Ploting is much slower on Linux than on Windows

[mjchow]

I'm using the same hardware configuration and the same parameters on both Linux and Windows

-b 6000 -r 8

the CPU utilization on Windows is 400%
CPU utilization on Linux can't even reach 200%

[heus12]

no，linux is 10% faster than windows

if you compile and optimization linux kernel, that can 20%-40% faster than windows

[mjchow]

Which linux distribution do you use?

[heus12]

Which linux distribution do you use?

Every linux distribution can use, but you need to compile and optimize the Linux kernel yourself

[mjchow]

F1 complete, time: 659.182 seconds. CPU (56.44%) this is in ubuntu20.04
F1 complete, time: 200.342 seconds. CPU (189.32%) this is in windows 10

Why is CPU utilization so different for the same plot parameter ？ my plot parameter is (-k 32 -b 8000 -u 128 -r 8)

[loppefaaret]

Windows wil by default ensure that any SSD based media gets TRIM.
On the other hand, under most Linux distro's, you need to specify this yourself, with adding the mount option discard to an SSD based media.
Perhaps this is an issue you are facing ?

[mjchow]

@mjchow ntfs on Linux is slow. You should use native file system on *nix systems (ie ext4)

thank u. @leonidcan
After using ext4 in my SSD, the speed of plot is improved, but it is still slower than that under windows

[loppefaaret]

did you use discard with ext4 ?

[mjchow]

@loppefaaret
I don't seem to be using discard
How to check if I used discard

i typed mount | grep discard is no return

[loppefaaret]

if there is no return, there is no drive mounted with discard, so mount them with discard, as above

[mjchow]

i mount with discard
but there is no effect

[leonidcan]

I found that my linux is faster than windows (used on the same machine) but I am using software raid, how do I apply TRIM to RAID0 in mdadm ? Can't find good answer but I notice that plotting time go up overtime :(

[mjchow]

I feel that CPU utilization on Linux is lower than that on windows
But I don't know why
My CPU is e5-2695v3

[BrandtH22]

@mjchow , since this ticket was opened a number of years ago without further comments it seems this issue has been resolved.
We will be closing this ticket but if we have closed this ticket in error do not hesitate to reach out to us again with any followup questions or comments.

The best place to reach our support team is on Discord (https://discord.gg/chia) or by reopening this ticket.