SysPrecision controls how many significant digits Equinox uses to convert numbers from the binary floating point format to decimal display format. If this number is set too high, small errors can occur, as demonstrated in the example below.

The valid range for SysPrecision is 1 to 15, or zero to get back to the Equinox default, currently 11.

Note that Equinox number declarations always allow up to 15 significant digits (integral plus decimal), eg number x(12,3). SysPrecision just affects the actual math.

Example

This example shows what can happen when floating point rounding is carried out in too many significant digits for the Microsoft floating point library to resolve correctly.

SysPrecision = 13
print SysPrecision, 1046.13 - 1046.14, 9385 - 9384.99

SysPrecision = 12
print SysPrecision, 1046.13 - 1046.14, 9385 - 9384.99

SysPrecision = 0 | the default
print SysPrecision, 1046.13 - 1046.14, 9385 - 9384.99