System Performance: Miscellaneous Workloads

Standardized benchmarks such as UL's PCMark 10 and BAPCo's SYSmark take a holistic view of the system and process a wide range of workloads to arrive at a single score. Some systems are required to excel at specific tasks - so it is often helpful to see how a computer performs in specific scenarios such as rendering, transcoding, JavaScript execution (web browsing), etc. This section presents focused benchmark numbers for specific application scenarios.

3D Rendering - CINEBENCH R23

We use CINEBENCH R23 for 3D rendering evaluation. R23 provides two benchmark modes - single threaded and multi-threaded. Evaluation of different PC configurations in both supported modes provided us the following results.

3D Rendering - CINEBENCH R23 - Single Thread

3D Rendering - CINEBENCH R23 - Multiple Threads

As expected, Alder Lake-P retains the single threaded crown, and the Cezanne system delivers multi-threaded performance that is way ahead of any system in the list based on an Intel processor.

Transcoding: Handbrake 1.5.1

Handbrake is one of the most user-friendly open source transcoding front-ends in the market. It allows users to opt for either software-based higher quality processing or hardware-based fast processing in their transcoding jobs. Our new test suite uses the 'Tears of Steel' 4K AVC video as input and transcodes it with a quality setting of 19 to create a 720p AVC stream and a 1080p HEVC stream.

Transcoding - x264

Transcoding - x265_10bit

Software transcoding loves a lot of high-performance cores, and we can see the 8C/16T configurataion with Zen 3 cores handily besting the 4P / 8E combination in Alder Lake-P.

Transcoding - QuickSync H.264

Transcoding - QuickSync H.265 10bit

The QuickSync AVC / HEVC engine continues to remain the same between Tiger Lake and Alder Lake, except for a minor bump in clock speeds that get reflected in the transcoding rates above.

Transcoding - QuickSync H.264

Transcoding - QuickSync H.265 10bit

On the other hand, the VCE performance remains essentially the same between Renoir and Cezanne for the flagship SKUs.

Archiving: 7-Zip 21.7

The 7-Zip benchmark is carried over from our previous test suite with an update to the latest version of the open source compression / decompression software.

7-Zip Compression Rate

7-Zip Decompression Rate

Multiple cores help get compression and decompression done faster, and that is reflected in the significant lead enjoyed by the 4X4 BOX-5800U in the above benchmarks.

Web Browsing: JetStream, Speedometer, and Principled Technologies WebXPRT4

Web browser-based workloads have emerged as a major component of the typical home and business PC usage scenarios. For headless systems, many applications based on JavaScript are becoming relevant too. In order to evaluate systems for their JavaScript execution efficiency, we are carrying over the browser-focused benchmarks from the WebKit developers used in our notebook reviews. Hosted at BrowserBench, JetStream 2.0 benchmarks JavaScript and WebAssembly performance, while Speedometer measures web application responsiveness.

BrowserBench - Jetstream 2.0

BrowserBench - Speedometer 2.0

From a real-life workload perspective, we also process WebXPRT4 from Principled Technologies. WebXPRT4 benchmarks the performance of some popular JavaScript libraries that are widely used in websites.

Principled Technologies WebXPRT4

Single threaded performance helps in these benchmarks, and the Alder Lake-P system gets a slight edge across all three browser benchmark components.

Application Startup: GIMP 2.10.30

A new addition to our systems test suite is AppTimer - a benchmark that loads up a program and determines how long it takes for it to accept user inputs. We use GIMP 2.10.30 with a 50MB multi-layered xcf file as input. What we test here is the first run as well as the cached run - normally on the first time a user loads the GIMP package from a fresh install, the system has to configure a few dozen files that remain optimized on subsequent opening. For our test we delete those configured optimized files in order to force a fresh load every second time the software is run.

AppTimer: GIMP 2.10.30 Startup

As it turns out, GIMP does optimizations for every CPU thread in the system, which requires that higher thread-count processors take a lot longer to run. So the test runs quick on systems with fewer threads, however fast cores are also needed. The Alder Lake-P system gets the overall edge in this case.

