Thursday, August 18th 2016

AMD Demos Breakthrough Performance of the ZEN CPU Core

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

#51
Steevo
NokironJust to nitpick, it's 4Ghz not 3,7. Turbo Boost 3.0.
ark.intel.com/products/94196/Intel-Core-i7-6900K-Processor-20M-Cache-up-to-3_70-GHz
Yes, if you install the TBD 3.0 driver and have a board that supports it, and only for some cores if they are efficient enough, and only during single threaded applications running.

So all else being equal, in modern games that utilize 4 cores or more, 3.7 is as high as it will go under the right conditions, and if all cores are fully loaded and the TDP is hit it may be 3.2Ghz for all cores.


But the big gripe against AMD has been

1) IPC
2) cache latency
3) pipeline stalls
4) math performance
5) Power consumption


For years its been, pick one, and we will improve that some. Now they seem **SEEM** to have fixed their issue and are trying to show that clock for clock against a current generation Intel, they are faster in a real world scenario/benchmark that is common.
Caring1I posted this on the previous page, he obviously didn't read the thread and was eager to post his opinion.
I did miss your comment. I was busy looking up exactly what you had posted to see if AMD was using biased hardware and software to show their offering better.
Posted on Reply
#53
Fx
I have been waiting for years to build a new gaming rig. It will be Zen-powered.
Posted on Reply
#54
dwade
Couldn't care less about clock for clock comparison since they are 2 different architectures. Only metric that matters is performance per watt.
Posted on Reply
#55
iO
Still waiting for someone to comment: "They used the open sourced Blender so they can cripple the code on the Intel system!!!1!"
Posted on Reply
#56
GhostRyder
Good for AMD, if this all turns out to be completely true. One thing I still await is how its going to be clocked and how it overclocks. However, if it is a little below it when overclocked and at stock yet a much lower price it will be a game changer since the 8 core/16 thread CPU's from Intel cost quite alot. Still though, this needs to do well otherwise they may not have a chance to recover.
Posted on Reply
#57
lanlagger
"oh it is lsower than broadwelll because it is 3.0 ghz, but broadwell goes..." lets clear one thing - performance dos NOT matter! what matter is price/performance in a segment... till this day AMD had nothing in i3 and above segment - so intel only competed with him self (tried to lure in someone to upgrade their older intel system - for basically no gains)... now I do not believe that 8c/16t zen will compete with 8c/16 broadwell... but you damn right it will compete with 4c/8t i7's and anything below... and it looks good for AMD - because prices are also inflated (slowly) in that segment too... and forget about product pricing as "welll - there is costs and RD, and sales expenses then we add up all this , plus taxes, logistics etc" - that is what they tell you, but in reality it is more like: "so we have a product... now how high can we push the prices so potential clients will still buy? what are competittion prices?"
AMD will give us noting for free ( it wont sell us 4c/8t i7 killer for existing FX price), but it will make intel get back to earth with his pricing... look how out of hand intel got with his latest HEDT - zero gains (even overclocks worse than previous gen) and price bumps from +10 to +60%... why? - obviously not because "RD, taxes, logistics and production" and other bs, because it is not even the latest artchitecture and same architecture in xenon line did not receive any price bumps (but still got some performance bumps because of core count increase)
Posted on Reply
#58
Fouquin
MelvisBecause AMD always have no matter what! even when AMD was king of the hill they still sold there CPU's cheaper.

I think you are the one that needs a history lesson :p
Sorry, but you are completely and provably wrong. Here's a wonderful price breakdown from June 2005, a year after AMD launched the 939 FX chips, showing their less-than consumer friendly pricing structure when they were on top in performance:

Posted on Reply
#59
noname00
iOStill waiting for someone to comment: "They used the open sourced Blender so they can cripple the code on the Intel system!!!1!"
There are plenty of ways to cripple the Intel CPU. They can use DDR4 memory in single or dual channel mode, clocked at 1000 MHz and CL16 instead of quad channel at 2400 MHz and CL14. Or they could just simply lie about the frequencies used, as no frequency was showed at all in the video.

Let's just wait for Zen to be launched, and to see some actual reviews. That's what I will be doing. Kaby Lake and Zen shootout.

I also want to see how the quad core version performs.
Posted on Reply
#60
Melvis
FouquinSorry, but you are completely and provably wrong. Here's a wonderful price breakdown from June 2005, a year after AMD launched the 939 FX chips, showing their less-than consumer friendly pricing structure when they were on top in performance:

Erm no, hate to tell you and about 10 000 other people on this forum will tell you just the same thing that AMD was cheaper on the 939 socket, the FX57 and 60 was around $1000-$1200 and the Pentium EE was around $1500 AUS, its just a fact :)

Anyone can post a pic like you did that is from god knows what source, and from even before launch date where in actual fact I never saw any CPU from AMD (price to performance comparison) be more then Intel.
Posted on Reply
#61
noname00
FouquinSorry, but you are completely and provably wrong. Here's a wonderful price breakdown from June 2005, a year after AMD launched the 939 FX chips, showing their less-than consumer friendly pricing structure when they were on top in performance:

