chicken-hackers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Chicken-hackers] nursery size


From: felix winkelmann
Subject: Re: [Chicken-hackers] nursery size
Date: Mon, 6 Apr 2009 00:51:50 +0200

On Fri, Apr 3, 2009 at 6:36 AM, Jim Ursetto <address@hidden> wrote:
> I found I got a substantial performance boost on this machine--about
> 35% in one app--by setting the nursery stack size at runtime via
> -:s1024k.  I tried rebuilding the compiler to make this the default,
> using NURSERY=1048576, and rebuilding a few eggs.  Afterward, csi
> confirms a nursery size of 1024k, but no performance boost was
> observed at all.  If I then use -:s1024k on this new compiler, the
> performance boost returns.
>
> So, I assume -:s1024k is doing something more than simply setting the
> default target stack size.  But I can't figure out what from perusing
> the code.  I would like to know if I can make the behavior caused by
> -:s1024k the default, rather than setting it at runtime each time.

Normally, this should work. The logic that sets the nursery size,
depending on the various options and switches is quite a mess.
It is possible that something doesn't get carried through. Does
"-:d" show anything interesting?


cheers,
felix




reply via email to

[Prev in Thread] Current Thread [Next in Thread]