AMD Ryzen Architecture Features Revealed
John Williamson / 4 years ago
As I’m sure you’re aware, AMD recently hosted an event called New Horizon which included details about the company’s new CPU architecture, codenamed Ryzen. In addition to performance comparisons with the Intel Core i7-6900K, AMD discussed key parts of the architecture which make for very interesting reading. The SenseMI technology consists of Pure Power, Precision Boost, Extended Frequency Range, Neural Net Prediction and Smart Prefetch. The PurePower monitors the chip’s thermal readings, speed and voltage to use energy more wisely which leads to lower power consumption without impacting the performance.
The Precision Boost feature works alongside Pure Power to optimise performance and enact on-the-fly clock adjustments without queue drains which provide higher performance at the same power draw. Also, the CPU can be tuned with 25MHz increments. Next up is the Extended Frequency Range (XFR) which automatically scales the frequency based on the thermal loads and cooling apparatus while permitting frequencies beyond the standard Precision Boost limits. The Neural Net Prediction involves a true artificial network inside every CPU and anticipates pre-load instructions to choose the most effective path through the CPU. This system is based on a model of software code execution. Finally, the Smart Prefetch employs sophisticated learning algorithms to learn application data access patterns. In turn, the CPU prefetches data into the local cache to make it ready for immediate usage.
AMD’s new architecture is extremely advanced with a fantastic set of features. This bodes well for the future and I’m looking forward to hearing more about the inner workings of the Ryzen.