Expand description
§honeycomb-benches
This crate contains all benchmarks of the project. It also contains simple binaries used to profile and further optimize the implementation.
§Binary
The package provides a single binary, hc-bench
, which exposes several benchmarks as
subcommands. For details on options and arguments, run:
cargo run --bin hc-bench -- --help
Benchmarks are described in the documentation of their respective modules.
§Available benchmarks
§Criterion-based
builder
- grid building routines at fixed sizebuilder-grid-size
- grid building routines over a range of grid sizesfetch_icells
-CMap2::iter_<CELL>
methodsgrisubal
- grisubal kernel with a fixed size gridgrisubal-grid-size
- grisubal kernel over a range of grid granularitytriangulate-quads
- triangulate all cells of a mixed-mesh
§Iai-callgrind-based
prof-dim2-basic
-CMap2
basic operations benchmarksprof-dim2-build
-CMap2
constructor & building functions benchmarksprof-dim2-sewing-unsewing
-CMap2
(un)sewing & (un)linking methods benchmarks
Modules§
- cut_
edges cut-edges
benchmark- grid_
gen - Grid generation benchmark
- grisubal
grisubal
benchmark- remesh
- 2D remshing pipeline benchmark
- shift
shift
benchmark
Macros§
- prof_
init - Attempt to open a fifo at the path
/tmp/hc_perf_control
. - prof_
start - Write to the
/tmp/hc_perf_control
to enable perf sampling if${$var}
is defined. - prof_
stop - Write to the
/tmp/hc_perf_control
to disable perf sampling if${$var}
is defined.