Thursday, August 18th 2016

AMD Demos Breakthrough Performance of the ZEN CPU Core

Press Release by

Aug 18th, 2016 09:05 Discuss (187 Comments)

At an event last night in San Francisco, AMD provided additional architectural details and a first look at the performance of its next-generation, high-performance "Zen" processor core. AMD demonstrated the "Zen" core achieving a 40% generational improvement in instructions per clock, delivering a landmark increase in processor performance.

During the event, AMD demonstrated an 8-core, 16-thread "Summit Ridge" desktop processor (featuring AMD's "Zen" core) outperforming a similarly configured 8-core, 16-thread Intel "Broadwell-E" processor when running the multi-threaded Blender rendering software with both CPUs set to the same clock speed. AMD also conducted the first public demonstration of its upcoming 32-core, 64-thread "Zen"-based server processor, codenamed "Naples," in a dual processor server running the Windows Server operating system.

"The performance and efficiency of our 'Zen' core showcases AMD at its best," said Dr. Lisa Su, president and CEO of AMD. "Over the last four years we have made significant investments to develop a high-performance, multi-generation CPU roadmap that will power leadership products. Customer excitement for 'Zen' continues to grow as we make significant progress towards the launch of new products that will span from the datacenter to high-end PCs."

The "Zen" processor core features multiple architectural advances designed to increase the performance, throughput, and efficiency of AMD's future products. "Zen" is based on a clean-sheet design and features a new cache hierarchy, improved branch prediction and simultaneous multithreading (SMT). These advances will allow the "Zen" core to scale to meet the needs of a broad range of applications, including fanless 2-in-1s, embedded systems, high-performance computing, and the datacenter.

"An engineer may get one chance in their career to work on a project of this size and scope, and maybe never one with as much potential to impact the future as much as 'Zen,'" said Mark Papermaster, senior vice president and chief technology officer at AMD. "With 'Zen' we aim to do what many never thought possible - deliver a 40 percent generational improvement in instructions per clock while maintaining power requirements in line with our previous generation technology."

"AMD invested where it counts, with an x86 core that can scale from PCs to high-performance servers," said Linley Gwennap, principal analyst, Linley Group. "Consumers today expect to get the most out of their systems to create transformative experiences. The versatile design of 'Zen' delivers highly-efficient performance that should provide increased computing capabilities across the spectrum - from devices to cloud computing."

Expected to launch first, the "Zen"-based "Summit Ridge" desktops will utilize the AMD AM4 socket, a new unified socket infrastructure that is compatible with 7th Generation AMD A-Series desktop processors - previously codenamed "Bristol Ridge" - for exceptional performance and connectivity scalability required by AMD partners and customers. The first desktop systems featuring 7th Generation AMD A-Series processors and new AM4 sockets are scheduled to ship in the second half of 2016 in OEM PC designs.

With dedicated PCIe lanes for cutting-edge USB, graphics, data and other I/O, the AMD AM4 platform will not steal lanes from other devices and components. This allows users to enjoy systems with improved responsiveness and benefit from future-ready technologies that the AM4 platform provides with a powerful, scalable and reliable computing solution.

AMD AM4 platform key technology features include:

DDR4 Memory
PCIe Gen 3
USB 3.1 Gen2 10Gbps
NVMe
SATA Express

Additional "Zen" architectural features will be detailed next week in a presentation at Hot Chips 28.

Add your own comment

187 Comments on AMD Demos Breakthrough Performance of the ZEN CPU Core

#151

Nordic

Let's say this lives up to the hype. AMD did it. They matched or are negligibly close in performance with a great price. Is anyone here at all curious what intel will put out in response? Amd hasn't been close to intel in many years. What might they be willing to pull out of their sleeve given the a little competition?

#152

ZoneDymo

james888Let's say this lives up to the hype. AMD did it. They matched or are negligibly close in performance with a great price. Is anyone here at all curious what intel will put out in response? Amd hasn't been close to intel in many years. What might they be willing to pull out of their sleeve given the a little competition?

probably a ton considering they have given us copy paste since the 2600k....
but that should not matter, support AMD if this Zen works out.

#153

dalekdukesboy

FrickVery true.

Hey, don't change my wording :) I said how LOW his expectations were. I reiterate, I'm sorry but I can't help but laugh at how radically everyone is reacting as this little snippet of performance by a demo sent to us by AMD (very impartial source) either is the best thing since sliced bread or is proof the processor will be as good as a square wheel...look, I think as some here said it looks good, be nice if it at least is competitive but a decade of history precedes this where AMD has been woefully outdone, that's just the facts maam. So as I said I hope it's competitive and gives an alternative and maybe gives intel real competition so they can't sell their latest cpus for 2k or so but it's only a demo and we just have to wait and see. But it's promising.

