Posted on by Jean-Luc Aufranc (CNXSoft) - 13 Comments on Maxtang MTN-FP750 business mini PC review – Part 3: Ubuntu 22.04

Maxtang MTN-FP750 business mini PC review – Part 3: Ubuntu 22.04

We’ve already checked out the hardware of the Maxtang MTN-FP750 mini PC in the first part of the review with an unboxing and a teardown, and tested the AMD Ryzen 7 7735HS Mini PC with 32GB DDR5 and a 512GB SSD in Windows 11 Pro in the second part. In the third part of the review, we will report our experience with the Ubuntu 22.04 Linux operating system. This includes features testing, several benchmarks, storage and networking performance testing, a stress test to check thermal performance, as well as fan noise and power consumption measurements.

Maxtang MTN-FP750 review Ubuntu 22.04

Ubuntu 22.04 installation

Since the plan is to install Ubuntu 22.04 alongside Windows 11, so we will shrink the Windows 11 partition by about half first, before inserting the Ubuntu 22.04.3 USB drive to install the Linux distribution.

Disk partition Windows / Ubuntu dual boot

The installation went smoothly with no issues to report on.

Ubuntu 22.04 System information

Going to Settings->About confirms we have Ubuntu 22.04.3 64-bit running on the Maxtang MTN-FP750 mini PC with 32GB RAM and an AMD Ryzen 7 7735HS CPU with AMD Radeon Graphics, and a total of 638.8GB storage (512GB M.2 SSD included in the system, plus our own 2.5-inch 128GB SATA SSD)

Maxtang MTN-FP750 Ubuntu 22.04 system info

We can get a few more details from the command line:


Inxi utility provides a few more details about the system as well:


It shows an AMD Ryzen 7 7735HS processor with 16 threads clocked up to 4829 MHz,  a Foresee SU04Ge M.2 NVMe SSD, and a 118GB CJ225128TC SATA SSD, as well as an RTL8125 2.5GbE controller. The CPU temperature is clearly wrong, but the GPU temperature is shown to be 47°C.

Ubuntu 22.04 benchmarks with Maxtang MTN-FP750 mini PC

We’ll start the Ubuntu 22.04 benchmarks for the MTN-FP750 mini PC with Thomas Kaiser’s sbc-bench.sh script:


The maximum CPU temperature was 82.5°C, and no CPU throttling occurred. The 7-zip benchmarks score actually increased between runs starting at 54,174 MIPS, then 54,746 points, andfinally 55,305 with an average of 54,740 points.

Let’s check out the power limits with Ryzenadj:


All main power limits are the same:

  • Sustained Power Limit (STAPM LIMIT) – 35 Watts
  • Actual Power Limit (PPT LIMIT FAST) – 35 watts
  • Average Power Limit (PPT LIMIT SLOW) – 35 watts

Now let’s run CPU single and multi-core benchmarks with Geekbench 6.2.2.

Maxtang MTN-FP750 Geekbench Ubuntu 22.04

The single-core score is 2131 points, and the multi-core one is 9,665 points. You’ll find the full details on Geekbench website.

Next up is GPU performance with Unigine Heaven Benchmark 4.0 where the Maxtang MTN-FP750 mini PC achieves 67.8 fps on average and a score of 1,707 points at 1920×1080 resolution.

Unigine Heaven Benchmark 4.0 Linux AMD Ryzen 7 7735HS

We’ll now test YouTube videos streaming at both 4K and 8K resolution in Firefox.

Maxtang MTN-FP750 Youtube 4k p30 Ubuntu Linux

We played a 4Kp30 video for about 5 minutes and it was smooth with no frames dropped at all.

Maxtang MTN-FP750 Youtube 4k p60 Ubuntu Linux

Switching to 4Kp60 was fine too with only 41 frames dropped out of 37,737 when playing the video for a little over 10 minutes.

Maxtang MTN-FP750 Youtube 8k p30

An 8K YouTube video at 30 FPS played great with no frame dropped at all.

Maxtang MTN-FP750 Youtube 8k p60

8Kp60 was great too with 18 frames dropped out of 36,972 while playing the video for a little over 10 minutes.

4K video playback performance on Ubuntu 22.04 (Firefox) and Windows 11 (Chrome) is similar, but 8K 60 FPS was somehow much better on Ubuntu 22.04 since most frames were dropped at 8Kp60 on  Windows 11.

Speedometer 2.0 was used in Firefox to evaluate web browsing performance.

Speedometer 2.0 Maxtang MTN-FP750 Ubuntu

The score was 247 runs per minute with some variations between iterations with the score ranging from 233.4 to 250.3.

Speedometer 2.0 detailed results in Firefox

Comparing Maxtang MTN-FP750 performance in Ubuntu 22.04 with other mini PCs

Let’s now compare the Maxtang MTN-FP750 against some of its peers to get a better understanding of the performance in Ubuntu 22.04. We’ll use GEEKOM A5 (AMD Ryzen 7 5800H), GEEKOM Mini IT11 (11th Gen Core i7-11390H), GEEKOM AS 6 (AMD Ryzen 9 6900HX), and GEEKOM Mini IT13 as reference points.

Before checking out the benchmark results, we’ll have a look at the key features and specifications of each system.

Maxtang MTN-FP750GEEKOM A5GEEKOM Mini IT11GEEKOM AS 6GEEKOM Mini IT13
SoCAMD Ryzen 7 7735HSAMD Ryzen 7 5800HIntel Core i7-11390HAMD Ryzen 9 6900HXIntel Core i9-13900H
CPU8-core/16-thread up to 4.75 GHz 8-core/16-thread processor up to 4.4 GHz4-core/8-thread up to 5.0 GHz8-core/16-thread up to 4.9 GHz14-core/20-thread up to 5.4 GHz
GPUAMD Radeon 680M GraphicsAMD Radeon Vega 8 Graphics96 EU Intel Iris Xe Graphics up to 1.4 GHzAMD Radeon Graphics 680M96 EU Intel Iris Xe Graphics up to 1.5 GHz
Memory32GB DDR5-480032GB DDR4-320032GB DDR432GB DDR5-480032GB DDR4-3200
Storage512GB NVMe SSD512GB NVMe SSD1TB NVMe SSD1TB NVMe SSD2TB NVMe SSD
Default OSWindows 11 ProWindows 11 ProWindows 11 ProWindows 11 ProWindows 11 Pro

Here are the benchmark results for each mini PC in Ubuntu 22.04.

Maxtang MTN-FP750GEEKOM A5GEEKOM Mini IT11GEEKOM AS 6GEEKOM Mini IT13
sbc-bench.sh
- memcpy19,252.7 MB/s18,717.0 MB/s19,734 MB/s19,131.7 MB/s24,014.4 (P-core)
- memset18,055.7 MB/s43,837.0 MB/s45,636.7 MB/s16,781.4 MB/s26,647.9 (P-Core)
- 7-zip (average)54,74053,61025,16754,59256,540
- 7-zip (top result)55,30554,85025,40156,25160,981
- OpenSSL AES-256 16K1,297,694.72k1,202,869.59k1,707,917.31k1,249,203.54k1,844,401.49k (P-Core)
Geekbench 6 Single2,1312,0021,9771,9922,745
Geekbench 6 Multi9,6659,3475,7299,53511,974
Unigine Heaven score17078901,0791,5531,333
Speedometer (Firefox)247218-202273

Overall performance is pretty good, and quite similar to the GEEKOM AS 6, although the Maxtang MTN-FP750 has the best 3D graphics score (in Unigine Benchmark 4.0) of the mini PCs above. It’s not too far from the Intel Core i9-powered Mini IT13 either, and offers a better performance/price ratio albeit some of that is due to the lower capacity 512GB SSD.

Storage testing and USB ports

We tested the performance of the preinstalled 512GB NVMe SSD with iozone3:


The performance is acceptable with a sequential read speed of 2.7 GB/s and a sequential write speed of 2.0 GB/s. For reference, the same SSD achieved 3.4 GB/s and 2.5 GB/s read/write speeds in Windows 11 using CrystalDiskMark.

We then connected an ORICO M234C3-U4 “USB4” M.2 NVMe SSD enclosure to double-check each of the USB port’s speed using lsusb and iozone3 with an EXT-4 partition. For example, here are the results for the front left USB 3.2 port of the Maxtang MTN-FP750:


Here are the results for each of the five ports (from left to right):

  • Front panel
    • USB-A #1 – USB 3.2 – USB 3.1 SuperSpeedPlus (10 Gbps) – 830 MB/s
    • USB-A #2 – USB 3.2 – USB 3.1 SuperSpeedPlus (10 Gbps) – 833 MB/s
    • USB-C  –  USB 3.2 – USB 3.1 SuperSpeedPlus (10 Gbps)  – 854 MB/s
  • Rear panel
    • USB-A #1 (top) – USB 2.0 – USB 2.0 Hight-Speed (480 Mbps) – 42 MB/s
    • USB-A #2 (bottom) – USB 3.2 – USB 3.1 SuperSpeedPlus (10 Gbps)  – 856 MB/s

We also tested the performance of the SATA interface with the 128GB SSD (CHUANG JIU CJ225128TC) we installed in the mini PC:


That would be 234 MB/s read speed and 126 MB/s write speed, so the SATA port works as expected.

2.5GbE networking

The Maxtang MTN-FP750 does not ship with a wireless module (although it could be fitted with one), so we only tested the 2.5GbE port with iperf3 utility and UP Xtreme i11 mini PC on the other end

  • Download:

  • Upload:

  • Full-duplex (bidirectional):


Everything works as it should, so no issue here.

Stress test and CPU temperature testing

We ran a stress test on the 16 threads of the AMD Ryzen 7 7735HS CPU to evaluate thermal performance by monitoring the CPU temperature with psensor and the CPU frequency with the sbc-bench.sh script.

AMD Ryzen 7 7735HS temperature in Linux (stress test)

The CPU did not throttle during our test with a maximum temperature of around 86°C in a room at about 28°C. The CPU frequency stabilized at around 3,400 MHz.

Fan noise

The main downside of the Maxtang MT-FP750 mini PC is the fan noise. It’s already noisy at idle, but under load, the fan can be heard 10 meters away. It’s a great feature if you’re afraid your staff may fall asleep during work… OK. I’m exaggerating a bit, but it’s the noisiest mini PC we’ve reviewed so far.

We measure the fan noise with a sound level meter placed at around 5 centimeters from the top of the enclosure:

  • Idle and web browsing – 40.1 to 56.5 dBA
  • Stress test on all 16 threads – 56.9 – 58.5 dBA

For reference, the meter measures 38 – 39 dBA in a quiet room.

Maxtang MTN-FP750 power consumption in Ubuntu 22.04

We measured the power consumption with a wall power meter:

  • Power off – 0.7 Watt
  • Idle – 9.0 – 9.3 Watts
  • Video playback – 59 – 62.2 Watts (Youtube 8K 60fps in Firefox)
  • CPU stress test (stress -c 16) – 57.2 – 58.5 Watts

During the measurements, the mini PC was connected to Ethernet (through a WiFi 6 router), one RF dongle for a mouse and keyboard combo, and a VGA monitor through an HDMI to VGA adapter.

Conclusion

Maxtang MTN-FP750 mini PC worked quite well in Ubuntu 22.04 with no major issues, good performance while browsing, 4Kp60 and 8Kp60 YouTube videos played smoothly, higher 3D graphics performance than most other mini PCs, and 2.5GbE networking working to expectation. It also provides some expansion capabilities an M.2 socket for a wireless module (sadly not included in our sample), five USB ports, and support for up to three independent displays. It performed better than in Windows 11 where our system struggled with 8Kp60 videos on YouTube.

A small downside is the power limit set to a conservative 35W (PL1/PL2) meaning the MTN-FP750 will not fully leverage the performance of the AMD Ryzen 7 7735HS octa-core processor by default, but this would be changed by the user if needed. The main downside is that the fan is quite louder than on other mini PCs we’ve reviewed. I can hear it even about ten meters away in another room and there does not even need to be a heavy load. Having said that the cooling solution is working fine and we did not notice CPU throttling during our tests, which may be important for a machine that’s supposed to run all day or even 24/7 as a digital signage or retail kiosk. In those settings (e.g. busy airports or shopping malls), the fan noise may not even matter at all.

We’d like to thank Maxtang for sending the MTN-FP750 mini PC with AMD Ryzen 7 7735HS CPU (NUC-7735HS-A16 model) with 32 GB of DDR5 RAM, and a 512GB M.2 SSD. The mini PC can be purchased for $387.03 and up on Aliexpress, as well as on the company’s online store where the model with 32GB RAM and a 512GB SSD as reviewed here goes for $483, but you can also select a barebone model for $400, and they offer a discount for orders of 4 pieces.

CNXSoft: This article is a translation – with some additional insights – of the review on CNX Software Thailand by Suthinee Kerdkaew.

Share this:
FacebookTwitterHacker NewsSlashdotRedditLinkedInPinterestFlipboardMeWeLineEmailShare

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

ROCK 5 ITX RK3588 mini-ITX motherboard

13 Replies to “Maxtang MTN-FP750 business mini PC review – Part 3: Ubuntu 22.04”

  1. Thanks for the great review, as per your usual.

    I detect an inconsistency, tho: you first state that it “does not ship with a wireless module (although it could be fitted with one), so we only tested the 2.5GbE port” but later you wrote “the mini PC was connected to WiFi 6”, so what gives?

    As an additional note, I checked and AMD says the 7735H does support ECC RAM, needing only “platform support” (which I think means the board having the traces for the additional pins”), but I can see no mention of ECC support anywhere from Maxtang. This is really sad… :-/

    Reply

    1. Sorry about that. The mini PC was connected over Ethernet to a WiFi 6 router in repeater mode during the power measurement tests.

      With regard to ECC memory, the mini PC shipped with off-the-shelf Crucial 16 GB non-ECC DDR5-5600 memory sticks, and it’s not clear whether ECC is supported. One would assume if it was Maxtang would have indicated it.

      Reply

      1. No prob, and thanks for the clarification re: WiFi6.

        Re: ECC, it would be really great if it was supported… as you have contact with Maxtang, perhaps you could ask and let us know?

