snoop: benchmark in parallel
Context
This MR allows to perform benchmarks in parallel. The tool automatically pins the master process and worker processes on distinct cores.
Some notes:
- this only works on Linux (dynamic checks prevent using this new feature on non-Linux platforms)
- Depends on !5496 (merged)
TODO:
-
Assess adverse effects
Benchmarking in parallel should be avoided, except for long-running workloads where the added variability is negligible in front of the total execution time. Another reason to merge this MR despite disappointing results is that performing measurements in a dedicated process separated from the main one reduces variability.
Manually testing the MR
./tezos-snoop benchmark Blake2b_example and save to blake2b_full.workload --bench-num 300 --nsamples 50 --ncores 4
Checklist
-
Document the interface of any function added or modified (see the coding guidelines) -
Document any change to the user interface, including configuration parameters (see node configuration) -
Provide automatic testing (see the testing guide). -
For new features and bug fixes, add an item in the appropriate changelog ( docs/protocols/alpha.rst
for the protocol and the environment,CHANGES.rst
at the root of the repository for everything else). -
Select suitable reviewers using the Reviewers
field below. -
Select as Assignee
the next person who should take action on that MR
Edited by Ilias Garnier