AMD "Zen" Core Architecture

The Revolutionary Architecture Behind AMD Ryzen™ and EPYC™ Processors

Performance

With a core engine that supports simultaneous multi-threading for future-looking workloads; a leading-edge cache system and neural-net prediction, to help lower effective latency; a sharp eye on efficiency, for impressive performance-per-watt leadership, “Zen” is a scalable architecture that can be continuously improved. 

Scalability

AMD employed a radical new idea for x86 processors: chiplets. Instead of building larger monolithic dies, AMD invested in a strategy to use processor building blocks called chiplets. Each chiplet houses a number of “Zen”-based cores, and more chiplets can be added to a package to create a higher performance model processor. 

Efficiency

At AMD, our core design is an undertaking of continuous optimization. Decoupling our core and I/O development processes enabled us to shrink the CPU die and optimize variants for performance or energy efficiency. The placement of every transistor and the allocation of every microwatt of power from processor to platform demonstrates the AMD commitment to efficiency.

“Zen 5” Architecture

Cutting-edge 4nm manufacturing technology enables the world’s most powerful – and most efficient – desktop processor lineup with AMD Ryzen 9000 Series processors. Improvements include Improved branch prediction accuracy and latency; higher throughput with wider pipelines and vectors; and deeper window size across design for more parallelism. As a result, single-threaded IPC is increased by about ~XX% gen-over-gen.

“Zen 4” Architecture

AMD Ryzen created on leading 5nm manufacturing technology, AMD Ryzen 7000 Series processors boast a maximum clock speed up to an impressive 5.7 GHz⁷. Thanks to major redesigns of key portions of the chip like the front end, execution engine, load/store hierarchy, and a generationally-doubled L2 cache on each core, the chip can deliver up to a 13% IPC increase over its predecessor. When combined with the 800 MHz clock increase over last gen, this can add up to 29% more single-thread performance.

4^th Generation AMD EPYC processors include up to 128 “Zen 4” or “Zen 4c” cores with exceptional memory bandwidth and capacity.  The innovative AMD chiplet architecture enables high performance, energy-efficient solutions optimized for different computing needs. These cores represent a significant advancement from the last generation, with new support for highly complex machine learning and inferencing applications.

“Zen 3”

“Zen 3” debuted in AMD Ryzen 5000 Series desktop processors and increased the maximum clock speed to 4.9 GHz. This comprehensive design overhaul delivered a further 19% IPC. It also transitioned to a new "unified complex" design which dramatically reduced core-to-core and core-to-cache latencies. Latency-sensitive tasks like PC gaming especially benefited from this change, as tasks now have direct access to twice as much L3 cache versus "Zen 2."

“Zen 3+” transitioned to the newer 6nm manufacturing process. Focused on mobile users, efficiency was a strong focus, along with delivering high performance and performance-per-watt. This manifested in the AMD Ryzen 6000 Series processors for mobile. Laptops with these processors have demonstrated up to 29 hours of video playback on battery.8 In addition, they deliver exceptional performance in thin and light laptops.

“Zen 2”

The Ryzen 3000 Series desktop processors benefited from a major core redesign, doubling up the L3 cache capacity (up to 32MB), floating point throughput (to 256-bit), OpCache capacity (to 4K), and Infinity Fabric bandwidth (to 512-bit). It also featured a new TAGE branch predictor. All of these improvements contributed to a very substantial 15% IPC increase, and with these processors benefitting from the new 7nm manufacturing node, maximum clock speeds climbed to 4.7 GHz. 

“Zen”

The original launch of the "Zen” architecture in the Ryzen 1000 Series desktop processors featured clock speeds up to 4 GHz, and were manufactured on the 14nm manufacturing node. This was followed the next year with the Ryzen 2000 Series featuring updated “Zen+” architecture, which was die-shrunk to the 12nm node and delivered higher clock speeds with about 3% higher IPC (instructions per clock) compared to its predecessor. Despite this modest increase, it delivered up to 15% higher gaming performance due to updates like Precision Boost 2 and XFR 2, thanks in part to a clock speed increase up to 4.3 GHz.