You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
I use TLP to optimize the energy efficiency of VM host machines. Right now, the settings are applied across the entire topology of the CPU(s), so when it determines that "powersave" should be enabled, it does it on all cores.
This is generally fine, but I pin critical tasks like I/O to core 0, and therefore I'd like that core to run at a higher minimum state, to reduce latency.
Describe the solution you'd like
I'd like it if there were a way to omit specific cores from the processor settings (easier solution), or apply different settings for certain cores (more complex solution).
Describe alternatives you've considered
Is there the possibility run hook scripts at the events where TLP applies these changes, so I can run scripts override TLP's work on specific cores?
The text was updated successfully, but these errors were encountered:
Hi, I can imagine the simple solution with a cpu core blacklist.
The complex solution would mean that all core-related settings, i.e. CPU_SCALING_* and CPU_ENERGY_* need to be extended analogously to DISK_*. Since cpus nowadays tend to have many cores/threads, the configuration file would become confusing for users. I want to avoid this.
Is your feature request related to a problem? Please describe.
I use TLP to optimize the energy efficiency of VM host machines. Right now, the settings are applied across the entire topology of the CPU(s), so when it determines that "powersave" should be enabled, it does it on all cores.
This is generally fine, but I pin critical tasks like I/O to core 0, and therefore I'd like that core to run at a higher minimum state, to reduce latency.
Describe the solution you'd like
I'd like it if there were a way to omit specific cores from the processor settings (easier solution), or apply different settings for certain cores (more complex solution).
Describe alternatives you've considered
Is there the possibility run hook scripts at the events where TLP applies these changes, so I can run scripts override TLP's work on specific cores?
The text was updated successfully, but these errors were encountered: