Scheduled Maintenance: We are aware of an issue with Google, AOL, and Yahoo services as email providers which are blocking new registrations. We are trying to fix the issue and we have several internal and external support tickets in process to resolve the issue. Please see:
viewtopic.php?t=158230
Linux Kernel, Network, and Services configuration.
dcihon
Posts: 279 Joined: 2012-09-23 16:18
Been thanked: 3 times
#1
Post
by dcihon » 2018-07-12 10:34
What do you think of using these commands to test the performance of a system:
Code: Select all
systemd-analyze
systemd-analyze critical-chain
systemd-analyze blame
Mine looks like this:
Code: Select all
cihonm@cihonm:/media/cdrom0$ systemd-analyze
Startup finished in 5.563s (kernel) + 32.257s (userspace) = 37.820s
graphical.target reached after 32.243s in userspace
Code: Select all
cihonm@cihonm:/media/cdrom0$ systemd-analyze critical-chain
The time after the unit is active or started is printed after the "@" character.
The time the unit takes to start is printed after the "+" character.
graphical.target @32.243s
└─multi-user.target @32.243s
└─exim4.service @28.361s +3.881s
└─network-online.target @28.360s
└─NetworkManager-wait-online.service @17.174s +11.186s
└─NetworkManager.service @12.623s +4.550s
└─dbus.service @12.314s
└─basic.target @12.298s
└─sockets.target @12.298s
└─virtlogd-admin.socket @12.298s
└─sysinit.target @12.225s
└─haveged.service @12.225s
└─apparmor.service @10.349s +1.863s
└─local-fs.target @10.347s
└─home.mount @10.064s +282ms
└─systemd-fsck@dev-disk-by\x2duuid-1ee2b8d8\x2dab62\x2d4a09\x2d870a\x2d0c37e5308d5b.service @9.623s +417ms
└─dev-disk-by\x2duuid-1ee2b8d8\x2dab62\x2d4a09\x2d870a\x2d0c37e5308d5b.device @9.621s
Code: Select all
cihonm@cihonm:/media/cdrom0$ systemd-analyze blame
48.470s apt-daily-upgrade.service
30.178s apt-daily.service
15.911s udisks2.service
11.186s NetworkManager-wait-online.service
8.986s tor@default.service
8.621s certbot.service
7.630s syslog-ng.service
6.797s dev-sda6.device
6.229s loadcpufreq.service
4.899s libvirtd.service
4.550s NetworkManager.service
4.272s lm-sensors.service
4.230s avahi-daemon.service
4.022s networking.service
3.881s exim4.service
3.878s accounts-daemon.service
3.530s vboxdrv.service
3.362s keyboard-setup.service
3.208s snapper-boot.service
3.197s ModemManager.service
3.129s autofs.service
3.001s wpa_supplicant.service
2.925s teamviewerd.service
2.785s systemd-logind.service
2.701s rtkit-daemon.service
2.036s dev-loop0.device
1.968s dev-loop3.device
1.963s vpnagentd.service
1.942s dev-loop2.device
1.941s lightdm.service
1.863s apparmor.service
1.752s upower.service
1.745s dev-loop1.device
1.606s lvm2-monitor.service
1.516s systemd-udevd.service
1.500s colord.service
1.357s snap-core-4206.mount
1.289s snap-core-4407.mount
1.287s snap-acestreamplayer-7.mount
1.259s snap-core-4650.mount
1.048s bluetooth.service
983ms systemd-modules-load.service
874ms ssh.service
707ms binfmt-support.service
685ms ebtables.service
547ms run-rpc_pipefs.mount
536ms systemd-journal-flush.service
521ms alsa-restore.service
502ms systemd-timesyncd.service
