Skip to contents

Use 32bit float precision in computations

Source: R/rkeops_options.R
rkeops_use_float32.Rd

Set up rkeops options to use 32bit float precision in computation when calling user-defined operators.

Usage

rkeops_use_float32()

Value

None

Details

Note: Default behavior is to use 32bit float precision in computation.

Important: 32bit float precision computations are faster than 64bit float, however the lower precision may have a huge effect on the accuracy of your computation and validity of your results in certain applications.

Since R only manages 64bit float (a.k.a "double") numbers, the input and output are converted to 32bit float before computation and back to 64bit float after computation.

See also

rkeops_use_float64(), set_rkeops_options()

Author

Ghislain Durif

Examples

if (FALSE) {
rkeops_use_float32()
}