simple PRNG?

[Noob]

Re: simple PRNG?

George Marsaglia wrote:

[snip PRNG code and insightful comments]

It's great to have you post to this group again.

Regards.

[Chris Torek]

Re: simple PRNG?

In article <rtGdnf5giO88in janZ2dnUVZ_tSkn Z2d@comcast.com >
George Marsaglia <geo@stat.fsu.e duwrote:

>/* initialize with any 32-bit seed x and any 32-bit y not 0 */
>
>static unsigned long x=2282008, y=362436069;
>
>#define sK ( x=69069*x+123, y^=y<<13, y^=y>>17, y^=y<<5, x+y )

[snippage]

>Use of sK in any expression will produce a random 32-bit integer,
>(unsigned), so that, for example, u=sK*2.32830643 7e-10 will
>produce a uniform real u in the interval [0,1) ...

If "unsigned long" is 64 bits (and it is on a number of systems),
this will produce a 64-bit result instead of a 32-bit result. I
think[%] -- but have not attempted to prove -- that it is OK simply
to reduce the result here mod 2^32 with:

#define sK \
((x=69069*x+123 , y^=y<<13, y^=y>>17, y^=y<<5, x+y) & 0xffffffffUL)

The mask (against 0xffffffffUL) should be optimized out by any good
compiler in which "unsigned long" is already 32 bits. It is
necessary if you expect the result to be limited to no more than
32 bits, though.

[% I am not sure whether extra retained bits in x and/or y are a
problem. If they are, add more masks with 0xffffffffUL. Again,
the generated code should be the same wherever the masks are a
no-op, as on any machine with 32-bit "unsigned long". Alternatively,
one could use "uint32_t" on C99 systems, but masking is completely
portable, even to C89-only systems.]
--
In-Real-Life: Chris Torek, Wind River Systems
Salt Lake City, UT, USA (40°39.22'N, 111°50.29'W) +1 801 277 2603
email: gmail (figure it out) http://web.torek.net/torek/index.html

[Thad Smith]

Re: simple PRNG?

Chris Torek wrote:

In article <rtGdnf5giO88in janZ2dnUVZ_tSkn Z2d@comcast.com >
George Marsaglia <geo@stat.fsu.e duwrote:

>/* initialize with any 32-bit seed x and any 32-bit y not 0 */
>>
>static unsigned long x=2282008, y=362436069;
>>
>#define sK ( x=69069*x+123, y^=y<<13, y^=y>>17, y^=y<<5, x+y )

[snippage]

>Use of sK in any expression will produce a random 32-bit integer,
>(unsigned), so that, for example, u=sK*2.32830643 7e-10 will
>produce a uniform real u in the interval [0,1) ...

>
If "unsigned long" is 64 bits (and it is on a number of systems),
this will produce a 64-bit result instead of a 32-bit result. I
think[%] -- but have not attempted to prove -- that it is OK simply
to reduce the result here mod 2^32 with:
>
#define sK \
((x=69069*x+123 , y^=y<<13, y^=y>>17, y^=y<<5, x+y) & 0xffffffffUL)

The calculation of x should be OK, but y^=y>>17 would yield a different
value. Try

#define sK \
(x=69069*x+123, y=(y^y<<13) & 0xffffffffUL, y^=y>>17, \
y^=y<<5, (x+y) & 0xffffffffUL)

[% I am not sure whether extra retained bits in x and/or y are a
problem. If they are, add more masks with 0xffffffffUL. Again,
the generated code should be the same wherever the masks are a
no-op, as on any machine with 32-bit "unsigned long". Alternatively,
one could use "uint32_t" on C99 systems, but masking is completely
portable, even to C89-only systems.]

Also, since uint32_t is not guaranteed to exist the masked approach is more
general.

--
Thad

[pete]

Re: simple PRNG?

Walter Roberson wrote:

>
In article <fs057r$7rd$1@a ioe.org>,
jacob navia <jacob@nospam.o rgwrote:
>

unsigned long random(void)
{
register long x, hi, lo, t;

/*
* Compute x[n + 1] = (7^5 * x[n]) mod (2^31 - 1).
* From "Random number generators:
good ones are hard to find",
* Park and Miller, Communications of the ACM,
vol. 31, no. 10,
* October 1988, p. 1195.
*/
x = _randseed;
hi = x / 127773;
lo = x % 127773;
t = 16807 * lo - 2836 * hi;
if (t <= 0)
t += 0x7fffffff;
_randseed = t;
return (t);
}

>
It would be safer (or at least clearer) if most of your constants
were suffixed with L to avoid their being interpreted (if only
mentally) as overflowing the range of UINT_MAX on 16 bit systems.

Why are there any signed type objects at all
in that function definition?

register long unsigned fixes everything.

But I wouldn't publish purportedly portable code
with the keyword "register" in it, either.