495ms cpufrequtils.service
494ms systemd-journald.service
485ms systemd-tmpfiles-setup-dev.service
450ms dev-mqueue.mount
449ms dev-hugepages.mount
438ms systemd-remount-fs.service
420ms packagekit.service
419ms sys-kernel-debug.mount
417ms systemd-fsck@dev-disk-by\x2duuid-1ee2b8d8\x2dab62\x2d4a09\x2d870a\x2d0c37e5308d5b.service
409ms resolvconf.service
408ms systemd-sysusers.service
357ms systemd-sysctl.service
346ms rpcbind.service
330ms proc-sys-fs-binfmt_misc.mount
293ms plymouth-read-write.service
282ms home.mount
256ms systemd-tmpfiles-setup.service
230ms systemd-udev-trigger.service
209ms clamav-daemon.service
204ms pppd-dns.service
202ms dev-disk-by\x2duuid-42e5932b\x2dd210\x2d427f\x2daef0\x2d28aac6826539.swap
201ms tor.service
197ms systemd-user-sessions.service
194ms libvirt-guests.service
170ms systemd-tmpfiles-clean.service
160ms ifupdown-wait-online.service
137ms systemd-update-utmp.service
131ms kmod-static-nodes.service
130ms blk-availability.service
129ms systemd-rfkill.service
98ms rc-local.service
92ms console-setup.service
91ms hddtemp.service
86ms systemd-random-seed.service
84ms nfs-config.service
83ms polkit.service
80ms user@1000.service
60ms vboxweb-service.service
60ms vboxballoonctrl-service.service
56ms vboxautostart-service.service
18ms plymouth-start.service
11ms systemd-update-utmp-runlevel.service
3ms plymouth-quit-wait.service
2ms openvpn.service
2ms sys-fs-fuse-connections.mount
bw123
Posts: 4015 Joined: 2011-05-09 06:02
Has thanked: 1 time
Been thanked: 28 times
#2
Post
by bw123 » 2018-07-12 17:20
That's not a good test of system performance or speed, that's only bootup time?
resigned by AI ChatGPT
Hallvor
Global Moderator
Posts: 2029 Joined: 2009-04-16 18:35
Location: Kristiansand, Norway
Has thanked: 139 times
Been thanked: 206 times
#4
Post
by Hallvor » 2018-07-13 09:44
Try installing hardinfo. It has its own toolbox of benchmarks, and you'll get a nice report on your system.
My numbers:
Code: Select all
Benchmarks
CPU Blowfish
CPU Blowfish
This Machine 1366 MHz 3,637
Intel(R) Celeron(R) M processor 1.50GHz (null) 26.1876862
PowerPC 740/750 (280.00MHz) (null) 172.816713
CPU CryptoHash
CPU CryptoHash
This Machine 1366 MHz 552,391
CPU Fibonacci
CPU Fibonacci
This Machine 1366 MHz 1,415
Intel(R) Celeron(R) M processor 1.50GHz (null) 8.1375674
PowerPC 740/750 (280.00MHz) (null) 58.07682
CPU N-Queens
CPU N-Queens
This Machine 1366 MHz 4,782
FPU FFT
FPU FFT
This Machine 1366 MHz 0,893
FPU Raytracing
FPU Raytracing
This Machine 1366 MHz 3,525
Intel(R) Celeron(R) M processor 1.50GHz (null) 40.8816714
PowerPC 740/750 (280.00MHz)
dcihon
Posts: 279 Joined: 2012-09-23 16:18
Been thanked: 3 times
#5
Post
by dcihon » 2018-07-13 10:32
Thanks for the info.
I will try these suggestions and post back the results.
NFT5
df -h | grep > 20TiB
Posts: 597 Joined: 2014-10-10 11:38
Location: Canberra, Australia
Has thanked: 10 times
Been thanked: 43 times
#6
Post
by NFT5 » 2018-07-13 16:53
I think that those systemd-analyze commands are good for reviewing possible problems in your startup, but the speed of your system will also reflect in the results.
