What Is a Shared vCPU? Cores, Threads and VPS Performance

A shared vCPU is a virtual processor scheduled on physical CPU resources also used by other workloads. It is not a promise of one dedicated physical core, and its performance depends on both hardware and contention.

Why core counts can mislead

A physical processor core may expose multiple hardware threads. Those threads share execution resources. Allocating more virtual CPUs can help parallel work, but it does not automatically multiply single-thread speed or remove host scheduling delays.

Clock speed, processor generation, cache, memory access and storage also matter. A newer six-core server may outperform an older system with many more threads for a latency-sensitive application. Compare the workload, not only the advertised number of cores.

What to measure in Linux

Use uptime for load averages and top for process activity. With sysstat installed, mpstat 1 5 gives a short sampling window. Steal time can indicate that a virtual CPU waited for the hypervisor, but one sample cannot identify every source of slowness.

Measure request latency, queue depth and error rate alongside CPU. A blocked disk task can increase load while CPU utilization remains low.

Match the plan to the workload

Autdo's currently listed plans use shared vCPU. Choose enough memory for the application and leave room for the OS. Do not design around continuously monopolizing shared CPU, or assume an automatic CPU policy that has not been published.

Read CPU load troubleshooting and performance plan selection. Live prices and availability are in the catalog.

Reference: Linux proc filesystem documentation.