In the article
https://learn.microsoft.com/en-US/office/troubleshoot/excel/floating-point-arithmetic-inaccurate-result discusses calculation errors in Excel related to the subtraction of close floating point numbers. But decimal arithmetic, implemented in Excel based on the IEEE754 standard, does not guarantee correct calculations over the entire range of declared decimal values, even for simple arithmetic operations.
Here are some examples:
Addition
A1 = 39579678828828700000
A2 = 100000000000451000
A1 + A2 = 39679678828829100000
While the correct answer is 39679678828829200000
Multiplication
A1=100000000052251
A2 = 999
A1*A2 = 99900000052198800
While the correct answer is 99900000052198700
Division
A1=100000000052252
A2=193
A1/A2 = 518134715296.643000000
While the correct answer is 518134715296.642
The subtraction operation is the most dangerous in Excel. This applies not only to the subtraction of close numbers. A subtraction of even distant numbers in the sequence of arithmetic operations is likely to result in an error. For example
A1=1234567890.11
A2 = 1230000000
(A1-A2)*999 = 4563322219.88990
While accurate to 15 significant digits, the correct answer is 4563322219.89000
The question arises, is it possible to use Excel for arithmetic calculations?
The solution to this problem in https://www.techrxiv.org/articles/preprint/The\_arithmetic\_of\_binary\_equivalents\_of\_decimal\_numbers/19294511