[SOLVED] Arch boot is too slow. Windows 10 boots faster.

vero2513 · 2021-02-13 15:47:21

seth wrote:

You can use dd to measure the disk performance (read/write from /dev/zero and into /dev/null) and for a start "cpupower frequency-info" to see whether your CPU hangs in a low frequency.
https://wiki.archlinux.org/index.php/CP … g#cpupower
Edit: the updated journal does not suggest an early microcode load.

This is the output of cpupower frequency-info

cpupower frequency-info
analisi della CPU 0:
  driver: intel_cpufreq
  CPU che operano alla stessa frequenza hardware: 0 
  CPU che è necessario siano coordinate dal software: 0
  latenza massima durante la transizione: 20.0 us
  limiti hardware: 1.20 GHz - 3.40 GHz
  gestori disponibili: conservative ondemand userspace powersave performance schedutil
  gestore attuale: la frequenza deve mantenersi tra 1.20 GHz e 3.40 GHz.
                   Il gestore "schedutil" può decidere quale velocità usare
                  in questo intervallo.
  current CPU frequency: Unable to call hardware
  current CPU frequency: 1.29 GHz (asserted by call to kernel)
  boost state support:
    Supported: yes
    Active: yes
    25500 MHz max turbo 4 active cores
    25500 MHz max turbo 3 active cores
    25500 MHz max turbo 2 active cores
    25500 MHz max turbo 1 active cores

seth · 2021-02-13 15:58:53

Should I remove this highlighted?

No, but the grub.cfg is (hopefully) capped, it's still "i915.enable_fbc=1" and please check

dmesg | grep "microcode updated early"

Also put the CPU under some stress and check the frequency again while it's supposed to do something, https://wiki.archlinux.org/index.php/St … and_Memory

Last edited by seth (2021-02-13 16:00:39)

vero2513 · 2021-02-13 16:01:28

seth wrote:

Should I remove this highlighted?
No, but the grub.cfg is (hopefully) capped, it's still "i915.enable_fbc=1" and please check
dmesg | grep "microcode updated early"
Also put the CPU under some stress and check the frequency again while it's supposed to do something, https://wiki.archlinux.org/index.php/St … and_Memory

I did

dmesg | grep "microcode updated early"

but it returns nothing

loqs · 2021-02-13 16:06:37

vero2513 is the aim to get to a boot time of around 6 second as you experienced under Windows? If so then have you investigated suspend to disk? Which is comparable to Windows fast boot.

vero2513 · 2021-02-13 16:09:39

loqs wrote:

vero2513 is the aim to get to a boot time of around 6 second as you experienced under Windows? If so then have you investigated suspend to disk? Which is comparable to Windows fast boot.

Yes..It's way too complicate though..maybe I should swtich to manjaro considering that it configures everything and I probably (?) won't have problem?

euromatlox · 2021-02-13 16:09:52

Is there entry for swap in fstab ? I don't know it it's required or not, but if it's just a swap (no important files), why not have it there?

UUID=xxxxxxxxxxxxx	none		swap		defaults	0 0

xxxxxxxxxxxxx should be replaced by real uuid of swap partition. Be careful to select correct uuid.

As a superuser you can easily get a list of uuids, like this for example:

blkid | grep UUID=

Also check Arch Wiki of swap.

vero2513 · 2021-02-13 16:11:53

euromatlox wrote:

Is there entry for swap in fstab ? I don't know it it's required or not, but if it's just a swap (no important files), why not have it there?
UUID=xxxxxxxxxxxxx	none		swap		defaults	0 0
xxxxxxxxxxxxx should be replaced by real uuid of swap partition. Be careful to select correct uuid.
As a superuser you can easily get a list of uuids, like this for example:
blkid | grep UUID=
Also check Arch Wiki of swap.

Yes I already added the swap partition to the fstab and got the UUID by using that command

euromatlox · 2021-02-13 16:26:21

There's this older post of random numbers / entropy.

Linux provides device interfaces (/dev/random and /dev/urandom) to a pool of random numbers collected from system interrupt service routines. On some systems, especially on those systems with high needs or limited user interaction, the standard collection mechanism cannot meet demand...more here.

As Seth says, very low possibility that this is the case, but good to know that this kind of behavior can happen too.

