Tolerances
The absolute tolerance used for floating point comparisons is hard-coded in the project to 1e-10
for Float64
and to 1f-5
for Float32
. You can use ScopedValues.jl to customize these tolerance values in specific computations:
using Meshes
using ScopedValues
with(Meshes.ATOL64 => 1e-9, Meshes.ATOL32 => 1f-4) do
# do your computations with custom tolerances
end