C Reference ยท stdlib.h
C rand() and srand() functions
Reference for C rand(), srand() and RAND_MAX: how the generator works, seeding, picking numbers in a range without bias, and why rand is not for security.
rand() returns the next number of a pseudo-random sequence, an int from 0 to RAND_MAX. srand() sets the seed, which decides where the sequence starts. The same seed always produces the same sequence on the same implementation, which is handy for reproducible tests and useless as a source of unpredictability.
If you never call srand(), the program behaves as if it had called srand(1).
#include <stdlib.h> int rand(void); void srand(unsigned int seed);
Try it
Press Run (or Ctrl + Enter) to see the result here.
Program input ยท what you'd type while the program runs
Type the answers before you press Run, one per line. Each time the program reads a line (with scanf or fgets), it gets the next line from this box.
Press Run (or Ctrl + Enter) to see the result here.
Program input ยท what you'd type while the program runs
Type the answers before you press Run, one per line. Each time the program reads a line (with scanf or fgets), it gets the next line from this box.
Syntax
int rand(void); void srand(unsigned int seed); #define RAND_MAX /* implementation-defined, at least 32767 */
| Name | Description |
|---|---|
seed |
Starting value for the sequence. Equal seeds give equal sequences. |
Returns: rand() returns an integer in [0, RAND_MAX]. srand() returns nothing.
Errors / undefined behavior: none. Whether rand() is thread-safe is implementation-specific (the standard allows a data race, glibc uses a lock).
Usage notes
- Seed once, near the start of
main. Typical seeds aretime(NULL)or a value read from the operating system. Callingsrand(time(NULL))repeatedly inside a loop restarts the sequence with the same value and yields repeated numbers. - The sequence is platform specific. For seed 7 this runner's C library (glibc) printed
4 4 6 2 6 4for six dice rolls; another library gives other numbers for the same seed. Never hard-code expected random values in code that must run elsewhere. RAND_MAXis 2147483647 with glibc but only 32767 with Microsoft's C library. Code that assumes the large value breaks on Windows.rand() % nis biased whenRAND_MAX + 1is not a multiple ofn: small results occur slightly more often. The rejection loop above removes the bias.- Low bits can be weak in old generators. Prefer scaling (
rand() / (RAND_MAX + 1.0) * n) over the remainder whennis large relative toRAND_MAX. - Not for security. The sequence is predictable from a few outputs. Keys, tokens and passwords need an operating-system source such as
getrandom()or/dev/urandom(POSIX) or a cryptographic library. - Shuffling: use Fisher-Yates with an unbiased
random_below, not repeated random swaps. - Alternatives: POSIX
random()/srandom()has a larger state; for Monte Carlo work write a small generator such as xorshift or PCG so results are identical on every platform.
Related
time,qsort, Integer division- Lessons: Random numbers, Useful stdlib.h functions