AMD started off its Zen 3 launch event by promising a 19 percent Instruction Per Clock improvement over Zen 2. That’s quite the improvement. Indeed AMD is promising to have the fastest single-threaded performance overall. This coupled with the Ryzen 5000 chips higher frequency means that AMD looks like it’s gunning to wrestle the overall performance crown from Intel.
This impressive improvement comes from a number of areas, although the benefits of the unified 8-Core Complex (CCX) is certainly key, which leads to reduced latency and better access to the L3 cache.
There are also performance improvements from increasing the number of Load and Stores compared to Zen 2. AMD has also added more branch prediction bandwidth to tackle back to back predictions—something it is calling ‘zero bubble’. It has also made the Zen 3 floating point and integer execution wider.
Lisa Su went on to point out that ‘It’s our largest increase generation over generation since we launched the Zen family.’ This will have an impact on the performance in serious workloads, but as IPC is so important to gaming, it should also mean that AMD could finally close ground with Intel and potentially overtake it.
Obviously we’ll have to wait and see how the new Ryzen 5000 chips perform once they’re in our labs.
Available globally on November 5, the new Ryzen 5000 series is made up of the Ryzen 5 5600X, Ryzen 7 5800X, Ryzen 9 5900X, and Ryzen 9 5950X.