The latency of the instructions is one of the key elements for the performance of any processor. Depending on where the data to be processed through an instruction is located, its resolution speed is higher or lower and therefore a processor is faster or slower. Today, the systems for managing dozens of software processes on the hardware are managed by the operating system and improper operation can lead to a considerable loss of performance.
This has been the problem that Ryzen cores of different generations have had to face in Windows 11, which has been seen as the L3 cache at least it tripled and with it the performance of these processors was cut in the new Microsoft operating system. Luckily, the patch for Windows 11 that will be released in a few days is already at least in AMD users enrolled in the Insider Program and we have been able to know how they improve the performance of Ryzen CPUs.
The improvement of the AMD Ryzen in Windows 11
Although according to information from Advanced Micro Devices itself, this problem especially affects the performance of CPUs with more than 8 cores and above 65 W of consumption, the reality is that all CPUs in the Ryzen 5000 range share the same CCD Chiplet and therefore also do with the problem.
This first screenshot made under AIDA64 corresponds to a Ryzen 5600X CPU without the Windows 11 patch for AMD CPUs, as you can read the L3 cache has a bandwidth of 106.64 GB / s for reading, 53 GB / s for writing and 63 GB / s for copying or simultaneous reading and writing. All this thanks to a 39 nanosecond exaggerated latency. It must be taken into account that latency and bandwidth are related, since the more delay in the transmission of data by logic, the less of them are transmitted per unit of time, thus affecting the bandwidth.
This capture is instead made with the version of Windows 11 with the patch for AMD’s Ryzen and as can be seen under the same processor, Ryzen 5 5600X, thanks to a decreased latency from 39 to 14.5 nanoseconds you can see how the bandwidth for reading with the L3 cache has increased to the 373.29 GB / s Reading, 378.55 GB / s of writing and in terms of copying to 437.91 GB / s. This is therefore a significant performance increase under the same processor.
Remember that from the moment the L3 cache is inside the CCD these clock speeds occur in the communication of the cores within each chiplet and not between chiplets, where the latency is greater due to the distance.
Fairer comparisons between Intel and AMD CPUs
Before the release of this patch, the artificial handicap for AMD processors made a fair comparison between both CPUs impossible, since the Ryzen worked with a limitation that did not allow them to get their full potential under Windows 11. With the performance gained thanks to By implementing this patch, we will finally be able to get a more concrete idea of how both processors compare under the new version of the world’s most widely used operating system.
Which architecture will win the contest, AMD’s Zen 3 or Intel Core 12? There is little left and now we know that we can compare both architectures fairly under Windows 11. It is expected that this patch will be implemented in little more than a week for any PC that has this OS, including of course those that are not Insider.