From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from b-painless.mh.aa.net.uk ([81.187.30.52]) by ewsd; Tue Sep 24 05:53:45 EDT 2019
Received: from 132.198.187.81.in-addr.arpa ([81.187.198.132] helo=quintile.net)
	by b-painless.mh.aa.net.uk with esmtp (Exim 4.92)
	(envelope-from <steve@quintile.net>)
	id 1iChVf-00042r-Py
	for 9front@9front.org; Tue, 24 Sep 2019 10:53:27 +0100
Received: from [10.15.173.220] ([185.69.145.33]) by quintile.net; Tue Sep 24 10:53:26 BST 2019
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=utf-8
From: Steve Simon <steve@quintile.net>
Mime-Version: 1.0 (1.0)
Date: Tue, 24 Sep 2019 10:53:25 +0100
Message-Id: <08256730-9364-4F53-A4A4-C5F4CD2A6F8A@quintile.net>
Subject: Re: [9front] Quick PRNG question
References: <d02cb03b-725a-4032-89e7-47262671f46e@www.fastmail.com> <alpine.LNX.2.00.1909241144280.7981@phi>
In-Reply-To: <alpine.LNX.2.00.1909241144280.7981@phi>
To: 9front@9front.org
X-Mailer: iPhone Mail (16G102)
List-ID: <9front.9front.org>
List-Help: <http://lists.9front.org>
X-Glyph: ➈
X-Bullshit: just-in-time-scale software singleton 


i would normally defer to =E2=80=9Cnumerical recipes=E2=80=9D in such matter=
s. if you just want a square distribution then xor and a shift register is n=
ice and simple. the book also has a slower algorithm which reshapes this to a=
 gaussian without much effort.

-Steve


> On 24 Sep 2019, at 10:45 am, Julius Schmidt <aiju@phicode.de> wrote:
>=20
> Look up xorshift and its variants
>=20
> It's about a million times simpler than Mersenne twister and some of the v=
ariants are much better at passing statistical tests.
>=20
>> On Tue, 24 Sep 2019, Ethan Gardener wrote:
>>=20
>> Hi Aiju! I remember some time back you laughed at Mersenne twister PRNG, c=
iting another PRNG as simpler and better.  What was the better PRNG?
>>=20