System Performance: UL and BAPCo Benchmarks GPU Performance
POST A COMMENT

25 Comments

View All Comments

  • yannigr2 - Friday, August 5, 2022 - link

    Nice review thanks.
    Considering Intel's optimizations for 3D benchmarks, 1-2 games are a necessity for closer to real life results.
    Reply
  • Dante Verizon - Friday, August 5, 2022 - link

    Yes, some games and real world benchmarks... Reply
  • PeachNCream - Saturday, August 6, 2022 - link

    Probably costs too much in terms of time and money to use real world applications. :( Reply
  • xol - Friday, August 5, 2022 - link

    the color legend on the web/javascript seems to be messed up/makes no sense

    ..

    also I think to call the Xe gpu in the intel box as "new" is not quite right - it's not the new architecture (ie Arc) or even close, just shares the branding - it's clearly from the same series that gave us tehe HD 400 way back, and the HD 770 (such as found in the i5-12500) - the difference is that this box has 3x the EU (@96) compared to the i5-12500 (32 EU)... hence the good performance.
    Reply
  • xol - Friday, August 5, 2022 - link

    postscript I'm just gonna say that using the old HD graphics architecture is no bad thing .. at least the drivers will work ! (jokes on Arc for now_ Reply
  • abufrejoval - Friday, August 5, 2022 - link

    Whatever the issues with ARC drivers might be, the iGPU drivers for Linux work quite well also with the newer Xe based variants. The worst I had to do was to force the i915 drivers to accept the unknown Xe PCI device ID via a boot parameter for the kernel.

    No issues on Windows 10/11 either, while there could be trouble with AMD GPUs on Windows server editions because AMD likes to save money on driver signatures there. I used to run Windows server on earlier APUs (Richmond/Kaveri) and had to fiddle hard to get them working anyway.
    Reply
  • abufrejoval - Friday, August 5, 2022 - link

    I'm afraid there is no Thunderbolt in the Intel variant either..., I checked all references I could find.

    And that's really too bad, because for this to work as a µ-server I'd use the TB connector to attach 10GBase-T Ethernet based on AQC107 e.g. as sold by Sabrent. The main advantage is really solid Linux support since years, much better than for the various 2.5 variants from RealTek and Intel.

    The AQC113 chip is out there (hopefully fully AQC107 driver compatible), please ASRock add it to the base board on both devices and I shall buy one of each at least!

    2.5Gbit/s is a long overdue improvement over Gbit, but no longer adequate either. And while it will a few Watts when used at max speed (I think about 3 with Green Ethernet), these boxes aren't running on batteries.
    Reply
  • abufrejoval - Friday, August 5, 2022 - link

    Ok, now finshed reading the review ;-)

    I guess the 2nd set of PCIe x4 would become allocated to Thunderbolt on the Intel variant, if that's really working. AFAIK at least some re-timer chips or similar are both required and in short supply, which is why I'd want working proof.

    AFAIK the AQC113 can do 10GBase-T out of PCIe 4.0 x1 or PCIe 3.0 x2 (or even PCIe 2.0 x4, like the AQC107). So just by dropping the 2nd Gbit, they'd gain all the resources required on Cezanne.

    Yes, a couple of millimeters in height and a Noctua cooler would make all the difference in conjunction with open power and fan settings in the BIOS.

    Since these mainboard are so similar, perhaps somebody (even ASRock) could come up with an alternate chassis?

    I have zero software issues with various kinds of Linux on my 5800U based notebook, while I'm pretty sure all that P/E drama isn't yet sorted out in enterprise Linux. I'm pretty sure that E/P won't reduce the energy footprint on such a NUC in my operations, nor provide better performance under load.

    But with systems so closely matched, at least now I could find out.
    Reply
  • ganeshts - Friday, August 5, 2022 - link

    Yes, the Type-C port close to the Type-A one is indeed Thunderbolt 4. I tested out by connecting the Plugable Thunderbolt 4 Hub to it, and then connecting a Thunderbolt 3-only SSD to one of the downstream ports. I made sure that the TB3 SSD delivered PCIe performance with a quick CrystalDiskMark workload.

    ASRock also mentions it in their block diagram..
    Reply
  • AdrianBc - Saturday, August 6, 2022 - link

    Regarding the height of the case, there is no need to increase it. Intel has proved, already many years ago, that it is possible to cool without problems even a CPU with PL1 = 45 W, without increasing the volume over the classic high variant of NUC, i.e. with a volume under 0.7 L (as proven by the Skull Canyon NUC, which unfortunately did not have any successor, because Intel replaced it with much larger "NUC" computers intended for gaming).

    For better cooling in the same case, it would be enough to remove the possibility of installing a 2.5" HDD/SDD, while keeping the same size for the case.

    The freed height could be used for a taller cooler, with enough capacity to cool even a 45 W CPU, not only a 28 W CPU.
    Reply

Log in

Don't have an account? Sign up now