#154

Caring1

noname00Let's not forget the other dirty trick that AMD used in the past - selling a small number of graphic cards/CPUs just after the launch where you could unlock a few more cores. And it didn't always worked, but some of us bought them hoping their CPU could be unlocked.

There was no dirty tricks on their behalf, and there was never a guarantee unlocking would work with those chips.
They were intentionally sold gimped as three core CPU's, with the chance that one could be unlocked, it was a risk many purchasers took, not all were lucky.

#155

cdawall

where the hell are my stars

qubitAs AMD have to underclock - ie hamstring - the Intel CPU to beat it, then it doesn't look good at all for them. They should compare stock with stock running at stock, as that's the true comparison, not a gimped competition.

If the -E CPU is too strong for it, then they should have compared it with the mainstream models instead for a fair comparison.

It looks to me like there will be some useful performance improvements with Zen, but AMD are gonna continue playing catch-up with Intel, like always. Shame AMD squandered their lead with the A64 series a decade ago. They really blew it. :ohwell:

It's somewhat hard to compare an 8 core 16 thread mainstream chip when your competitor doesn't have one. Every single person would have cried if AMD ran it against a 6700k in a multithreading benchmark. Double the cores and thread need to compete with the same no?

#156

HumanSmoke

cyneaterAll I know Zen was designed by Jim Keller who designed the K8 x86-64 the Athlon 64 .. so maybe there is some hope.

I'm continually amazed how many people get the development of this pivotal product in AMD's history wrong. Keller worked on the original K8 project. This was shelved and the actual K8 was designed by a team led by AMD's chief architect Fred Weber.
What did Keller actually achieve at AMD? He led the team at DEC that developed the EV6 bus. AMD later purchased the EV6 IP and Keller continued his work. EV6 now named HyperTransport.
Keller was also a part of the team that developed AMD64, along with Dirk Meyer, David Cutler and Robert Short (both working at Microsoft), and SUSE - who developed the compiler,

cadaveca1. I think that if they could do more than that 3 GHz reliably, they would have. That is, if they are smart. I think you could realistically predict that in order to do what they have seem to have done, the CPU pipeline is going to be relatively short, cache will get hit often, and run hot, kind of like Intels CPUs. That means it's all up to the process, and not AMD. :p

Which brings us to Intel's second gen 14nmFF process versus Glofo's 14nm. A process they have already had issues with meeting clock/power envelope targets with the RX480. Glofo's yield, ramp, and inability to supply top binned parts on previous process nodes is near legendary.

theoneandonlymrkfinally Intel's micro opp and wide core rescources for 1thread tactic is getting countered, hold onto your pants cos imho Kaby lakes getting served its own ass.

I think you'll find that Intel has been spending more resources developing an improved thread scheduling engine. It might not be ready for Kaby Lake which is already shipping, but Cannonlake will surely have it.

To the argument over pricing of Athlon X2 64 and Pentium EE's, pricing was highly variable based on availability since AMD had supply issues. You can find instances where pricing changed on a monthly or weekly basis. Here's a single snapshot I quickly found from my mountain of old magazines

#157

qubit

Overclocked quantum bit

cdawallIt's somewhat hard to compare an 8 core 16 thread mainstream chip when your competitor doesn't have one. Every single person would have cried if AMD ran it against a 6700k in a multithreading benchmark. Double the cores and thread need to compete with the same no?

Yeah, you'd think so, but we've seen 4 core Intel CPUs handily beat AMD's "8-core" siamesed CPUs even without using HT since the IPC is so much better.

In general companies should show fair performance comparisons between their products and the competition's by choosing a competing product that's in a similar price / performance segment. It looks like here they've gone for Intel's very top end, know then can't really compete with it in a fair head-to-head comparison, so have put out some fudged stats to try and look better which are making people like me, @Tatty_One and others suspicious that we have another Bulldozer underperforming fiasco on our hands.

Of course, in the end, it doesn't matter what we speculate here, the official benchmarks at product launch will reveal the truth.

#158

uuuaaaaaa

qubitYeah, you'd think so, but we've seen 4 core Intel CPUs handily beat AMD's "8-core" siamesed CPUs even without using HT since the IPC is so much better.

In general companies should show fair performance comparisons between their products and the competition's by choosing a competing product that's in a similar price / performance segment. It looks like here they've gone for Intel's very top end, know then can't really compete with it in a fair head-to-head comparison, so have put out some fudged stats to try and look better which are making people like me, @Tatty_One and others suspicious that we have another Bulldozer underperforming fiasco on our hands.

Of course, in the end, it doesn't matter what we speculate here, the official benchmarks at product launch will reveal the truth.

