AMD details which is the best core for Ryzen Master and which is the best core for Windows 10

Explained by AMD why the best core of the Ryzen 3000 in the Ryzen Master tool and in Windows 10 is not the same.

Problems with AMD Ryzen 3000 processors have become a constant in recent months. The company has had to go out several times to indicate express corrections to some problems, which on the other hand, were minor. The most important was that some Ryzen did not reach the frequencies set by the company. Now they have had to go out and explain how Windows detects better cores.

Windows 10 was updated to optimize the performance of the Ryzen 3000, something that is appreciated. This update indicated which were the best cores and the preferred processor cores. A distinction that had not been explained so far by anyone, but Robert Hallock explains how it works and indicates that the readings are correct.

No products found.

Ryzen Master tells what the cores are like to Windows 10

Hallock has technically explained that the readings provided by Ryzen Master are correct. Starting this process in AMD itself, who classifies all the processor cores.

Each Ryzen processor has internal firmware to detect the fastest core. Then the Ryzen Master software as well as Windows 10 read the firmware table to see what each core looks like.

The thing is that Windows 10 and the Ryzen Master tool do not use the same criteria to differentiate the cores. This is causing a data discrepancy and user anger. So much so that Hallock indicates that they work to adjust the parameters and that Ryzen Master and Windows 10 have the least possible discrepancy. What is relevant is that to explain it they have given detailed data on the reasons why this happens.

Technical explanation of reading Ryzen Master and Windows 10

1. Inside the processor is the firmware that identifies the performance and classifies the cores. The list is compiled in the factory based on voltage and frequency readings for each core.
2. Windows 10 and Ryzen Master read the performance table stored in the processor firmware where it is determined which core is the fastest.
3. Sorting the kernels with the 'sysfs' command on Linux shows the kernels sorted with a difference of 3%. This does not have to indicate that one kernel is faster than another.
4. CPPC2 (also called 'preferred core' or 'faster core') is based on an abstract scale and without units. This classification tells which is the fastest kernel, but without specifying how much. AMD to avoid this distinguishes them with 3% of difference. It allows the operating system easy detection with no room for rounding.
5. The problem is in the way of operating the operating system, much more complex. Windows chooses and prioritizes the fastest kernel specified in the firmware. Windows also adds another criterion, which is to select the second fastest kernel from the CCX set. Programmers look at this data to avoid all work revolving around a single core, spreading the load. So Windows 10 when it takes the faster core, it chooses a second core to balance the load and meet the optimal performance criteria.
6. Ryzen Master uses the same firmware data choosing the best voltage and frequency curve. The fastest kernel is set in this case from an overclocking perspective. The core therefore does not have to be the one that gives the most performance, but the one that can achieve the most frequency. We see how there are many users who seek to obtain the highest possible frequency in a single core. Therefore it is the fastest kernel, not the one with the best performance.
7. So Windows and Ryzen Master correctly choose the fastest kernel. The problem is that there is a disparity in criteria between the two, even if they are based on the same data table.
8. AMD will modify the Ryzen Master criteria to conform to what Windows indicates. This will allow users to be confident that the selected kernel is the best and not just the fastest.

Source: Reddit