Last edited by euromatlox (2021-02-13 16:35:05)

seth · 2021-02-13 16:30:46

Check the journals he posted for crng, also there're no stalls.
This isn't an entropy issue.

loqs · 2021-02-13 16:32:31

vero2513 wrote:

loqs wrote:
vero2513 is the aim to get to a boot time of around 6 second as you experienced under Windows? If so then have you investigated suspend to disk? Which is comparable to Windows fast boot.
Yes..It's way too complicate though..maybe I should swtich to manjaro considering that it configures everything and I probably (?) won't have problem?

You have a swap file which you have now added to the fstab so you would need to add one kernel parameter referencing the swap file by its UUID which you also mention knowing but that is way too complicated?
How is that more complicated than all the other options you added to the boot loader?

Suspend_and_hibernate#Required_kernel_parameters

vero2513 · 2021-02-13 16:35:18

loqs wrote:

vero2513 wrote:
loqs wrote:
vero2513 is the aim to get to a boot time of around 6 second as you experienced under Windows? If so then have you investigated suspend to disk? Which is comparable to Windows fast boot.
Yes..It's way too complicate though..maybe I should swtich to manjaro considering that it configures everything and I probably (?) won't have problem?
You have a swap file which you have now added to the fstab so you would need to add one kernel parameter referencing the swap file by its UUID which you also mention knowing but that is way too complicated?
How is that more complicated than all the other options you added to the boot loader?
Suspend_and_hibernate#Required_kernel_parameters

Should I add it to the /etc/default/grub or to /boot/grub/grub.cfg?

loqs · 2021-02-13 16:38:14

Any changes you make to /boot/grub/grub.cfg will be overwritten the next time you run:

# grub-mkconfig -o /boot/grub/grub.cfg

d_fajardo · 2021-02-13 16:38:28

# Static information about the filesystems.
# See fstab(5) for details.
# <file system> <dir> <type> <options> <dump> <pass>
# /dev/sda3
UUID=ac89ab0d-4661-49e0-9974-d4ce498f4501 / ext4 rw,relatime 0 1
# /dev/sda2
UUID="50030b78-9dd7-4cda-91c4-deb0e467793a" none swap sw 0 0

It's probably also going to help if there'san entry of your /efi partition here in fstab.

loqs · 2021-02-13 16:40:17

d_fajardo wrote:

It's probably also going to help if there'san entry of your /efi partition here in fstab.

How would that help? What during boot needs to access the ESP after the kernel has loaded?

finoderi · 2021-02-13 16:41:36

vero2513 wrote:

Should I add it to the /etc/default/grub or to /boot/grub/grub.cfg?

https://wiki.archlinux.org/index.php/Ke … eters#GRUB

vero2513 · 2021-02-13 17:03:04

I've followed this wiki and added the swap on kernel parameters
https://wiki.archlinux.org/index.php/Po … parameters
I rebooted and wrote the systemd-analyze critical-chain in the terminal..the output is this:

systemd-analyze critical-chain
The time when unit became active or started is printed after the "@" character.
The time the unit took to start is printed after the "+" character.

graphical.target @16.839s
└─sddm.service @16.838s
  └─systemd-user-sessions.service @16.803s +31ms
    └─network.target @16.799s
      └─wpa_supplicant.service @21.238s +350ms
        └─dbus.service @14.181s
          └─basic.target @14.150s
            └─sockets.target @14.150s
              └─pcscd.socket @14.150s
                └─sysinit.target @14.126s
lines 1-13...skipping...
The time when unit became active or started is printed after the "@" character.
The time the unit took to start is printed after the "+" character.

graphical.target @16.839s
└─sddm.service @16.838s
  └─systemd-user-sessions.service @16.803s +31ms
    └─network.target @16.799s
      └─wpa_supplicant.service @21.238s +350ms
        └─dbus.service @14.181s
          └─basic.target @14.150s
            └─sockets.target @14.150s
              └─pcscd.socket @14.150s
                └─sysinit.target @14.126s
                  └─systemd-update-done.service @14.079s +46ms
                    └─ldconfig.service @11.622s +2.455s
                      └─local-fs.target @11.621s
                        └─tmp.mount @11.617s +3ms
                          └─swap.target @11.585s
                            └─dev-disk-by\x2duuid-50030b78\x2d9dd7\x2d4cda\x2d91c4\x2ddeb0e467793a.swap @11.423s +159ms
                              └─dev-disk-by\x2duuid-50030b78\x2d9dd7\x2d4cda\x2d91c4\x2ddeb0e467793a.device @11.422s

It looks like the swap takes even more to load..maybe I did something wrong..but I can't understand what

loqs · 2021-02-13 17:06:45

How did you suspend the system? Please post the updated journal.

vero2513 · 2021-02-13 17:13:43

loqs wrote:

How did you suspend the system? Please post the updated journal.

