[Date Index] [Thread Index] [Author Index]

Re: problem with RandomInteger

• To: mathgroup at smc.vnet.net
• Subject: [mg112221] Re: problem with RandomInteger
• From: Mark Adler <madler at alumni.caltech.edu>
• Date: Sun, 5 Sep 2010 05:28:00 -0400 (EDT)
• References: <i5qhhg$pkm$1@smc.vnet.net> <i5sues$ph0$1@smc.vnet.net>

On 2010-09-04 01:02:36 -0700, Albert Retey said:
> I think this is because Table does compile if it thinks that will
> improve speed.

That definitely explains the abrupt change in behavior at 250:

SystemOptions[CompileOptions -> TableCompileLength]
{"CompileOptions" -> {"TableCompileLength" -> 250}}

Though when I try to compile either the function or the function in
Table, it all works fine.  So it seems that when Table compiles, it is
using some internal specialized version of Compile that has the bug.

Mark