPUs (RX 470 and 480), it seems much cheaper than the 1060, for similar performance, but for CPUs Intel hold all the gaming performance.

[potis]

Same boat as everyone.

Would love to see what they can bring, but not because i plan to buy or support them, cause maybe if there is competition, things can do back to progressing faster and not costing extra because there is no competition

Rumor wise, they are 2 years behind, so if they do the pricing correctly, as example i5 4570/4590/6400 performance for 150$ and not 220$ it might be interesting.

[Evildeffy]

Rather than making uninformed conjectures as most of you are doing right now...

You could always attempt to read up on these things, most of this has already been discussed and whilst the info isn't final you can read it here:

http://wccftech.com/find/?q=zen

With more specific information here:
http://wccftech.com/amd-zen-summit-r...etails-leaked/
http://wccftech.com/amd-zen-summit-r...verclock-cpus/
http://wccftech.com/amd-zen-summit-r...-launch-rumor/

The current information that is known is that it will take on a Broadwell-E chip head to head and edge it out slightly.
At the same speeds and core amount.

Considering that Broadwell-E is UNDER 2% of Sky Lake in regards to IPC it is a vast improvement and if the rest holds true it should, logically speaking, close in on Sky Lake (and therefore Kaby Lake as there is NO IPC difference between Kaby Lake and Sky Lake) whilst offering double the amount of cores for the same money.

In essence if they can hold it true at the currently specified numbers, which AMD is still sticking to, you'll be getting an 8C/16T with a 95W TDP that is equal to an i7-6850K CPU for the price of a Sky Lake i7-6700K but potentially double the performance in multi-threaded applications for the same price as aforementioned Core i7-6700K.
Gaming performance of course can be extrapolated by the same amount as it is tied to the IPC and speed of the processor and since they are close to each other it will be pretty much in the same field as Sky Lake but double the cores so the content creators will get a free boost for the same money as a Sky Lake.

As far as overclocking goes... every single Hexa/Octa-core has a physical wall of frequency and heat generation it has to deal with, Intel's Hexa/Octa-cores are fireballs when overclocked I expect ZEN to be no different, this has little to do with brand and more with actual physical limits.

ANY tech nerd should be extremely excited about ZEN IF the information and promises that have been shown and said so far are real.
It is a market disruptor and it is something Intel wouldn't have an answer for for 2 years, they'd have their Deca-Core only and considering that goes for 1.700 USD/EUR vs. an Octa-Core that goes for 350 - 500 USD/EUR the choice SHOULD be easy for most people and businesses.

Dismissing it just because it's AMD is literally the most stupid thing you could do.
You WANT AMD to succeed, you NEED that competition so prices get lowered and YOU get the benefits of such.

Quick addition:
Current plans for AMD are 3 "model ranges" much like Intel's Model range of i3 / i5 / i7

The AMD variants will be called SR3 / SR5 / SR7.

SR3 = 4C/8T which has the same performance potential as the i7-6X00(K) editions (Sky Lake), clock for clock and core for core whilst prices much like the Intel Core i3 range. (Sky Lake)
SR5 = 6C/12T which has the same performance potential as the i7-68X0 editions (Broadwell-E), clock for clock and core for core whilst prices much like the Intel Core i5 range. (Sky Lake)
SR5 = 8C/16T which has the same performance potential as the i7-68X0 editions (Broadwell-E), clock for clock and core for core whilst prices much like the Intel Core i7 range. (Sky Lake to entry Broadwell-E)

How does this not get you excited at the possibility?
If it fails like before then yes it'd suck ass and Intel will milk us for a long ass time but if it succeeds, and AMD has said so far that their ZEN architecture is performing beyond what they expected, it is a new and exciting time for whooping ass as choices open up and diversity returns.

[Fascinate]

Like i said in another thread, i want a 8 core 16 thread zen with haswell IPC for 300 bucks. If they can do that, they got a winner.

[Artorius]

Like i said in another thread, i want a 8 core 16 thread zen with haswell IPC for 300 bucks. If they can do that, they got a winner.

The IPC you want is pretty much irrelevant when they've already shown us that they can edge out BW-E, and for the pricing I'd say that's unrealistic even for the locked SKUs if those exist. The 6900K costs 1000 USD, why in the world would AMD sell their equivalent CPU for less than a third of the price? It doesn't make any sense whatsoever.

Sure, you guys perhaps argue that Blender is a best case scenario for Zen or that they had to downclock the 6900K to make the comparison which might be a bad sign. But they were using ES silicon which is always clocked lower than the final product for debugging reasons.

I'm going to quote someone because it's pointless to write the same stuff again:

Zen has similar pipeline stages than BD, probabily same or lower FO4. They simplified the INT scheduler: from 4 queue switch to 6 single queues. Why they did that? They did that and the FO4 is increased? Are they crazy? I don't think so... If the FO4 is the same or similar, the FMAX should be the same on 28nm. But the 14nm has higher transconductance, lower leakage and parasitic capacitance. And AMD declared same consumption than excavator core at same frequency. Given the known data i think that it is very plausible. So FMAX in the same ballpark than XV (4.3GHz). You don't buy this only because INTEL does not reach these clocks with 8 core... But AMD managed to have a 4.1/4.3 core on the 28nm BULK in 95W... I have an i7 3820 (3.6/3.9) on my pc at work, with 130W TDP. On 32nm BUKLK INTEL couldn't reach even 4GHz in turbo... Why? 28nm TSMC is better than INTEL 32nm HKMG? No. Excavator has lower FO4. And also Zen. If you don't get this, you will not ever agree with me...

8/16 probabily only FMAX is on the 4-4.3GHz... The FX 8370E on 32nm SOI, is 95W and is 3.3/4.3 turbo max. I expect at least this for the 8c Zen (3.3 base and 4.3 turbo), maybe higher base and slightly higher turbomax. We are talking of 2 full nodes... I think that the FMAX could be higer than 4.3... For the base frequency, it depends on the energy saving techniques. 14nm FF has very low leakage. It's estimated than in each instant, 85-90% of the transistors are off and drain only leakage. If the leakage is low, the power budget can be given to raise the base clock... I think that at least 3.5/4.5 for the 95W and 4.0/4.6 for the 125W, if it will be ever produced...

[zeropeorth]

Since I have a z170 board, irrelevant for me, but I've had amd before when they were the better chip. They don't have to fit into your requirements though to succeed. They just need to be competitive outside the lower end/value market for their own margins.

If I was in the market, I'd consider it after I see some proper reviews and benchmarks. I don't understand why people get so brand obsessed; buy what makes most sense for you as long as it has a reasonable build quality, shoulldnt have to be intel/nvidia and at one time amd/ati had the lead and they should again. I want to see amd give both intel and nvidia strong competition because it is much better for us, the consumers in both prices and capabilities.