Linux and Memory Performance on an Intel NUC 11 Enthusiast Phantom Canyon NUC11PHKi7C

I’ve already looked at Windows performance on the NUC11PHKi7C Enthusiast Phantom Canyon which is Intel’s latest NUC 11 flagship product specifically targeting gamers as it includes an NVIDIA RTX 2060 GPU. Now it is the turn of Linux and like before I will compare performance against Intel’s previous NUC with a discrete GPU: the NUC 9 Extreme Ghost Canyon. I will also briefly revisit Windows performance by looking at the impact of using 3200MHz memory as opposed to the 2400MHz used during the previous review.

Hardware Overview

As a reminder, the NUC11PHKi7C physically consists of a 221 x 142 x 42 mm (8.70 x 5.59 x 1.65 inches) rectangular plastic case. It is an actively cooled mini PC and uses Intel’s 10 nm Core i7-1165G7 Tiger Lake processor which is a quad-core 8-thread 2.80 GHz processor boosting to 4.70 GHz with Intel’s Iris Xe Graphics and the NUC also includes an NVIDIA N18E-G1-B notebook graphics card which is a GeForce RTX 2060 mobile GPU. As a Max-P (max-performance) GPU, its power is limited to 115W as opposed to 160W for the desktop equivalent and consequently has lower clock and memory speeds.

Having purchased the NUC11PHKi7C barebones model I’ve added a 2TB M.2 2280 NVMe drive from addlink (S70) and 64GB (2 x 32GB) DDR4 3200MHz memory from G.SKILL. For the comparison, the NUC 9 Extreme I’ve used my NUC9i7QNX which has a 2TB M.2 2280 NVMe drive from ADATA (XPG 8200 Pro), 64GB (2 x 32GB) of Team Group’s Team Elite DDR4 3200MHz memory, and an EVGA GeForce RTX 2060 KO ULTRA GAMING GPU.

Linux Installation Issues

One of the benchmarks requires compiling its code which is written in C with the ‘-march=native’ option. However, the default 9.3.0 version of ‘gcc’ installed in Ubuntu 20.04 fails to recognize the ‘tigerlake’ architecture of this new CPU so the solution is to install and use the later 10.2.0 version of ‘gcc’ as this simple example shows:

Running the game Grand Theft Auto V under Proton on Steam also encountered some issues. The first problem is that after running an initial benchmark the game resets the graphics to some unknown state where the framerate subsequently fluctuates excessively and performance is degraded. For example during an initial run the framerate for the following scene is over 187:

whereas subsequent runs can result in a framerate dipping into the 30s:

The second problem is when connecting to an HDMI monitor via a Thunderbolt port and using PRIME to set the GPU to NVIDIA:

the game refuses to load:

No working solution for either issue has been found so far.

Intel NUC 11 “Enthusiast Phantom Canyon” Linux Performance

As I had already installed Windows I first shrank the Windows partition in half and created a new partition into which I installed Linux using an Ubuntu 20.04.2 ISO as dual boot. After installation and updates, a brief check showed working Wi-Fi, Bluetooth, and Ethernet.

The key hardware information under Ubuntu 20.04.2 is as follows:

First I set performance to ‘maximum’ and fan mode to ‘cool’ on both the NUC 9 Extreme and NUC 11 Enthusiast through the BIOS. I then booted and set the CPU Scaling Governor to ‘performance’ and ran my Linux benchmarks for which the majority of the results are text-based but the graphical ones for the NUC11PHKi7C included:

and also my specific set of Phoronix Test Suite tests:

All of the results for the NUC 11 Enthusiast can be compared to those from the NUC 9 Extreme:

The key issue with the NUC 9 Extreme is the slow read performance for NVMe drives connected via the Compute Element compared to when connected via the baseboard and this is reflected in the results. Fortunately, this isn’t the case for the NUC 11 Extreme with performance aligned with the manufacturer’s specification. Again the impact of using a mobile RTX 2060 can be seen in graphical benchmarks such as Furmark which is 20% lower than when using the desktop version.

Gaming on Intel NUC 11 with Linux

For both NUC, I performed a quick test of three games under Steam in Ubuntu (Counter-Strike: Global Offensive, Grand Theft Auto V, and Shadow Of The Tomb Raider) at 1920×1080 resolution. I used the default settings for CS:GO and GTA V together with each of the graphic presets in SOTTR:

Note that on Ubuntu I relied on Steam’s in-game FPS counter in CS:GO and GTA V’s in-game FPS counter to estimate the average FPS as I did not have tools equivalent to MSI AFterburner/Rivatuner available.

Again the impact of the mobile GPU was evident:

However, the most notable difference between the two devices is actually the fan noise. The fans reach their loudest during the market scene in the SOTTR benchmark and using a hand-held sound meter I measured around 54dBA for the NUC9i7QNX when it was best described as a loud whooshing. In contrast, I measured only around 44 dBA for the NUC11PHKi7C and it sounded like a low hum which was very acceptable. The ambient sound was around 33 dBA.

Windows Revisited

As a result of my 3200MHz Team Group’s Team Elite memory in the NUC11PHKi7C failing to boot I initially had to use some 2400MHz memory from Samsung for the previous Windows performance tests. Now that I have replaced this with some 3200MHz G.SKILL memory I have rerun all the Windows benchmarks. Interestingly, this only saw minor improvements in only a few benchmarks. Specifically, the Passmark score improved by 3% due to the Memory Mark improving by 17%. Productivity in PCMark 10 improved by 7% and finally Geekbench 5 scores for single and multi-core improved by 3% and 5% respectively.

Then I changed the fan mode from ‘balanced’ to ‘cool’ using Intel NUC Software Studio and reran the benchmarks. However, the impact was negligible other than reducing the productivity gain in PCMark 10 from 7% to 5% indicating that the margin of error when running the tests is probably a factor in the initial improvement. For reference across all the benchmarks, the average improvement of upgrading from 32GB of 2400MHz memory to 64GB of 3200MHz memory was 1.6% per result which increased to 1.7% in conjunction with the ‘cool’ fan mode.

The updated Windows benchmark results are:

I also repeated the gaming tests on Windows:

which showed a similar minor improvement from upgrading the memory of between 1 to 5 FPS depending on game and settings.

Final Observations

As a ‘gaming’ mini PC, although it uses a mobile GPU resulting in the graphics performance being lower than the equivalent desktop GPU, it does meet the brief. It is somewhat pricey but so too are graphics cards currently and the NUC11PHKi7C includes one making it a better value proposition right now. Given its physical size both the CPU and GPU performance are very good and the fact that it runs near silent makes this an outstanding mini PC.

Ian W MORRISON (Linuxium)

Ian is interested in mini PCs and helps with reviews of mini PCs running Windows, Ubuntu and other Linux operating systems. You can follow him on Facebook or Twitter.

Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress