On 2007-02-19, Wolf Behrenhoff <NoSpamPleaseButThisIsValid3@[EMAIL PROTECTED]
> wrote:
> Marco van de Voort schrieb:
>> On 2007-02-19, Wolf Behrenhoff <NoSpamPleaseButThisIsValid3@[EMAIL PROTECTED]
>
wrote:
>>> Richard Engebretson schrieb:
>>>> I am running your test on a linux virtual console using nr:longint.
>>>> Still running after over 125,000 recursions. Don't know ?? 175,000
...
>>>> snorrr :-)
>>> Oh! That's interesting!
>>>
>>> I did the test under Windows XP (str=string[100])
>>>
>>> Turbo Pascal: 140 recursions (depends on the -M setting)
>>> Free Pascal: 3455 recursions (depends on the -Cs setting)
>>>
>>> But 175,000? A difference in three orders of magnitude? Something must
>>> be VERY different!
>>
>> Which FPC compiler ? win32 I hope?
>
> Of course. Version 2.0.2. I didn't update for a long time...
-Cs probably. I manually tried what FPC (or the windows call behind it)
accepts as stack size, and it is about 64MiB. (67million something bytes).
540564 iterations, running 9.6s with or without optimization.
|
