Linux 5.10 Scheduler Updates Bring SMT Balancing Tweaks
Ingo Molnar as usual is quite quick in submitting his changes for the new kernel merge window in the areas he oversees. With the scheduler changes for Linux 5.10 there are some changes worth mentioning.
- Cache hotness is now ignored for SMT migration since they share the same core and in turn the same caches.
- A set of patches to improve the fairness between CFS tasks. The work aims to improve the fairness of rebalancing of the system and benchmarks have shown a 0.5~2.7% improvement for the Hackbench scheduler test on ARM64. In another test with realtime application threads have been a ~2% improvement.
- NUMA balancing improvements.
- A new scheduler debug tracepoint for tracking the CPU capacity that is useful for Arm Energy Aware Scheduling and related task placement / load balancing optimizations. There are also updates to the Arm EAS code.
- A new MEMBARRIER_CMD_PRIVATE_EXPEDITED_RSEQ flag for the Restartable Sequences (Rseq) system call. This is based on Google-internal work.
More details on these scheduler changes for Linux 5.10 via this pull request.
- Cache hotness is now ignored for SMT migration since they share the same core and in turn the same caches.
- A set of patches to improve the fairness between CFS tasks. The work aims to improve the fairness of rebalancing of the system and benchmarks have shown a 0.5~2.7% improvement for the Hackbench scheduler test on ARM64. In another test with realtime application threads have been a ~2% improvement.
- NUMA balancing improvements.
- A new scheduler debug tracepoint for tracking the CPU capacity that is useful for Arm Energy Aware Scheduling and related task placement / load balancing optimizations. There are also updates to the Arm EAS code.
- A new MEMBARRIER_CMD_PRIVATE_EXPEDITED_RSEQ flag for the Restartable Sequences (Rseq) system call. This is based on Google-internal work.
More details on these scheduler changes for Linux 5.10 via this pull request.
6 Comments