--
pete

[CBFalconer]

Re: simple PRNG?

pete wrote:

>

.... snip ...

>
Why are there any signed type objects at all
in that function definition?
>
register long unsigned fixes everything.
>
But I wouldn't publish purportedly portable code
with the keyword "register" in it, either.

Why not? register is still a reserved word. All it is guaranteed
to do is prevent taking the address. Although it seems totally
ridiculous here.

--
[mail]: Chuck F (cbfalconer at maineline dot net)
[page]: <http://cbfalconer.home .att.net>
Try the download section.



--
Posted via a free Usenet account from http://www.teranews.com

[pete]

Re: simple PRNG?

CBFalconer wrote:

>
pete wrote:

... snip ...

Why are there any signed type objects at all
in that function definition?

register long unsigned fixes everything.

But I wouldn't publish purportedly portable code
with the keyword "register" in it, either.

>
Why not?

Because, ...

register is still a reserved word. All it is guaranteed
to do is prevent taking the address. Although it seems totally
ridiculous here.

.... I can't imagine that there ever might be
a situation where I would want to prevent taking the address.

--
pete

[Walter Roberson]

Re: simple PRNG?

In article <47EDA13E.4777@ mindspring.com> ,
pete <pfiland@mindsp ring.comwrote:

>CBFalconer wrote:

>register is still a reserved word. All it is guaranteed
>to do is prevent taking the address. Although it seems totally
>ridiculous here.

>... I can't imagine that there ever might be
>a situation where I would want to prevent taking the address.

It becomes an optimization hint to a compiler: if there are
pointers in the code, the optimizer can know that none of them
point to the object declared as 'register', which fact might
allow it to use tighter code in some cases.
--
"Eightly percent of the people in the world are fools and the
rest of us are in danger of contamination." -- Walter Matthau

[Keith Thompson]

Re: simple PRNG?

roberson@ibd.nr c-cnrc.gc.ca (Walter Roberson) writes:

In article <47EDA13E.4777@ mindspring.com> ,
pete <pfiland@mindsp ring.comwrote:

>>CBFalconer wrote:

>

>>register is still a reserved word. All it is guaranteed
>>to do is prevent taking the address. Although it seems totally
>>ridiculous here.

>

>>... I can't imagine that there ever might be
>>a situation where I would want to prevent taking the address.

>
It becomes an optimization hint to a compiler: if there are
pointers in the code, the optimizer can know that none of them
point to the object declared as 'register', which fact might
allow it to use tighter code in some cases.

On the other hand, since "register" can only be applied an object
declared inside a function, an optimizing compiler probably already
knows that the object's address is never taken.

On the other other hand, the code in question was published in 1988;
at the time, "register" might well have been useful as an optimization
hint.

--
Keith Thompson (The_Other_Keit h) <kst-u@mib.org>
Nokia
"We must do something. This is something. Therefore, we must do this."
-- Antony Jay and Jonathan Lynn, "Yes Minister"

[pete]

Re: simple PRNG?

Walter Roberson wrote:

>
In article <47EDA13E.4777@ mindspring.com> ,
pete <pfiland@mindsp ring.comwrote:

CBFalconer wrote:

>

register is still a reserved word. All it is guaranteed
to do is prevent taking the address. Although it seems totally
ridiculous here.

>

... I can't imagine that there ever might be
a situation where I would want to prevent taking the address.

>
It becomes an optimization hint to a compiler: if there are
pointers in the code, the optimizer can know that none of them
point to the object declared as 'register', which fact might
allow it to use tighter code in some cases.

Is that something that you've done for that reason,
or is that an example of something that you can imagine
that I can't?

--
pete

[Richard Tobin]

Re: simple PRNG?

In article <fsk9qp$6i0$1@c anopus.cc.umani toba.ca>,
Walter Roberson <roberson@ibd.n rc-cnrc.gc.cawrote :

>It becomes an optimization hint to a compiler: if there are
>pointers in the code, the optimizer can know that none of them
>point to the object declared as 'register'

I would have thought most compilers would be able to tell whether
the address of a variable is taken, so that this hint doesn't
provide any information they don't already have.

-- Richard
--
:wq

[Richard]

Re: simple PRNG?

roberson@ibd.nr c-cnrc.gc.ca (Walter Roberson) writes:

In article <47EDA13E.4777@ mindspring.com> ,
pete <pfiland@mindsp ring.comwrote:

>>CBFalconer wrote:

>

>>register is still a reserved word. All it is guaranteed
>>to do is prevent taking the address. Although it seems totally
>>ridiculous here.

>

>>... I can't imagine that there ever might be
>>a situation where I would want to prevent taking the address.

>
It becomes an optimization hint to a compiler: if there are
pointers in the code, the optimizer can know that none of them
point to the object declared as 'register', which fact might
allow it to use tighter code in some cases.

Sounds strange.