I'm fairly happy with mine:
Code: Select all
chris@BOSSDESK:~$ systemd-analyze
Startup finished in 2.070s (kernel) + 2.233s (userspace) = 4.303s
chris@BOSSDESK:~$ systemd-analyze blame
659ms nmbd.service
609ms samba-ad-dc.service
534ms vboxdrv.service
478ms home-chris-DATA.mount
293ms NetworkManager.service
285ms ModemManager.service
270ms exim4.service
249ms loadcpufreq.service
201ms binfmt-support.service
190ms smbd.service
166ms nfs-common.service
157ms systemd-logind.service
147ms keyboard-setup.service
139ms alsa-restore.service
133ms rc-local.service
133ms pppd-dns.service
132ms ssh.service
126ms minidlna.service
124ms rsyslog.service
99ms avahi-daemon.service
82ms console-setup.service
63ms lightdm.service
63ms networking.service
62ms systemd-fsck@dev-disk-by\x2duuid-09a78e50\x2d1716\x2d49c5\x2d84
56ms udisks2.service
56ms systemd-fsck@dev-disk-by\x2duuid-C3A8\x2dEEA4.service
51ms home-chris-VBOX\x2dVMS.mount
50ms packagekit.service
49ms rtkit-daemon.service
46ms speech-dispatcher.service
45ms colord.service
44ms irqbalance.service
43ms systemd-user-sessions.service
42ms polkitd.service
39ms rpcbind.service
34ms kbd.service
34ms systemd-tmpfiles-setup-dev.service
33ms hddtemp.service
32ms home-chris-MYDOCS\x2dVM.mount
32ms systemd-modules-load.service
29ms upower.service
23ms systemd-tmpfiles-setup.service
22ms systemd-setup-dgram-qlen.service
21ms systemd-udev-trigger.service
20ms sys-kernel-debug.mount
19ms boot-efi.mount
18ms proc-sys-fs-binfmt_misc.mount
17ms dev-mqueue.mount
17ms user@117.service
17ms hdparm.service
17ms dev-hugepages.mount
16ms home.mount
15ms cpufrequtils.service
15ms dev-sda8.swap
11ms systemd-journal-flush.service
10ms user@1000.service
10ms systemd-remount-fs.service
8ms systemd-update-utmp.service
6ms systemd-tmpfiles-clean.service
6ms kmod-static-nodes.service
6ms vboxballoonctrl-service.service
6ms vboxautostart-service.service
5ms vboxweb-service.service
3ms systemd-sysctl.service
3ms udev-finish.service
3ms systemd-update-utmp-runlevel.service
3ms systemd-random-seed.service
3ms systemd-udevd.service
2ms systemd-rfkill@rfkill0.service
1ms sys-fs-fuse-connections.mount
lines 48-70/70 (END)
hardinfo is also good for looking at system performance, but makes comparisons against a machine so old as to be irrelevant.
CwF
Global Moderator
Posts: 2636 Joined: 2018-06-20 15:16
Location: Colorado
Has thanked: 41 times
Been thanked: 192 times
#7
Post
by CwF » 2018-07-13 22:03
My hardinfo sampling...
| X5687 x2 | E5540 | E5-1650 | x4vm | E3-1225 | i386vm |
| CPU Blowfish | 0.94 | 2.32 | 1.18 | 2.53 | 2.87 | 2.81 |
| CPU CryptoHash | 1304.78 | 708.39 | 1404.05 | 757.44 | 771.58 | 798.69 |
| CPU FIbonacci | 1.76 | 2.29 | 1.56 | 1.53 | 1.69 | 2.56 |
| CPU N-Queens | 0.76 | 0.97 | 0.58 | 5.63 | 4.92 | 4.96 |
| FPU FFT | 1.07 | 1.38 | 1.00 | .93 | .97 | .91 |
| FPU Raytracing | 7.31 | 4.27 | 3.08 | 4.22 | 3.94 | 13.50 |
CwF
Global Moderator
Posts: 2636 Joined: 2018-06-20 15:16
Location: Colorado
Has thanked: 41 times
Been thanked: 192 times
#8
Post
by CwF » 2018-07-20 23:20
CwF wrote: My hardinfo sampling...
Ahh, I was hoping to see a few comps.
Bulkley
Posts: 6383 Joined: 2006-02-11 18:35
Has thanked: 2 times
Been thanked: 39 times
#9
Post
by Bulkley » 2018-07-20 23:40
The speed of any computer I use is directly related to how fast I can type, thus something less than 30wpm.
Seriously, I think that when we try to test the speed of a system we fool ourselves. We used to use glxgears until we learned that all it is useful for is to watch the gears go around. Etc. Real bench testing requires test gear and a fair amount of knowledge. Even if we can get it done properly all we get is bragging rights. It still doesn't improve ones typing.
Head_on_a_Stick
Posts: 14114 Joined: 2014-06-01 17:46
Location: London, England
Has thanked: 81 times
Been thanked: 132 times
#10
Post
by Head_on_a_Stick » 2018-07-21 11:15
CwF wrote: Ahh, I was hoping to see a few comps.