Thank you.
Posted on Reply
#62
Fouquin
MelvisErm no, hate to tell you and about 10 000 other people on this forum will tell you just the same thing that AMD was cheaper on the 939 socket, the FX57 and 60 was around $1000-$1200 and the Pentium EE was around $1500 AUS, its just a fact :)

Anyone can post a pic like you did that is from god knows what source, and from even before launch date where in actual fact I never saw any CPU from AMD (price to performance comparison) be more then Intel.
It's just a completely unsupported fact, or no wait that's called a conjecture. Also, we're not discussing AUS pricing, we're discussing USD pricing. Your pricing system is never a reflection of retail costs in the US, nothing against Australia but you almost never see proper retail pricing on anything.
Posted on Reply
#63
Nokiron
MelvisErm no, hate to tell you and about 10 000 other people on this forum will tell you just the same thing that AMD was cheaper on the 939 socket, the FX57 and 60 was around $1000-$1200 and the Pentium EE was around $1500 AUS, its just a fact :)

Anyone can post a pic like you did that is from god knows what source, and from even before launch date where in actual fact I never saw any CPU from AMD (price to performance comparison) be more then Intel.
965EE was cheaper here, not by much. As said though, MSRP was lower.

That source is rechreport.
Posted on Reply
#64
Grings
nobody gave a shit about the flagship overpriced processors from either platform back then, as they could overclock the cheap ones easier
Posted on Reply
#65
Fouquin
Gringsnobody gave a shit about the flagship overpriced processors from either platform back then, as they could overclock the cheap ones easier
Everybody gave a shit because it jacked prices on the cheaper chips.
Posted on Reply
#66
acperience7
Gotta admit, I'm on the hype train with this one. I'm itching for a full system upgrade.
Posted on Reply
#67
cdawall
where the hell are my stars
Still has roughly half the cache of the Intel chips out right now. I'm betting these will compete, but not quite beat them overall. It has me worried that they limited it to 3ghz, I wonder how they will clock.
Posted on Reply
#68
alexsubri
AMD Stocks are rising up to ~5% today! Glad I bought my shares cheap ^___^
Posted on Reply
#69
OneMoar
There is Always Moar
remember people don't trust AMD they have a history of creative truth bending

watch for real and proper benchmarks before believing a single syllable of anything they say

multi thread benchmarks are a TRAP it doesn't show what the real performance improvement are and makes AMD's chips look really good
Posted on Reply
#70
$ReaPeR$
well.. finally. this looks very promising but truth be told, with no independent reviews this is just hype. i really hope AMD did it right this time.
Posted on Reply
#71
OneMoar
There is Always Moar
$ReaPeR$well.. finally. this looks very promising but truth be told, with no independent reviews this is just hype. i really hope AMD did it right this time.
I doubt it the last so called 'leaked' benchmark was ashes and we all know how much that (and I use this term very loosely)`game` favors lots of threads
theres been no superpi or luxmark single thread or cinbench
Posted on Reply
#72
Jism
OneMoarremember people don't trust AMD they have a history of creative truth bending

watch for real and proper benchmarks before believing a single syllable of anything they say

multi thread benchmarks are a TRAP it doesn't show what the real performance improvement are and makes AMD's chips look really good
Then what is a solid test then? If anyone is buying an 8 core intel or AMD equivalent, it's not for their single-core-performance but their multithreaded, where maximum amount of cores and / or threads count, not just 2 or 4 cores alone.

AMD did a test, and proved, clock by clock to be a little faster then a recent Broadwell-E model running on 3Ghz. Now how that test has performed does'nt really matter. You cant really cheat a multi-threaded test unless you cripple some cores, multipliers and choice of memory. If AMD does that in order to make the AMD product look better, then they will fail eventually and proberly be sued for promissing something they did'nt deliver.

Architect on K10 (Jim keller) was lead on that project for a few years. They defenitly know what they are doing my friend.
Posted on Reply
#73
alexsubri
Here is the YouTube demo of Zen playing the new Deus Ex @ 4k resolution with an R9 GPU.

[youtube]
[/youtube]
Posted on Reply
#74
OneMoar
There is Always Moar
alexsubriHere is the YouTube demo of Zen playing the new Deus Ex @ 4k resolution with an R9 GPU.

[youtube]
[/youtube]
O look a DX12 title where cpu performance doesn't matter nearly as much

don't be dense people

always assume that AMD(or any vendor) is lying they have a history of it they are not to be trusted they are guilty until proven innocent
full stop on that hype train please
Posted on Reply
#75
$ReaPeR$
OneMoarI doubt it the last so called 'leaked' benchmark was ashes and we all know how much that (and I use this term very loosely)`game` favors lots of threads
theres been no superpi or luxmark single thread or cinbench
look, even marketing idiocy has its limits, there has to be some truth at least. i doubt that zen will be faster, but imo it will be within 10% of the corresponding intel counterpart. at this point that will have to do.
Posted on Reply
Add your own comment
Jan 23rd, 2025 20:28 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts