Systemd on (some) older hardware

[paul1149]

I've been putting together an older Toshiba with an AMD e-300 apu, a pathetic apu from the day it was manufactured. Going to Dreamer's LXQT respin helped a lot, and I thought I finally had a decent system working.

Then this morning I fired it up for a final check, to find that performance was on the edge of not acceptable. I pondered why this was, and then finally I remember a change I had made late the night before. I went into MX Boot Options and specced Systemd.

So I rebooted, this time with sysvinit, and I was back to my working system again. I would guess the performance was 35-50% better on average.

I did a little research, and it seems this is a problem on some, not all, older hardware, Something to be aware of.

[j2mcgreg]

It's generally known that there is at least a slight performance hit when using MX with SystemD on any system regardless of age or complexity.

[Nokkaelaein]

It's generally known that there is at least a slight performance hit when using MX with SystemD on any system regardless of age or complexity.

Hmm. In my work, I'm extremely performance conscious, and I didn't notice anything when switching to mainlining systemd some months ago. Granted, I'm not exactly running MX, but a specialized version of it with some quite different innards, and running threaded interrupt handlers, and so on. Nevertheless, this is intriguing and makes me pretty curious if there's some benchmark data available? "Regardless of age or complexity", hmm.. What areas of operation suffer a performance hit in stock MX when running systemd, and how? As in, what is the context of the "performance hit" in this case? (General processing / FPU calculation resources? General system responsiveness / latency? I/O, startup/shutdown, some specific tools like some particular known applications or the MX tools suite or other? Can the cause be pinpointed and has it been measured?)

[paul1149]

> What areas of operation suffer a performance hit in stock MX when running systemd, and how?

In my case it was all manner of sluggishness, from playing a video from YT at constant 95%+ CPU to clicking on a program close button and waiting 20+ seconds for it to disappear. The machine was borderline unusable.

[Nokkaelaein]

In my case it was all manner of sluggishness, from playing a video from YT at constant 95%+ CPU to clicking on a program close button and waiting 20+ seconds for it to disappear. The machine was borderline unusable.

Yeah, that sounds like an extreme difference. Would be interesting to know the cause. (My very broad guess would be, the chip is so old, with only two concurrent threads to run stuff on, it can't cope with some particularly heavy hitting service parallelism. With old 1-2 thread systems like that, the way background tasks are scheduled, and what kind of background/maintenance jobs the system is set to launch and when, becomes a big deal.)

[retroD0d0]

I don't really notice a huge impact running systemd it has to be said on my 2009 potato, it even boots a second faster;) Also, how much/type RAM do you have? Systemd is more memory hungry, you're system may be overly depending on a swap? (cough QSI would help here) What secondary storage is it HDD or eMMC? I have found I/O is sometimes the bottleneck rather than the CPU.

[paul1149]

4GB, 7200rpm hard drive. Swappiness untouched.

Here you go.

Code: Select all

System:
  Kernel: 6.1.0-33-amd64 [6.1.133-1] arch: x86_64 bits: 64 compiler: gcc v: 12.2.0
    parameters: BOOT_IMAGE=/boot/vmlinuz-6.1.0-33-amd64 root=UUID=<filter> ro quiet splash
    resume=UUID=<filter> resume_offset=780288
  Desktop: Xfce v: 4.20.0 tk: Gtk v: 3.24.38 info: xfce4-panel wm: xfwm v: 4.20.0 vt: 7
    dm: LightDM v: 1.32.0 Distro: MX-23.6_x64 Libretto April 13  2025 base: Debian GNU/Linux 12
    (bookworm)
Machine:
  Type: Laptop System: TOSHIBA product: Satellite C855D v: PSCBQU-00600J
    serial: <superuser required>
  Mobo: TOSHIBA model: Portable PC v: MP serial: <superuser required> UEFI: Insyde v: 6.00
    date: 08/21/2012
Battery:
  ID-1: BAT0 charge: 44.1 Wh (100.0%) condition: 44.1/47.5 Wh (92.9%) volts: 10.8 min: 10.8
    model: PA5024U-1BRS type: Li-ion serial: <filter> status: full
CPU:
  Info: model: AMD E-300 APU with Radeon HD Graphics bits: 64 type: MCP arch: Bobcat level: v1
    built: 2011-13 process: GF 40nm family: 0x14 (20) model-id: 2 stepping: 0 microcode: 0x5000119
  Topology: cpus: 1x cores: 2 smt: <unsupported> cache: L1: 128 KiB desc: d-2x32 KiB; i-2x32 KiB
    L2: 1024 KiB desc: 2x512 KiB
  Speed (MHz): avg: 780 high: 781 min/max: 780/1300 boost: disabled scaling: driver: acpi-cpufreq
    governor: ondemand cores: 1: 781 2: 779 bogomips: 5191
  Flags: ht lm nx pae sse sse2 sse3 sse4a ssse3 svm
  Vulnerabilities:
  Type: gather_data_sampling status: Not affected
  Type: itlb_multihit status: Not affected
  Type: l1tf status: Not affected
  Type: mds status: Not affected
  Type: meltdown status: Not affected
  Type: mmio_stale_data status: Not affected
  Type: reg_file_data_sampling status: Not affected
  Type: retbleed status: Not affected
  Type: spec_rstack_overflow status: Not affected
  Type: spec_store_bypass status: Vulnerable
  Type: spectre_v1 mitigation: usercopy/swapgs barriers and __user pointer sanitization
  Type: spectre_v2 mitigation: Retpolines; STIBP: disabled; RSB filling; PBRSB-eIBRS: Not
    affected; BHI: Not affected
  Type: srbds status: Not affected
  Type: tsx_async_abort status: Not affected
Graphics:
  Device-1: AMD Wrestler [Radeon HD 6310] vendor: Toshiba driver: radeon v: kernel
    alternate: amdgpu arch: TeraScale-2 code: Evergreen process: TSMC 32-40nm built: 2009-15 ports:
    active: LVDS-1 empty: HDMI-A-1,VGA-1 bus-ID: 00:01.0 chip-ID: 1002:9802 class-ID: 0300
  Device-2: Importek TOSHIBA Web Camera - HD type: USB driver: uvcvideo bus-ID: 2-4:2
    chip-ID: 10f1:1a43 class-ID: 0e02
  Display: x11 server: X.Org v: 1.21.1.7 compositor: xfwm v: 4.20.0 driver: X: loaded: radeon
    unloaded: fbdev,modesetting,vesa dri: r600 gpu: radeon display-ID: :0.0 screens: 1
  Screen-1: 0 s-res: 1366x768 s-dpi: 96 s-size: 362x204mm (14.25x8.03") s-diag: 416mm (16.36")
  Monitor-1: LVDS-1 mapped: LVDS model: LG Display 0x033a built: 2012 res: 1366x768 hz: 60
    dpi: 101 gamma: 1.2 size: 344x194mm (13.54x7.64") diag: 395mm (15.5") ratio: 16:9 modes:
    max: 1366x768 min: 640x480
  API: OpenGL v: 4.5 Mesa 22.3.6 renderer: AMD PALM (DRM 2.50.0 / 6.1.0-33-amd64 LLVM 15.0.6)
    direct-render: Yes
Audio:
  Device-1: AMD Wrestler HDMI Audio vendor: Toshiba driver: snd_hda_intel v: kernel bus-ID: 00:01.1
    chip-ID: 1002:1314 class-ID: 0403
  Device-2: AMD FCH Azalia vendor: Toshiba driver: snd_hda_intel v: kernel bus-ID: 00:14.2
    chip-ID: 1022:780d class-ID: 0403
  API: ALSA v: k6.1.0-33-amd64 status: kernel-api tools: alsamixer,amixer
  Server-1: PipeWire v: 1.0.0 status: active with: 1: pipewire-pulse status: active
    2: wireplumber status: active 3: pipewire-alsa type: plugin 4: pw-jack type: plugin
    tools: pactl,pw-cat,pw-cli,wpctl
Network:
  Device-1: Realtek RTL8188CE 802.11b/g/n WiFi Adapter driver: rtl8192ce v: kernel modules: wl
    pcie: gen: 1 speed: 2.5 GT/s lanes: 1 port: 3000 bus-ID: 02:00.0 chip-ID: 10ec:8176
    class-ID: 0280
  IF: wlan0 state: up mac: <filter>
  Device-2: Realtek RTL810xE PCI Express Fast Ethernet vendor: Toshiba driver: r8169 v: kernel
    pcie: gen: 1 speed: 2.5 GT/s lanes: 1 port: 2000 bus-ID: 06:00.0 chip-ID: 10ec:8136
    class-ID: 0200
  IF: eth0 state: down mac: <filter>
Drives:
  Local Storage: total: 232.89 GiB used: 13.63 GiB (5.9%)
  SMART Message: Unable to run smartctl. Root privileges required.
  ID-1: /dev/sda maj-min: 8:0 vendor: Hitachi model: HTS725025A9A364 size: 232.89 GiB block-size:
    physical: 512 B logical: 512 B speed: 3.0 Gb/s type: HDD rpm: 7200 serial: <filter> rev: C72E
    scheme: GPT
Partition:
  ID-1: / raw-size: 232.63 GiB size: 227.92 GiB (97.97%) used: 13.63 GiB (6.0%) fs: ext4
    dev: /dev/sda2 maj-min: 8:2
  ID-2: /boot/efi raw-size: 256 MiB size: 252 MiB (98.46%) used: 274 KiB (0.1%) fs: vfat
    dev: /dev/sda1 maj-min: 8:1
Swap:
  Kernel: swappiness: 15 (default 60) cache-pressure: 100 (default)
  ID-1: swap-1 type: file size: 5.42 GiB used: 0 KiB (0.0%) priority: -2 file: /swap/swap
Sensors:
  System Temperatures: cpu: 52.1 C mobo: N/A gpu: radeon temp: 52.0 C
  Fan Speeds (RPM): N/A
Repos:
  Packages: pm: dpkg pkgs: 2139 libs: 1066 tools: apt,apt-get,aptitude,nala,synaptic pm: rpm
    pkgs: 0 pm: flatpak pkgs: 0
  No active apt repos in: /etc/apt/sources.list
  Active apt repos in: /etc/apt/sources.list.d/brave-browser-release.list
    1: deb [arch=amd64 signed-by=/usr/share/keyrings/brave-browser-archive-keyring.gpg] https://brave-browser-apt-release.s3.brave.com/ stable main
  Active apt repos in: /etc/apt/sources.list.d/debian-stable-updates.list
    1: deb http://deb.debian.org/debian bookworm-updates main contrib non-free non-free-firmware
  Active apt repos in: /etc/apt/sources.list.d/debian.list
    1: deb http://deb.debian.org/debian bookworm main contrib non-free non-free-firmware
    2: deb http://security.debian.org/debian-security bookworm-security main contrib non-free non-free-firmware
  Active apt repos in: /etc/apt/sources.list.d/mx.list
    1: deb http://mirror.math.princeton.edu/pub/mxlinux/mx/repo/ bookworm main non-free
Info:
  Processes: 180 Uptime: 1h 6m wakeups: 2 Memory: 3.42 GiB used: 1.11 GiB (32.6%) Init: SysVinit
  v: 3.06 runlevel: 5 default: graphical tool: systemctl Compilers: gcc: 12.2.0 alt: 12
  Client: shell wrapper v: 5.2.15-release inxi: 3.3.26
Boot Mode: UEFI

[retroD0d0]

OK, specs seem fine, respectable even, if you're not too snooty. I have a similar system with slightly higher clocks (e-350), It is currently running Windows 10 32bit! (though it is actually 64 bit capable) It's certainly not pretty, quite sluggish, but just about usable, as long as updates/defender isn't running in the background and I set the resolution to 1366*768. It can process fullHD video h.264 too. If it can run Win10 I'm sure it can eat MX. As bad of an early APU as it was, it actually outperformed the Intel Atoms it was designed to compete with.

Have you eliminated the possibility the CPU is throttling due to overheating issues? Thermal paste might need changing after all this time. Have you run SMART tests on your hdd?

I find it's modern web standards that are the biggest roadblock for older hardware. For instance, the newer video codecs that YT uses require a lot of CPU cycles to decompress. In fact, I use FreeTube and select legacy formats/codecs to get around this. Also the load by the sheer number of network connections that are made when accessing even a single website as it accesses resources/ads from all over the web, the other problem I have is with websites with irresponsible JavaScript workers that seem to background run even when doing nothing noticeable. Using script blockers helps here, or setting your browser to aggressively sleep inactive tabs.

I'm curious @paul1149 how does it perform when offline eg. Libreoffice, scrolling through rich pdfs etc?

[paul1149]

The temperature is fine. If it runs at high cpu for a while the keyboard is slightly warm, but not hot. The hard drive at the least passed SST, maybe more, not sure, but that's not a factor anyway, I believe, because it was the same hard drive for both sysd and sysv.

Ironically this unit had W10 home on it, and performed not catastrophically badly. But I was unwilling to go through the interminable 4.5GB windows upgrade. So I went to MX, assuming would be a cakewalk. But MX under SysD actually did worse than W10! I was perplexed until I made this discovery.

Now I'm tempted to go back, at least via usb boot, and see how xfce does without SysD on the machine.

I agree about javascript and web standards. Many tabs on the latest Vivaldi "shimmer", as they apparently never fully load. This very tab, here, right now, was fluctuating between 6 and 16% cpu on my Ryzen system. I was tempted to put NoScript on the laptop, but the guy I'm building it for probably would be lost with it. I chose firefox for the system because, surprisingly, it did better on resources than Brave or even PaleMoon. FF does hibernate tabs, but gives you no control over it. I will take a look at freetube -thanks.

The problem wasn't related to being online. All windows performed extremely poorly.

[retroD0d0]

But MX under SysD actually did worse than W10! I was perplexed until I made this discovery.

I was hoping to migrate my e-350 to MX too when W10 is EoL. Bit worried now . If you don't mind a bit of fan noise (and warm fingers), have you thought about changing the CPU governor? Those bobcat APU's are dog-slow, I won't lie, but the default 'OnDemand' governor in MX doesn't flatter it at all. If you switch to a performance profile you will notice an improvement in responsiveness at least.