Hmmm, yes, sorry:
Code: Select all
_______________| Clevo (Intel i4330M@2.4GHz, HD4600) | X201 (Debian) | X201 (Alpine Linux) | Compaq CQ58 (AMD E1@1.2GHz)
FPU Raytrace | 3.27s | 4.98s | 5.17s | 22.06s
FPU FFT | 0.76s | 1.41s | 1.4s | 9.33s
CPU N-Queens | 5.11s | 7.7s | 6.89s | 32.98s
CPU Fibonacci | 1.28s | 2.65s | 3.46s | 7.02s
CPU Cryptohash | 478.43MiB/s | 264.91MiB/s | 256.36MiB/s | 73.55MiB/s
CPU Blowfish | 3.32s | 4.74s | 6.54s | 16.43s
The ThinkPad is an Intel i5 M520 (Ironlake) running at 2.4GHz (4MiB RAM).
That Compaq is dog slow...
EDIT: the Clevo is running Arch Linux and the Compaq is running Debian stable.
deadbang
CwF
Global Moderator
Posts: 2636 Joined: 2018-06-20 15:16
Location: Colorado
Has thanked: 41 times
Been thanked: 192 times
#11
Post
by CwF » 2018-07-21 13:56
Bulkley wrote: Seriously, I think that when we try to test the speed of a system we fool ourselves.
Sure. It's fun though. Big long extensive benchmarks I don't ever take the time. Many are so narrow in scope they're almost period or generation specific. On the other hand little quickies, and especially video specific test can give some guidance and reveal things otherwise missed. Your 30 words per minute was eclipsed somewhere back in the DX2 days, or before, and not so helpful. Setting up a regular machine to do regular things which seems to operate in regular ways, yes, I probably wouldn't bother. Over the last decade, really since dynamic clocking came about I've come to use a few old test to help me adjust that clock profile. Setting up non-regular machines in recent years I've found these little benchmarks critical, since some cpu generations have acted different. I could have settled for a variable result I didn't understand, or find the 3-6X performance increases I've found. One test led me to discover and verify a passed gpu yielding acceleration help to QXL virtual gpu, so now for a single non-regular purpose I can have 2 screens passed from a VM and its third 'window' on the host screen also working. I rarely use that, but better than assuming it didn't work... Overall, benchmarks are usefull when they are repeatable and used to guage the software more so than the hardware.
bedtime
Posts: 146 Joined: 2012-12-16 19:34
Has thanked: 1 time
Been thanked: 6 times
#13
Post
by bedtime » 2018-07-22 16:58
Here's mine:
Code: Select all
$ systemd-analyze
Startup finished in 1.205s (kernel) + 680ms (userspace) = 1.885s
graphical.target reached after 516ms in userspace
$ systemd-analyze blame
193ms dev-sda1.device
165ms user@1000.service
104ms udisks2.service
73ms accounts-daemon.service
58ms NetworkManager.service
46ms systemd-journald.service
42ms systemd-logind.service
32ms systemd-udev-trigger.service
29ms systemd-udevd.service
22ms systemd-fsck@dev-disk-by\x2duuid-8da2c527\x2dd70c\x2d46cd\x2d870d\x2d137108a0e88e.service
20ms networking.service
19ms systemd-fsck@dev-disk-by\x2duuid-d7b55308\x2df3a7\x2d40e9\x2dbbaa\x2d885f38261385.service
18ms systemd-fsck@dev-disk-by\x2duuid-7d90d2dd\x2dde29\x2d447a\x2daa91\x2d1ea55339fdd0.service
14ms systemd-journal-flush.service
10ms var.mount
9ms systemd-tmpfiles-setup.service
9ms home.mount
8ms polkit.service
8ms dev-disk-by\x2duuid-f221cdb0\x2ddac5\x2d44f8\x2d9eea\x2d4b4bbcd92554.swap
8ms systemd-tmpfiles-setup-dev.service
8ms sys-kernel-debug.mount
7ms kmod-static-nodes.service
7ms systemd-sysusers.service
6ms sysstat.service
6ms systemd-modules-load.service
6ms systemd-remount-fs.service
5ms tmp.mount
5ms dev-mqueue.mount
5ms systemd-update-utmp.service
5ms systemd-sysctl.service
4ms systemd-random-seed.service
4ms dev-hugepages.mount
3ms systemd-update-utmp-runlevel.service
3ms systemd-user-sessions.service
I could probably get it under a second if I customized the kernel, but I'm fine with it for now.
CwF
Global Moderator
Posts: 2636 Joined: 2018-06-20 15:16
Location: Colorado
Has thanked: 41 times
Been thanked: 192 times
#14
Post
by CwF » 2018-07-22 17:40
bedtime wrote: I could probably get it under a second if I customized the kernel, but I'm fine with it for now.
Maybe I should try for under 10!
Code: Select all
# systemd-analyze
Startup finished in 11.140s (kernel) + 1.189s (userspace) = 12.329s
My -blame is to long with the top 8 >194ms up to 564ms.
Note it takes over a minute to get post, and that happens just once a month. In my info above this is the E5.
None1975
df -h | participant
Posts: 1388 Joined: 2015-11-29 18:23
Location: Russia, Kaliningrad
Has thanked: 45 times
Been thanked: 65 times
#15
Post
by None1975 » 2018-07-23 14:05
bedtime wrote: 3ms systemd-user-sessions.service.
Man, it is really fast. Lighting fast.
Here mine:
Code: Select all
005 ~ > systemd-analyze
Startup finished in 2.829s (kernel) + 7.388s (userspace) = 10.218s
141 ~ > systemd-analyze blame
3.545s ufw.service
3.014s dev-sda1.device
1.663s apt-daily.service
815ms ntp.service
667ms keyboard-setup.service
536ms rsyslog.service
523ms atopacct.service
469ms systemd-tmpfiles-setup-dev.service
453ms systemd-fsck@dev-disk-by\x2duuid-1b720753\x2dba60\x2d4ce9\x2d8a
434ms systemd-remount-fs.service
419ms dev-hugepages.mount
418ms dev-mqueue.mount
417ms sys-kernel-debug.mount
345ms apt-daily-upgrade.service
298ms systemd-modules-load.service
297ms dev-disk-by\x2duuid-82823ecd\x2d9cd0\x2d4e91\x2da0dd\x2d902e3f9
256ms networking.service
237ms kmod-static-nodes.service
235ms systemd-udevd.service
189ms systemd-tmpfiles-setup.service
167ms systemd-journald.service
156ms systemd-udev-trigger.service
148ms systemd-journal-flush.service
lines 1-23...skipping...
3.545s ufw.service
3.014s dev-sda1.device
1.663s apt-daily.service
815ms ntp.service
667ms keyboard-setup.service
536ms rsyslog.service
523ms atopacct.service
469ms systemd-tmpfiles-setup-dev.service
453ms systemd-fsck@dev-disk-by\x2duuid-1b720753\x2dba60\x2d4ce9\x2d8a12\x2d5bed214725aa.service
434ms systemd-remount-fs.service
419ms dev-hugepages.mount
418ms dev-mqueue.mount
417ms sys-kernel-debug.mount
345ms apt-daily-upgrade.service
298ms systemd-modules-load.service
297ms dev-disk-by\x2duuid-82823ecd\x2d9cd0\x2d4e91\x2da0dd\x2d902e3f947e96.swap
256ms networking.service
237ms kmod-static-nodes.service
235ms systemd-udevd.service
189ms systemd-tmpfiles-setup.service
167ms systemd-journald.service
156ms systemd-udev-trigger.service
148ms systemd-journal-flush.service
137ms systemd-random-seed.service
130ms systemd-user-sessions.service
121ms systemd-logind.service
109ms console-setup.service
84ms user@1000.service
82ms home.mount
75ms systemd-tmpfiles-clean.service
60ms systemd-update-utmp.service
26ms systemd-sysctl.service
5ms systemd-update-utmp-runlevel.service
5ms alsa-restore.service
bedtime
Posts: 146 Joined: 2012-12-16 19:34
Has thanked: 1 time
Been thanked: 6 times
#16
Post
by bedtime » 2018-07-23 15:03
None1975 wrote: bedtime wrote: 3ms systemd-user-sessions.service.
