| short_description: > |
| pmbench - paging and virtual memory benchmark |
| |
| description: > |
| pmbench is a micro-benchmark that profiles system paging performance by |
| measuring latencies of each memory access throughout the run and |
| reporting the statistics of measured latencies. |
| |
| homepage: |
| |
| parameters: |
| runtime: |
| mapsize: Specify virtual address map size in Mebibyte (2^20 bytes) unit. The default is 256. |
| setsize: Specify working set size in Mebibyte unit. It should be no greater than mapsize. The default is 128. |
| pattern: [linear, uniform, normal, pareto] |
| shape: Pattern-specific numeric parameter that determines the shape of the distribution. See Usage for details. |
| jobs: Number of concurrent worker threads to spawn for the benchmark. Some releases may not support this option. The default is 1. |
| timestamp: [rdtsc, rdtscp, perfc] |
| cold: [y, n] |
| access: [histo, touch] |
| |
| results: |
| pmbench.latency.ns.average: |
| pmbench.latency.ns.0-256%: |
| pmbench.latency.ns.256-512%: |
| pmbench.latency.ns.512-1K%: |
| pmbench.latency.ns.1K-2K%: |
| pmbench.latency.ns.2K-4K%: |
| pmbench.latency.ns.4K-8K%: |
| pmbench.latency.ns.8K-16K%: |
| pmbench.latency.ns.16K-32K%: |
| pmbench.latency.ns.32K-64K%: |
| pmbench.latency.ns.64K-128K%: |
| pmbench.latency.ns.128K-256K%: |
| pmbench.latency.ns.256K-512K%: |
| pmbench.latency.ns.512K-1M%: |
| pmbench.latency.ns.1M-2M%: |
| pmbench.latency.ns.2M-4M%: |
| pmbench.latency.ns.4M-8M%: |
| pmbench.latency.ns.8M-16M%: |
| pmbench.latency.ns.16M-32M%: |
| pmbench.latency.ns.32M-64M%: |
| pmbench.latency.ns.64M-128M%: |
| pmbench.latency.ns.128M-256M%: |
| pmbench.latency.ns.256M-512M%: |
| pmbench.latency.ns.512M-1G%: |
| pmbench.latency.ns.1G-inf%: |
