scx_rustland: improve scheduler cpu selection #57
Conversation
The current CPU selection logic in the scheduler presents some inefficiencies. When a task is drained from the BPF queue, the scheduler immediately checks whether the CPU previously assigned to the task is still idle, assigning it if it is. Otherwise, it iterates through available CPUs, always starting from CPU #0, and selects the first idle one without updating its state. This approach is consistently applied to the entire batch of tasks drained from the BPF queue, resulting in all of them being assigned to the same idle CPU (also with a higher likelihood of allocation to lower CPU ids rather than higher ones). While dispatching a batch of tasks to the same idle CPU is not necessarily problematic, a fairer distribution among the list of idle CPUs would be preferable. Therefore change the CPU selection logic to distribute tasks equally among the idle CPUs, still maintaining the preference for the previously used one. Additionally, apply the CPU selection logic just before tasks are dispatched, rather than assigning a CPU when tasks are drained from the BPF queue. This adjustment is important, because tasks may linger in the scheduler's internal structures for a bit and the idle state of the CPUs in the system may change during that period. Signed-off-by: Andrea Righi <[email protected]>
When the scheduler decides to assign a different CPU to the task always make sure the assignment is valid according to the task cpumask. If it's not valid simply dispatch the task to the global DSQ. This prevents the scheduler from exiting with errors like this: 09:11:02 [WARN] EXIT: SCX_DSQ_LOCAL[_ON] verdict target cpu 7 not allowed for gcc[440718] In the future we may want move this check directly into the user-space scheduler, but for now let's keep this check in the BPF dispatcher as a quick fix. Signed-off-by: Andrea Righi <[email protected]>
| // mitigate migration overhead), otherwise pick the next idle CPU available. | ||
| if !idle_cpus.contains(&task.cpu) { | ||
| task.cpu = *cpu; | ||
| } |
There was a problem hiding this comment.
This still can lead to multiple tasks being assigned to the same cpu, right? Something to improve in the future, I guess.
There was a problem hiding this comment.
@htejun exactly, we can still assign the same idle cpu to multiple tasks if their prev_cpu is still idle, that is not super bad, being too aggressive with migrations also comes with a cost. We could also speculate on the fact that tasks with a close vruntime should be spread across multiple CPUs, so that they can run pretty much at the same time in parallel, rather than piling them up on the same idle CPU (that is also something that I'd like to experiment).
But there is definitely room for improvements in the cpu selection logic and in the userspace scheduler in general.
And I'm happy to see that we are reaching a stage where the BPF component is becoming a more "stable" backend (requiring minimal changes), so that we can start channeling all the optimizations, improvements, tests and experiments in the user-space part. If that's the case it would be a nice success story for this little project IMHO. :)
Implement a more reasonable idle CPU selection logic in the user-space scheduler.