Today AMD introduced ZEN 3, 7nm processing power ready to roll in Q4 2020. ZEN 3 delivers higher max boost than ever before, significant IPC uplift, and a whole new core layout and new cache topology. AMD confirmed that ZEN 4 was on-track with a 5nm design as well – it would appear that, as one would expect, AMD will “not let up” going into the future.
AMD made a lot of claims today for ZEN 3. They suggested that ZEN 3 was 2.8x more efficient than Intel’s Core i9-10900K. They noted a 2.4x performance per watt boost over first-gen Ryzen hardware. The Zen 3 was claimed to be boosted 24% performance per watt over Zen 2. AMD claimed they’d accomplished increased loads and store over ZEN 2, wider execution capabilities, and more branch prediction bandwidth (“Zero Bubble” branch prediction).
ZMD claimed that they’ve improved “every aspect of the CPU” in the same 7nm node. This means higher frequency, higher IPC (19% higher than Zen 2), lower latency on top, and better core-to-core communication thanks to a new 8-core complex and a new processor layout.
With ZEN 3 core architecture, AMD said they’d accomplished 2x reduced memory latency for gaming with direct access L3 cache tech – and every core can address 32GB of L3 cache when needed. With this new layout, core and cache communication is accelerated for gaming – all 8 cores can get what they want of the 32MB of L3, when they want. Every core can now communicate to the cache on the chip without going off die.
Stay tuned today as we continue to follow AMD’s big event and watch the company blast of into the future with the latest in processor technology upgrades. This includes ZEN 3, teases of ZEN 4, and the Ryzen 5000 series.