I have enhanced wspy to read performance counters. It now has three different instrumentation methods: Reading ftrace logfiles from kernel subsystem. Currently reads the scheduler events for fork/exec/exit to construct process trees Reading /proc/stat on periodic basis (once per second) … Continue reading →
Two improvements have been added to the wspy program: Added a -z option that creates a zip archive with data files related to the wspy run Added a script that calls gnuplot to plot CPU usage over time. Below is … Continue reading →
Kicked off a run with the Phoronix CPU suite using likwid-perfctr. Benchmarks that were single-threaded were pinned to a single CPU, others to all CPUs. #!/bin/bash likwid-perfctr -a | tail +3 | awk ‘{ print $1 }’ | while read … Continue reading →
As one of the first steps inpriming the pump for Phoronix benchmarks, I ran the wspy program on 21 candidate CPU benchmarks. My goal was to start with a rough characterization, e.g. single-threaded vs. multi-threaded or cpu-bound vs not. #!/bin/bash … Continue reading →
Another useful tool from the likwid performance monitoring and benchmarking suite is likwid-topology. Provides thread (hyperthread), cache (L1, L2, L3) and NUMA (memory) topology together. Output below shows my i7-4770S CPU is a 4-core hyperthreaded processor with a 32KB L1D … Continue reading →
Below is a table that summarizes installation and run status of the Phoronix Test Suite CPU suite. Some background context of how this fits and where I hope to head from here… Where I’d like to head is comparing micro-architectural … Continue reading →
During a recent look at likwid-perfctr the performance counters didn’t look right in several aspects: CPU core to CPU core differences in what should be a symmetric benchmark and where wspy results showed processes balanced run to run differences with … Continue reading →
Created a simple monitoring tool named wspy for workload spy. Source is hosted at http://www.github.com/mvermeulen/wspy This is a wrapper program that collects data as the program runs. The initial program has two data collectors: trace – which uses the ftrace … Continue reading →
The likwid performance monitoring and benchmarking suite has multiple tools including: likwid-topology – print thread, cache and NUMA topology likwid-pin – pin threaded applications to processors likwid-bench – micro-benchmarking application likwid-agent – monitoring agent for hardware performance counters As of … Continue reading →