        Reply

        1. I’ve just asked them about ECC. We’ll probably need to wait for a while before getting an answer due to CNY 2024 holidays.

          Reply

          1. > I’ve just asked them about ECC

            Thanks, much appreciated! I will be sure to watch for your further posts here, in this comments section — or is there somewhere else I should monitor?

          1. Thanks for asking them and for relaying their answer. It’s really disappointing Maxtang don’t go the ‘extra inch’ (much less than the proverbial “mile”, since the processor already does, and plenty of ‘conventional’ mobo manufacturers provide it in the same situation, so it can’t be too hard nor expensive) to support ECC, tho… it woulf have made the FP750 perfect for my needs and it would have meant a guaranteed buy for me (and I suspect, for many others).

          2. > woulf have made the FP750 perfect for my needs

            Maybe adjusting the ‘needs’ is also an option? For example an awful lot of people still believe they need ECC RAM when they want to use ZFS.

            Which is just an urban myth brought up by some forum guy 14 years ago.

          3. > need ECC RAM when they want to use ZFS.

            Of course it depends on your definition of the word “need”.

            > Which is just an urban myth brought up by some forum guy 14 years ago.

            As someone who has personally lost data due to memory errors on non-ECC RAM, I can assure you it’s no “myth”: it does happen and more frequently than most people suspect (but without ECC RAM and without checking their data all the time like with ZFS, they end up never knowing).

            To sum it up, memory errors do happen, and when they do, ECC RAM makes the difference between losing data (often silently) and losing no data and having the errors corrected and logged so you can locate the bad RAM module and replace it. And got my definition of “need”, yes: I do need it as I value my data.

            But thanks for your opinion anyway.

          4. > As someone who has personally lost data due to memory errors on non-ECC RAM

            Well, that’s the urban myth part since with flipped bits you get data corruption but not data loss (and if you ‘value your data’ IMO you need a strategy based on data integrity like ZFS/btrfs and multiple copies on physically separated locations).

            People lost entire ZFS pools (and btrfs filesystems) with broken ‘write barrier semantics’ due to drives or HBAs lying about properly ‘syncing to disk’ (the OS wants data to be really written at the physical layer to ensure everything is in correct order but drives and /or HBA only fake this).

            From this perspective the invention of the ‘ZFS needs ECC RAM’ myth was useful since users believing into this ended up with buying server grade stuff as such no crappy HBAs and no shitty (ATA) drives.

            As for ‘memory errors do happen’ I second this since each and every server (ofc with ECC RAM since as a business you act kinda stupid trying to save a few bucks here) will be thrown immediatey into our monitoring and even machines that survived 72 hours burn-in test w/o any memory corruption show single bit errors later from time to time.

            BTW: one of our ZFS filers shows ‘Hardware Corrupted: 0.00001% – 32.00 kB of 251.05 GB RAM’ for over half a year now. The monthly scrubs though show no errors at all most probably since the kernel is mapping that memory area out once the 1st occurence in EDAC logs occurs.

          5. > with flipped bits you get data corruption but not data loss

            That’s at best a distinction without a difference. And with ZFS (which was my case) you lose the entire block where these flipped bits were, as ZFS will not return any data for a block with a wrong checksum and just return an error condition instead. And that’s just as well, because if the data is compressed (as per ZFS default) then the decompression would fail anyway.

            And I repeat, if not for your benefit, then for anyone else’s who’s reading this: it is *not* a “myth”. Memory errors do happen and without ECC, they *will* cause loss. I’ve seen it happen and it’s 100% a thing.

            I’m tuning out of this discussion now — not interested in debating this any further. You have a good one, and I hope your beliefs re: ECC RAM being a “myth” don’t ever cause you to lose data, or worse, cause someone else to lose theirs.

          6. > as ZFS will not return any data for a block with a wrong checksum and just return an error condition instead

            Which is highly unlikely if the flipped bit occured due to a corrupted memory cell since then data and checksum would either both be correct or ZFS would’ve already crashed/complained while writing.

            With ZFS as with all other FS all you get in such a situation is ‘standard’ data corruption. So whatever caused your data ‘loss’ it most probably had nothing to do with (non-ECC) RAM.

            > ECC RAM being a “myth”

            Again: I was solely talking about the myth ECC RAM being a requirement for ZFS which is still BS: https://jrs-s.net/2015/02/03/will-zfs-and-non-ecc-ram-kill-your-data/

            Asides that, yeah bit flips can happen (even with ‘standard’ ECC RAM which can correct only single bit flips but not multiple) and the consequences of this depend on the data affected. 🙂

Leave a Reply

Your email address will not be published. Required fields are marked *

Boardcon Rockchip and Allwinner SoM and SBC products
Boardcon Rockchip and Allwinner SoM and SBC products