I think that they were trying to show that they can now compete in IPC with Intel. About the clocks I agree on them to be conservative at this stage, I mean if they cannot deliver their target clocks (yields etc..) within a reasonable TDP, they will have to lower their target speed, so I think it kinda makes sense for AMD to be conservative with their clock speeds and claims.

#159

cdawall

where the hell are my stars

qubitYeah, you'd think so, but we've seen 4 core Intel CPUs handily beat AMD's "8-core" siamesed CPUs even without using HT since the IPC is so much better.

In general companies should show fair performance comparisons between their products and the competition's by choosing a competing product that's in a similar price / performance segment. It looks like here they've gone for Intel's very top end, know then can't really compete with it in a fair head-to-head comparison, so have put out some fudged stats to try and look better which are making people like me, @Tatty_One and others suspicious that we have another Bulldozer underperforming fiasco on our hands.

Of course, in the end, it doesn't matter what we speculate here, the official benchmarks at product launch will reveal the truth.

And we have seen single core amd chips out perform Intel dual cores (athlon 64 vs pentium d) using the past as a basis for future performance is ignorant in the tech world. This was again a single benchmark showing that they could be competitive on one front. People know how the rest performs we just aren't privy to the information yet. I'll hold judgment until real benchmarks and pricing hit.

#160

cadaveca

My name is Dave

HumanSmokeWhich brings us to Intel's second gen 14nmFF process versus Glofo's 14nm. A process they have already had issues with meeting clock/power envelope targets with the RX480. Glofo's yield, ramp, and inability to supply top binned parts on previous process nodes is near legendary.

Meh. You know what burned my ass was Bulldozer and a certain blonde lady saying 5 GHz. The architecture was capable, sure, but it took forever for yields to reach that level reliably (if they even are now). In hindsight, that blonde heard a good line and ran with it, but it did AMD more bad than good in the end.

I could care less about RX480, and what it does on similar silicon. It's cache organization that really kills a CPU's scaling, and a GPU doesn't really reveal too much about that, IMHO. I hear what you are saying, but I am going to choose to ignore that for now. ;)

What AMD needs, is order to gain consumer acceptance, if a chip that either beats 6700K or matches 6900K. There are two ways to do that. I surmise, however, that what we'll get will be directly between the two.

#161

64K

If AMD scores a touchdown with Zen then they should be rew

cadavecaWhat AMD needs, is order to gain consumer acceptance, if a chip that either beats 6700K or matches 6900K. There are two ways to do that. I surmise, however, that what we'll get will be directly between the two.

What AMD needs to survive is to get PC/Server manufacturers to buy more of their chips. That's all. End of story. It doesn't matter not a a fuck if Zen runs rings around Kaby Lake or not really in the long run. That's what we enthusiasts look at but that's a small blip on the sales radar.

#162

cdawall

where the hell are my stars

64KIf AMD scores a touchdown with Zen then they should be rew

What AMD needs to survive is to get PC/Server manufacturers to buy more of their chips. That's all. End of story. It doesn't matter not a a fuck if Zen runs rings around Kaby Lake or not really in the long run. That's what we enthusiasts look at but that's a small blip on the sales radar.

Bulldozer did that reasonably well for a short period of time before intel pulled their head out of their ass and released some products that could actually multithread.

#163

64K

Bulldozer was a fucking embarrassment for AMD but whatever.

I'm forward thinking to Zen and Kaby Lake and let's see the benches.

Bring on the benches and let's see what's what.

#164

cdawall

where the hell are my stars

64KBulldozer was a fucking embarrassment for AMD but whatever.

I would say it should never have been released to consumers, a throttled up jaguar core would have been a much better choice. In massively multithreaded loads bulldozer has not been touched by intel until haswell, in the server industry that is fine. In the consumer industry when the software developers are lazy its an issue.

#165

64K

Really, it's not about the performance. That's what we enthusiasts look at but it's about getting PC/server manufacturers to buy more of their chips an have a way of selling their product.

#166

Ubersonic

This is huge news, AMDs previous 8 cores could barely beat Intels 4C8T i7's, if they are beating Broadwell 8C16T CPUs clock for clock that's a massive boost and potentially a game changer.

To put this in perspective Intel charge £930 for an 8C16T Broadwell i7 that Turbos to 4GHz, after the FX-9590 fiasco AMD will not dare to charge anything remotely like that, and their top CPUs may even come it at more than 4GHz.

Potentially this could be the dawning of another Athlon XP era of value for money.

64KBulldozer was a fucking embarrassment for AMD

Indeed, it was their Netburst.

