4 ms·
I’m not a Java expert, but potentially due to the CPU/memory OS scheduler? Try running the workload with pinned CPUs assigned and not across NUMA nodes.
by plasma 5y ago
I’m not a Java expert, but potentially due to the CPU/memory OS scheduler?
Try running the workload with pinned CPUs assigned and not across NUMA nodes.