Add floating error correction algorythm

[BusinessDuck]

Instead of BigInt and other heavy libraries that other indicators using to solve floating error, we need to use https://docs.oracle.com/cd/E19957-01/806-3568/ncg_goldberg.html Goldberg correction algorithm

[BusinessDuck]

https://github.com/guipn/sinful.js/blob/master/sinful.js an exmple of multiplication and other math operations

[BusinessDuck]

Benchmarks:

https://medium.com/leaningtech/a-fast-bigint-js-in-an-evening-compiling-c-to-javascript-db61ae733512

Native BigInt also looks good

https://www.measurethat.net/Benchmarks/Show/10997/0/js-big-integers-v2#latest_results_block