GoogleCloudPlatform / GoogleCloudPlatform/PerfKitBenchmarker
Add hardinfo2 benchmark
- Dominant language
- Python
- Stars
- 2k
- Forks
- 562
- Avg merge
- 4h 55m
- Merged PRs (30d)
- 69
Description
https://github.com/hardinfo2/hardinfo2
https://hardinfo2.org/
This task adds hardinfo2 as a new benchmark for PKB. It can be divided into a few different components:
1) Add the benchmarking framework
- create hardinfo2_benchmark.py file in linux_benchmarks
- populate BENCHMARK_NAME and BENCHMARK_CONFIG constants so that the benchmark can be found by PKB
- create GetConfig, Prepare, Run and Cleanup functions handlers with pass/return [] as function content.
At this point you can run your new benchmark in PKB (though it will not do anything yet).
2) Install hardinfo2
- try installing on your machine using the guide on github.
- Create a new package called hardinfo2.py in linux_packages.
- create a new function, Install, that takes a vm as an input and installs hardinfo2 on that vm. You should be able to wrap shell commands as vm.RemoteCommand
add unit tests to make sure the remote commands are issued using mock to mock the vm.
- Installing on Ubuntu2404 is top priority, followed by other Linux distributions e.g. Debian, Rhel, Centos etc.
3) Run hardinfo2
- Download and run hardinfo2 locally, you don't need to use PKB to run hardinfo2. This part is about parsing the output into some sensible format.
- Store the hardinfo2 output as a file in the data directory under data/hardinfo2. This file will be your raw data for parsing and parsing unit tests.
- Add a function in hardinfo2.py that you added to linux_packages with a sensible name, e.g. ParseResults
- Parse results should take a str as input and produce a list of PKB Samples as output. You goal is to parse the output into useful samples, where each sample as a metric name, metric value, metric unit, metric metadata. Each row of hardinfo2's output should be a separate metric.
Test the parser function
Contributor guide
Assessment
This issue has not been assessed yet.