The version number of GSL: 2.1+dfsg-2
The hardware and operating system: Ubuntu 16.04.4 LTS The compiler used, including version number and compilation options: gcc 5.4.0 A description of the bug behavior: For a given mu, the average of the values returned by gsl_ran_exponential does not match the given mu. A short program which exercises the bug: Attached is a sample program which takes as input the number of iterations and the average to achieve. The program compares the average sleep time returned by the Exponential distribution (gsl_ran_exponential) and the Poisson distribution (gsl_ran_poisson). Some conversion has been applied in the code to keep the values as multiples of 10 Sample output with randomly chosen mean = 117 and different number of iterations is below: abhishek@ubuntu:~$ ./rand 10 117 (./rand <number of iterations> <mean to achieve>) max_poi = [130], min_poi = [90] avg_poi = [112.000] max_exp = [970], min_exp = [20] avg_exp = [226.000] abhishek@ubuntu:~$ ./rand 20 117 max_poi = [130], min_poi = [90] avg_poi = [112.000] max_exp = [970], min_exp = [20] avg_exp = [153.000] abhishek@ubuntu:~$ ./rand 50 117 max_poi = [140], min_poi = [90] avg_poi = [115.000] max_exp = [970], min_exp = [0] avg_exp = [110.000] abhishek@ubuntu:~$ ./rand 500 117 max_poi = [150], min_poi = [80] avg_poi = [117.000] max_exp = [1230], min_exp = [0] avg_exp = [118.000] As seen in the above results, the mean of all the values returned by gsl_ran_exponential (226, 153) is higher than the desired mean of 110 for num iterations les than 50
rand.c
Description: Binary data
