The Linux kernel’s I/O subsystem has evolved steadily to keep pace with modern storage hardware, with most layers gradually enhanced to exploit parallelism.
Fascinating to see writeback finally get a parallelism boost. Do you think per-allocation-group parallelization might still outperform the current inode-affinity approach?
Fascinating to see writeback finally get a parallelism boost. Do you think per-allocation-group parallelization might still outperform the current inode-affinity approach?