#167

Melvis

cdawallI would say it should never have been released to consumers, a throttled up jaguar core would have been a much better choice. In massively multithreaded loads bulldozer has not been touched by intel until haswell, in the server industry that is fine. In the consumer industry when the software developers are lazy its an issue.

I agree, they should of help out longer before bringing us Bulldozer as i found it to be a rush job, as Piledriver that came out like what? less then a yr later was what Bulldozer, or what AMD wanted to bring to the table the first time around, has been a decent performer for AMD. Patients is a virtue!

#168

Frick

Fishfaced Nincompoop

cadavecaWhat AMD needs, is order to gain consumer acceptance, if a chip that either beats 6700K or matches 6900K. There are two ways to do that. I surmise, however, that what we'll get will be directly between the two.

If they pull of that they have beat their own goals. Me I'm glad if they match Haswell.

#169

ZoneDymo

64KBulldozer was a fucking embarrassment for AMD but whatever.

I'm forward thinking to Zen and Kaby Lake and let's see the benches.

Bring on the benches and let's see what's what.

I dont agree, call me crazy but I appreciate/promote companies trying different things whether it works out or not at least they tried something

#170

RealNeil

64KBulldozer was a fucking embarrassment for AMD

Only due to the outrageous claims made by AMD's advertising dept. (didn't AMD fire most of them afterward?)
If they hadn't raised everyone's expectations so high, people wouldn't have been so pissed-off about it.
I had a few Bullys that met my lowered expectations without any problems.

#171

64K

RealNeilOnly due to the outrageous claims made by AMD's advertising dept. (didn't AMD fire most of them afterward?)
If they hadn't raised everyone's expectations so high, people wouldn't have been so pissed-off about it.
I had a few Bullys that met my lowered expectations without any problems.

I think they learned a lesson from that and aren't going to over-hype Zen before release and a part of the Bulldozer hype train was also AMD fans hyping it. Really I can't blame AMD for hyping their product before release. Most companies do it. I just take it with a grain of salt until I see actual reviews from tech sites like this one.

#172

rruff

64KReally, it's not about the performance. That's what we enthusiasts look at but it's about getting PC/server manufacturers to buy more of their chips an have a way of selling their product.

But why would manufacturer's buy them? There needs to be a performance/cost angle. Servers in particular care about efficiency, and I haven't seen too much on that score relating to Zen.

#173

RealNeil

64Kpart of the Bulldozer hype train was also AMD fans hyping it

Like what's happening now?
This is why I suggest a ~steady as she goes~ attitude until an actual release that we can sink our teeth into. Something that is measurable.

#174

64K

rruffBut why would manufacturer's buy them? There needs to be a performance/cost angle. Servers in particular care about efficiency, and I haven't seen too much on that score relating to Zen.

I don't know how AMD is going to increase their market share. In the past they have tried to sell their chips cheap but this is a failing strategy imo. For example do you remember when AMD landed the contracts for all 3 console makers? People were saying that this would turn AMD around and they would once again be profitable and yet they went into the red at record levels after that. I don't know how else that can be explained other than they were not charging MS, Sony and Nintendo enough for their chips. One could still make the argument that AMD did the right thing to gain more traction in the video game industry since most games are made for the console first and later ported to the PC.

The thing is that AMD needs to be charging more. It does little good to have the console market sewed up if they end up bankrupt from their efforts. If Zen is decent competition for Kaby Lake then I hope to see AMD being able to charge as much as Intel. The problem is that to gain market share they may have no other choice but to sell cheaper than Intel and not make a decent profit. It's a vicious cycle at this point and I don't know what they can do about it at this time.

#175

xenocide

64KI think they learned a lesson from that and aren't going to over-hype Zen before release and a part of the Bulldozer hype train was also AMD fans hyping it. Really I can't blame AMD for hyping their product before release. Most companies do it.

4 in 5 AMD fans suffer from Stockholm Syndrome.

Being serious though, there's a difference between promoting an upcoming product and misleading marketing which is what they did with Bulldozer... and Phenom II... and Phenom I... and almost all of their APU's. AMD has a history of using misleading slides and data to indicate their products will be substantially better than they actually are. Say what you will about Intel, but even if it's not the best news they stay reasonable with their claims, and are always pretty accurate.

Add your own comment

AMD Demos Breakthrough Performance of the ZEN CPU Core

187 Comments on AMD Demos Breakthrough Performance of the ZEN CPU Core

Latest GPU Drivers

New Forum Posts

Popular Reviews

Controversial News Posts

AMD Demos Breakthrough Performance of the ZEN CPU Core

Related News

187 Comments on AMD Demos Breakthrough Performance of the ZEN CPU Core

Latest GPU Drivers

New Forum Posts

Popular Reviews

Controversial News Posts