AMD Wants to Give Windows a Hidden 'Superpower' to Better Optimize Ryzen

AMD is preparing an update aimed at improving how operating systems will manage next-generation Ryzen processors. The innovation concerns support for a new feature called CPPC HighestFreq, integrated into the Collaborative Processor Performance Control (CPPC) framework, already widely used by Ryzen CPUs for dynamic performance management.

Currently, operating systems do not receive precise information directly from the firmware regarding the maximum frequencies achievable by individual cores. To determine which cores are the most performant, Windows and Linux rely on abstract performance values and interpolation models that estimate boost clock behavior. However, this approach is becoming less accurate with the latest Ryzen architectures. Modern AMD processors indeed adopt advanced and not perfectly linear boosting mechanisms, with behavioral differences among various cores. In practice, not all cores reach or maintain the same operating frequencies under the same conditions.

The CPPC HighestFreq function aims to address this limitation by allowing the firmware to communicate directly to operating systems the actual maximum frequencies supported by individual cores. The modification is already being prepared for the AMD P-State Linux driver and should become part of the upcoming ACPI 6.7 specification.

The main advantage relates to the operating system's scheduler, which can more accurately identify the faster cores and assign them workloads that are sensitive to frequencies. Applications like video games, interactive software, or particularly heavy threads could thus benefit from a better distribution of tasks.

AMD already uses the concept of "preferred core," which means selected cores are deemed more performant within the processor. However, without direct data on the actual frequencies, the scheduler does not always manage to choose the ideal core for every operational scenario.

The new implementation is not expected to revolutionize performance in absolute terms, but it could contribute to increasing the overall responsiveness of the system and the efficiency in thread management. In scenarios such as gaming, even small improvements in scheduling accuracy can translate into better performance consistency and lower latencies.