How this instrument works
This page checks a single identity: (a + b) mod m always equals ((a mod m) + (b mod m)) mod m, for any whole numbers a and b and any modulus m. People sometimes call this 'modulo being associative,' though the more precise description is that reduction mod m is compatible with addition — reduce each piece first, add the reduced pieces, and reduce once more, and the answer never changes.
The practical payoff is size control. Reducing a and b mod m before adding keeps every intermediate figure no larger than the modulus itself, which matters enormously once a and b are hundreds of digits long, as they routinely are in cryptographic arithmetic or rolling checksums. Rather than building a full sum that might run to dozens of digits and only then taking one large remainder, the calculation can proceed piece by piece, trading one big division for two small ones.
The equality holds no matter how large the two starting numbers are relative to the modulus — a=100 and b=50 against m=12 land on the same zero difference as the smaller pair 17 and 25 against m=7. The one genuine requirement is that the modulus stays fixed throughout both routes; swap in a different m partway through and the shortcut no longer applies, since the two sides would then be answering different questions.
- Enter the two numbers being combined into the a and b fields.
- Set the shared divisor in Modulus (m) — any positive number works, including non-whole values.
- Compare (a + b) mod m against ((a mod m) + (b mod m)) mod m — the two routes to the same reduced sum.
- Check Difference (should always be 0); anything else signals a data-entry slip, not a broken identity.
Worked example — 17 and 25, modulus 7
Take a=17, b=25, and modulus m=7. The direct route adds first: 17+25=42, and 42 mod 7 is 0, since 7 divides into 42 exactly six times with nothing left over. The piecewise route reduces first: 17 mod 7 is 3 (17 is 2×7 plus 3) and 25 mod 7 is 4 (25 is 3×7 plus 4); adding those reduced pieces gives 3+4=7, and 7 mod 7 is 0 again. Both routes land on check1=0 and check2=0, so the reported difference is 0, confirming the identity for this pair.
The same equality holds however the numbers are chosen: with a=10 and b=10 against m=3, both routes give 2; with a=100 and b=50 against m=12, both give 6. None of these are coincidences — reducing before adding, or adding before reducing, is guaranteed to land on the same final remainder every time, which is exactly why check1 and check2 sit side by side here rather than one being treated as the 'real' answer and the other as a mere shortcut.
Questions
Does modulo really distribute over addition?
Yes — (a + b) mod m always equals ((a mod m) + (b mod m)) mod m for any integers a and b and any positive m. This page's check1 and check2 columns compute the two sides independently; for a=17, b=25, m=7 both come out to 0, and the difference column confirms it with a flat zero, exactly as the identity predicts for every pair of numbers, not just this one.
Why is this sometimes called modulo's associative property?
The name is a loose borrowing rather than a precise match — associativity technically describes grouping within a single operation, like (a+b)+c versus a+(b+c). What is actually being tested here is that reduction mod m commutes with addition: reduce the pieces first or the sum first, and either order lands on the same remainder. The common label has stuck even though 'compatible with addition' describes the property more exactly.
Why reduce the pieces first instead of just adding?
Size. Reducing a and b mod m before adding keeps every number involved no bigger than the modulus, which is the difference between manageable arithmetic and unmanageable arithmetic once a and b run to dozens or hundreds of digits, as they do in cryptographic key operations and rolling hash functions. The full sum never has to exist in memory at its original, unreduced size.
Does the identity still hold for large numbers like 100 and 50?
Yes, and the size of a and b relative to the modulus makes no difference to the outcome. For a=100, b=50, m=12, adding first gives 150, and 150 mod 12 is 6; reducing first gives 100 mod 12 = 4 and 50 mod 12 = 2, and 4+2=6, then 6 mod 12 is still 6 — the same answer both ways, exactly as the identity guarantees regardless of scale.
Can the difference column ever come out nonzero?
Not for genuine whole-number inputs checked against one fixed modulus — the identity is a mathematical certainty rather than an approximation, so a nonzero difference would point to something like the modulus changing between the two calculations, or an input that was not a true integer, rather than an exception to the rule itself.