AMD CTO on Zen 3, 7nm+ CPU Architecture: Will Primarily Leverage Efficiency With Some Modest Performance Opportunities
AMD will be introducing their next-generation Zen 2 based processors next year which will first be featured on the EPYC Rome processors. These will be the first high-performance CPUs to feature a 7nm process node and while they are expected to deliver great improvements in performance and efficiency, AMD themselves have already started giving us hints about what the CPU architecture that comes after Zen 2 may look like.
AMD Zen 3 Will Use The 7nm+ EUV Process Node, Primarily Going To Leverage Power Efficiency With Modest Performance Uplift
It’s no doubt that AMD will have the first GPUs and CPUs utilizing the TSMC 7nm process node. Their Vega 20 “Instinct MI60” graphics accelerator is shipping later this year while EPYC Rome processors for the server market would be introduced next year. Now we got to hear a lot of architectural details from AMD regarding their Zen 2 architecture and the EPYC Rome CPUs that can be seen here. Zen 2 is currently looking to be one of the biggest CPU performance and CPU efficiency since Zen itself back in 2017.
After Zen was introduced to the market, a year later we got Zen+. A slightly efficient and optimized Zen architecture which was based on the 12nm process node instead of the 14nm process which Zen originally utilized. AMD’s latest roadmap now confirms that after Zen 2, we would get Zen 3, Zen 4 and even, Zen 5.
Details for Zen 2 are available but AMD’s CTO, Mark Papermaster, has unveiled what to expect from the Zen 3 CPU architecture. First and foremost, AMD states that their claimed performance and efficiency numbers are for a real product while TSMC uses their own methods. Hence, the difference between the power draw and performance uplift estimates from both companies. A node and an actual product are two varying things and you can’t compare them both.
AMD CPU Roadmap (2018-2020):
|Ryzen Family||Ryzen 1000 Series||Ryzen 2000 Series||Ryzen 3000 Series||Ryzen 4000 Series|
|Architecture||Zen (1)||Zen (1) / Zen+||Zen (2)||Zen (3)|
|Process Node||14nm||14nm / 12nm||7nm||7nm+|
|High End Server (SP3)||EPYC 'Naples'||EPYC 'Naples'||EPYC 'Rome'||EPYC 'Milan'|
|Max Server Cores / Threads||32/64||32/64||64/128||TBD|
|High End Desktop (TR4)||Ryzen Threadripper 1000 Series||Ryzen Threadripper 2000 Series||Ryzen Threadripper 3000 Series (Castle Peak)||Ryzen Threadripper 4000 Series|
|Max HEDT Cores / Threads||16/32||32/64||32/64?||TBD|
|Mainstream Desktop (AM4)||Ryzen 1000 Series (Summit Ridge)||Ryzen 2000 Series (Pinnacle Ridge)||Ryzen 3000 Series (Matisse)||Ryzen 4000 Series (Vermeer)|
|Max Mainstream Cores / Threads||8/16||8/16||8/16?
|Budget APU (AM4)||N/A||Ryzen 2000 Series (Raven Ridge)||Ryzen 3000 Series (Picasso) Zen+?||Ryzen 4000 Series (Renior)|
“TSMC may have been measuring a basic device like a ring oscillator — our claims are for a real product,”
“Moore’s Law is slowing down, semiconductor nodes are more expensive, and we’re not getting the frequency lift we used to get,” he said in a talk during the launch, calling the 7-nm migration “
Looking ahead, a 7-nm-plus node using extreme ultraviolet lithography (EUV) will “primarily leverage efficiency with some modest device performance opportunities,”
It is also stated that AMD would be utilizing the state of the art 7nm+ EUV (Extreme Ultraviolet Lithography) technology from TSMC to manufacturer their Zen 3 based processors. It makes sense since we expect the chips to launch around 2019-2020. The new process node along with the new optimized chip design is effectively stated to deliver better power efficiency, albeit with modest performance gains. Considering this would be the Zen+ of Zen 2, the statement makes sense. It’s still too early to tell what the final products would be like but AMD definitely would have a lot of experience with 7nm to work with when giving the final touches to next-generation Zen 3 CPUs.