Path: csiph.com!news.mixmin.net!eternal-september.org!reader01.eternal-september.org!.POSTED!not-for-mail
From: Tim Rentsch <tr.17687@z991.linuxsc.com>
Newsgroups: comp.lang.c
Subject: Re: Compute Unique Numbers in a Set
Date: Sun, 01 Jan 2023 23:53:52 -0800
Organization: A noiseless patient Spider
Lines: 63
Message-ID: <86a631qsmn.fsf@linuxsc.com>
References: <todbqe$16sg$1@gioia.aioe.org> <tot4eh$21qv0$1@news.xmission.com> <totti7$6if$1@gioia.aioe.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Injection-Info: reader01.eternal-september.org; posting-host="60a7a09099a9247382aee0cc66dfd4bb"; logging-data="1798533"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX19hTLSPf1xHVcKcb9H7F28yCFIuicStsIE="
User-Agent: Gnus/5.11 (Gnus v5.11) Emacs/22.4 (gnu/linux)
Cancel-Lock: sha1:CwBFX/+w081Fwjp3uYon9Kp+1rI= sha1:bansTK3OfC/I/ONNpIGHnFkrIh8=
Xref: csiph.com comp.lang.c:168705

jak <nospam@please.ty> writes:

> Il 02/01/2023 00:19, Kenny McCormack ha scritto:
>
>> In article <todbqe$16sg$1@gioia.aioe.org>, Albert wrote:
>>
>>> Is this the best way to generate unique random numbers in a set of
>>> 6 numbers?
>>>
>>>
>>> <******************************************************>
>>
>> Here's my contribution to this thread.
>>
>> Note that this is both gcc- and Linux- specific, as well as being
>> hard-coded to OP's problem specification.  Note also that I took
>> the liberty of assuming that OP actually, really, did want random
>> numbers between 0 and 59;  this would be suitable if, say, OP was
>> looking for a random number of minutes or seconds (and we could be
>> talking either about time intervals or about locations/distances
>> expressed in terms of latitude and/or longitude).
>>
>> --- Cut Here ---
>> #include <stdio.h>
>> #include <stdint.h>
>>
>> int main(void)
>> {
>>      uint64_t num = 0, j = 1;
>>      FILE *fp = fopen("/dev/urandom","r");
>>
>>      while (__builtin_popcount(num) < 6)
>> 	num |= 1 << (fgetc(fp) % 60);
>>
>>      for (int i=0; i<60; i++, j *= 2)
>> 	if (num & j)
>> 	    printf("i = %d, j = %ld\n",i,j);
>>      return 0;
>> }
>> --- Cut Here ---
>>
>> The values printed out of the variable "i" are your random numbers.
>
> HI,

> I state that I have not worried about the randomization algorithm
> chosen by the OP and, so I focused on data storage.  Your storage
> method is similar to mine because both algorithms use a lookup table
> with the difference that mine uses an array of characters while your
> use the bits of an uint64_t.  However I found a few Issues in your
> source code:
>
> The first concerns the use of the "__builtin_popcount" function, this
> function does not work with uint64_t but with unsigned int. [...]
>
> The second issue concerns that your program will print the values
> archived in numerical order and not in extraction order but perhaps
> this does not affect the OP even if reading his code this seemed
> important.

A third problem is that the way the [ 0 .. 59 ] values are
chosen is horribly biased:  some numbers are 25% more likely
than others.