Man, it is really fast. Lighting fast.
Here mine:
Code: Select all
005 ~ > systemd-analyze
Startup finished in 2.829s (kernel) + 7.388s (userspace) = 10.218s
141 ~ > systemd-analyze blame
3.545s ufw.service
3.014s dev-sda1.device
1.663s apt-daily.service
815ms ntp.service
667ms keyboard-setup.service
536ms rsyslog.service
523ms atopacct.service
469ms systemd-tmpfiles-setup-dev.service
453ms systemd-fsck@dev-disk-by\x2duuid-1b720753\x2dba60\x2d4ce9\x2d8a
434ms systemd-remount-fs.service
419ms dev-hugepages.mount
418ms dev-mqueue.mount
417ms sys-kernel-debug.mount
345ms apt-daily-upgrade.service
298ms systemd-modules-load.service
297ms dev-disk-by\x2duuid-82823ecd\x2d9cd0\x2d4e91\x2da0dd\x2d902e3f9
256ms networking.service
237ms kmod-static-nodes.service
235ms systemd-udevd.service
189ms systemd-tmpfiles-setup.service
167ms systemd-journald.service
156ms systemd-udev-trigger.service
148ms systemd-journal-flush.service
lines 1-23...skipping...
3.545s ufw.service
3.014s dev-sda1.device
1.663s apt-daily.service
815ms ntp.service
667ms keyboard-setup.service
536ms rsyslog.service
523ms atopacct.service
469ms systemd-tmpfiles-setup-dev.service
453ms systemd-fsck@dev-disk-by\x2duuid-1b720753\x2dba60\x2d4ce9\x2d8a12\x2d5bed214725aa.service
434ms systemd-remount-fs.service
419ms dev-hugepages.mount
418ms dev-mqueue.mount
417ms sys-kernel-debug.mount
345ms apt-daily-upgrade.service
298ms systemd-modules-load.service
297ms dev-disk-by\x2duuid-82823ecd\x2d9cd0\x2d4e91\x2da0dd\x2d902e3f947e96.swap
256ms networking.service
237ms kmod-static-nodes.service
235ms systemd-udevd.service
189ms systemd-tmpfiles-setup.service
167ms systemd-journald.service
156ms systemd-udev-trigger.service
148ms systemd-journal-flush.service
137ms systemd-random-seed.service
130ms systemd-user-sessions.service
121ms systemd-logind.service
109ms console-setup.service
84ms user@1000.service
82ms home.mount
75ms systemd-tmpfiles-clean.service
60ms systemd-update-utmp.service
26ms systemd-sysctl.service
5ms systemd-update-utmp-runlevel.service
5ms alsa-restore.service
I'm running an HP Prodesk 600, with an i7-7700, SSD, and 8g ram. *hugs his compy*
Note that I am running far less services than you, so that alone could be a big factor. Also, I running i3 window manager, so not much background services to be loaded (as opposed to gnome/kde...).
It seems that your apt services are slowing you down. And your ufw (firewall) service, I'm wondering if you could somehow load that in the background and not have systemd wait on it? Network connections used to slow my boot speed down as well until I told sysd to not wait on them.
This is impressive to me; I remember my alsa-restore taking several seconds to load, so I ditched it. I might reconsider adding it back now that I see this.
None1975
df -h | participant
Posts: 1388 Joined: 2015-11-29 18:23
Location: Russia, Kaliningrad
Has thanked: 45 times
Been thanked: 65 times
#17
Post
by None1975 » 2018-07-23 17:00
bedtime wrote: SSD, and 8g ram. *hugs his compy* :PAlso, I running i3 window manager, so not much background services to be loaded (as opposed to gnome/kde...).
I also running wm-xmonad. No DM, just startx. The system has a total of 1026 packages. But my system is on hdd, not on sdd. Also, my system has twice less RAM than yours. Cpu is Intel Core i7 CPU 920 @ 2.793GHz
t seems that your apt services are slowing you down. And your ufw (firewall) service, I'm wondering if you could somehow load that in the background and not have systemd wait on it?
Yes I'm thinking about this factor. No, I will not squander much of it. The system is not on the ssd. No matter... After four seconds, the total time will be six seconds.