Impact of operating system and memory on mini PC gaming performance

This article examines the effects of running a different operating system or more memory on Intel and AMD Mini PCs with similar specs when gaming.

It was inspired by building and testing a nickname ‘Steamdeck’ running Manjaro on an AMD-based mini PC with 16GB of memory, which made me wonder what performance would look like with Windows 11.

The first results were surprising because Windows appeared much slower. As I had previously heard of performance improvements when using 64GB of memory, I replaced the currently installed 16GB memory and immediately saw an improvement in results.

As I had never seen such a dramatic increase in performance on Intel mini-PCs simply by increasing the memory, I decided to explore further by testing gaming performance on similar Intel and AMD mini-PCs in using 16 GB or 64 GB of memory and comparing running Windows with Linux. Given the ‘Steamdeck’ used by Manjaro, I also wanted to test with Ubuntu to see if that made a difference.

Material under test

Recent AMD mini PCs have been notable for including the more powerful Radeon integrated graphics, while Intel mini PC iGPUs are generally much weaker, with the exception of the now dated Intel Iris Plus Graphics 655. Like many recent mini-PCs have been released with processors. with these integrated graphics it was the logical choice for my test intel devices. Limited by what I had available, the following four mini-PCs (Intel: GTi & NGC-5, AMD: GT-R & SER3) were selected for testing because they were the most similar:

Since memory was the key hardware component tested, I chose to reuse the same memory in each device to ensure consistency. Support for different memory speeds and the ability to overclock memory was limited in each device’s BIOS. Intel devices were limited to running memory at a maximum speed of 2400 MHz, but the AMD BIOS allowed a slight memory overclocking to be set to 2666 MHz. Operating in dual channel, I used two keys of Crucial 8 GB DDR4-2666 CL19 (CT8G4SFS6266) and two keys of 32 GB DDR4-3200 CL22 (F4-3200C22D-64GRS):

So, for Intel devices, the memory worked at 2400 MHz:

and for AMD devices it was running at 2666 MHz:

Noting that DDR4-3200 memory operates with CAS latency of 19 when clocked at 2666 MHz:

Software

New installations of each operating system were done on each device and updated to the latest versions, then benchmarking software was installed. Additionally, “RyzenAdj” was installed on AMD devices to configure power limits.

For Windows, Windows 11 Pro Version 21H2 build 22000.348 was used on each mini PC:

and for Ubuntu, Ubuntu 20.04.3 with generic 5.11.0-41 kernel was used:

Then, for Manjaro, Manjaro 21 KDE Plasma was used, but as Manjaro is a continuous version for the first round of tests on Intel NGC-5 and AMD SER3 mini-PCs, Manjaro 21.1.6 with kernel 5.13.19- 2 was used:

and for the second round of tests on Intel GTi and AMD GT-R, Manjaro 21.2rc1 with kernel 5.15.6-2 was used:

I also confirmed that changing the version point and kernels didn’t seem to influence the results by briefly running a few more benchmarks.

Finally, the Steam software from Valve and Heaven from Unigine were installed and used for testing as well as for installing the MSI Afterburner FPS monitoring software with Rivatuner Statistics Server on Windows and MangoHud on Linux.

System configuration

In Windows, the power mode has been set to “High performance” on each device:

and similarly, on Ubuntu and Manjaro, the CPU Scaling Governor was set to “performance”:

On both AMD devices, “RyzenAdj” was used to set the Actual Power Limit (PTT Limit Fast) to 45W, the Average Power Limit (PPT Limit Slow) to 40W, and the Slow PPT Constant Time (SlowPPTTimeConst) at 5 seconds:

for each operating system:

Finally, the “Display” resolution has been set to 1280 × 720 on each device:

Test methodology

Initially, I tested several games on Steam on Windows and Linux including Counter-Strike: Global Offensive (CS: GO), Grand Theft Auto V (GTA V), Horizon Zero Dawn (HZD) and Shadow Of The Tomb Raider ( SOTTR). Although I noticed consistent performance as per the findings below, I discontinued testing of CS: GO and GTA V in deference to using the more consistent game benchmarks of HZD and SOTTR. I also added tests with Heaven using the ‘OpenGL’ API:

because this is both available on Windows and Linux and was also always reproducible. However, I only tabulated the SOTTR and Heaven results as they sufficiently demonstrate the trends seen in all of the results.

Results

Operating system observations

A direct comparison of gaming performance between Windows and Linux cannot be drawn from such limited data and it should also be noted that some games run natively while others use compatibility tools like “Proton”. However, what was interesting was that on Intel mini PCs, SOTTR on Manjaro was much slower than on Ubuntu. This was not the case for AMD devices where performance was similar using the “OOTB” experience. There might just be a simple fix for this, but it highlights a common problem in “gaming” on Linux where it often seems necessary to look for fixes just to get things to work.

Observations of memory

The most obvious impact was that increasing memory from 16GB to 64GB on AMD devices resulted in a noticeable improvement in FPS. The benefits seem to favor Windows more than Linux which, while inferior, still experienced consistent increments. Conversely, there was effectively no consequence of the increased memory on Intel devices, with the few minor differences being within the range of test variance.

Conclusion

Gaming performance may differ between Windows and Linux, so the choice of operating system will likely depend on whether the desired games have “native” versions or are supported by an appropriate compatibility layer.

However, increasing memory appears to improve gaming performance on AMD mini-PCs with noticeable FPS increments, especially on Windows, while no noticeable improvement was seen on Intel mini-PCs. Whether these results justify the additional expense of purchasing more memory is debatable. However, if you have it available, it makes sense to use it.