Yes:
http://ix.io/2PiK

seth · 2021-02-13 17:16:16

"Yes" what?
The point of the exercise was to try a resume from hibernation - the journal doesn't reflect that.

vero2513 · 2021-02-13 17:19:54

seth wrote:

"Yes" what?
The point of the exercise was to try a resume from hibernation - the journal doesn't reflect that.

Ok I hibernated the laptop and redid the journalctl..this is the output:
http://ix.io/2PiO

loqs · 2021-02-13 17:52:45

feb 13 18:18:10 Arch-Linux systemd-sleep[8938]: System resumed.
....
feb 13 18:18:17 Arch-Linux kernel: wlo1: authenticated
....
feb 13 18:18:20 Arch-Linux NetworkManager[526]: <info>  [1613236700.4632] manager: NetworkManager state is now CONNECTED_GLOBAL

Seven or ten seconds on resume going off those the above timings. Does that match your experience vero2513?

euromatlox · 2021-02-13 21:10:39

I'll mention these two earlier posts here, might help or not:
KDE don't start at startup
KDE Plasma: plasmashell using 8GB after update

vero2513 · 2021-02-14 06:08:15

loqs wrote:

feb 13 18:18:10 Arch-Linux systemd-sleep[8938]: System resumed.
....
feb 13 18:18:17 Arch-Linux kernel: wlo1: authenticated
....
feb 13 18:18:20 Arch-Linux NetworkManager[526]: <info>  [1613236700.4632] manager: NetworkManager state is now CONNECTED_GLOBAL

Seven or ten seconds on resume going off those the above timings. Does that match your experience vero2513?

Yes..the system has blocked for 4 seconds then it shutdown and when I powered on the laptop it boot in 2/3 seconds

vero2513 · 2021-02-14 06:18:42

euromatlox wrote:

I'll mention these two earlier posts here, might help or not:
KDE don't start at startup
KDE Plasma: plasmashell using 8GB after update

I have no problem..plasmashell loads perfectly and is very fast (even though sometimes display has bugs and I don't know why..I also followed the archwiki to install plasma and I didn't install xf86-intel-video) but the problem is how much time it takes

Should I make a fresh install of Archlinux with no swap partition or with no hibernation and see if it boots faster?

Last edited by vero2513 (2021-02-14 06:28:50)

finoderi · 2021-02-14 06:32:53

vero2513 wrote:

Should I make a fresh install of Archlinux with no swap partition or with no hibernation and see if it boots faster?

And it'll boot faster because of what?
Maybe it would be better to read the wiki for better understanding what's going on during the system boot and why magic and miracles are so rare these days.

Arch Linux

#51 2021-02-13 15:47:21

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#52 2021-02-13 15:58:53

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#53 2021-02-13 16:01:28

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#54 2021-02-13 16:06:37

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#55 2021-02-13 16:09:39

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#56 2021-02-13 16:09:52

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#57 2021-02-13 16:11:53

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#58 2021-02-13 16:26:21

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#59 2021-02-13 16:30:46

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#60 2021-02-13 16:32:31

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#61 2021-02-13 16:35:18

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#62 2021-02-13 16:38:14

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#63 2021-02-13 16:38:28

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#64 2021-02-13 16:40:17

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#65 2021-02-13 16:41:36

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#66 2021-02-13 17:03:04

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#67 2021-02-13 17:06:45

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#68 2021-02-13 17:13:43

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#69 2021-02-13 17:16:16

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#70 2021-02-13 17:19:54

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#71 2021-02-13 17:52:45

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#72 2021-02-13 21:10:39

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#73 2021-02-14 06:08:15

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#74 2021-02-14 06:18:42

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

#75 2021-02-14 06:32:53

Re: [SOLVED] Arch boot is too slow. Windows 10 boots faster.

Board footer