qprimes is a fast console program, computing all prime numbers between a specified minimum and maximum value of range
- Build:
gcc main.c -O3 -lm -o qprimes
- Run
qprimes MIN MAX
to generate prime numbers between$[MIN, MAX]$ . -
$MIN, MAX$ can be expressed in decimal form or, if preceded by0x
, in haxadecimal form. - Run
qprimes
without arguments for more options.
$ git clone https://github.com/johsteffens/qprimes
$ cd qprimes
$ gcc main.c -O3 -lm -o qprimes
$ ./qprimes 10000000001099 10000000001199
10000000001141
10000000001161
10000000001177
10000000001191
4 primes between 10000000001099 and 10000000001199
Heap size: 197899 Bytes
With
- Processing time is about
$O( n\ log( log( n ) ) )$ for$r << n$ . - Memory requirement is about
$O( n )$ .
The maximum possible heap memory usage is around 270 MBytes.
Command: qprimes 0xFFFFFFFFFFFFFF00 0xFFFFFFFFFFFFFFFF
This test computes the last few primes below 264. It is the worst case for the given prime window size. Smaller primes will compute faster.
Platform | Time |
---|---|
AMD RyzenTM 9 7950X | 9 seconds |
IntelR CoreTM i7 6700 | 22 seconds |
Raspberry Pi 3 Model B+ | 250 seconds |
Raspberry Pi 2 Model B | 370 seconds |
qprimes uses a combination of sieving and paging.
We begin with the prime definition: An integer q > 1 is prime exactly when no integer p > 1 and p < q divides q.
Since any such p is either prime or composite, it is sufficient to test q by all primes below q.
If any
If a sequence of primes is needed, instead of explicitly testing for divisibility, it is generally much faster to simply cross out all non-primes in an interval by computing multiples of primes gathered so far. This approach is called sieving and we use it to collect all primes up to
Finally, we select an appropriate interval (called page) including the target range
Although the method has its roots in ancient times, it is still considered among the most efficient ways to generate a sequence of prime numbers.
Prime numbers are useful in various disciplines of numerical processing such as LCGs and hash tables.
I experimented especially with cuckoo hashing to develop the specific associative binding and runtime type awareness technique in beth. I prefer using LCGs for algorithm testing in monte carlo simulations.
I wrote this simple prime generator as tool for developing/improving above techniques.
© Johannes B. Steffens