Would be good to compare the efficiency of various operations (just basic arithmetic to start with) between:
- plain
f64 as a reference point
SciFloat
SciDecimal
rust_decimal::Decimalbigdecimal::BigDecimal
That way we can see (and inform people) what the performance cost of using decimal rather than binary floats and/or of tracking the uncertainty is, and we can start to identify bottlenecks.
Would be good to compare the efficiency of various operations (just basic arithmetic to start with) between:
f64as a reference pointSciFloatSciDecimalrust_decimal::Decimalbigdecimal::BigDecimalThat way we can see (and inform people) what the performance cost of using decimal rather than binary floats and/or of tracking the uncertainty is, and we can start to identify bottlenecks.