This is the mail archive of the
java@gcc.gnu.org
mailing list for the Java project.
Re: Synchronization cost regression
- To: java at gcc dot gnu dot org
- Subject: Re: Synchronization cost regression
- From: Loren James Rittle <rittle at latour dot rsch dot comm dot mot dot com>
- Date: Wed, 1 Aug 2001 17:33:10 -0500 (CDT)
- Cc: hans_boehm at hp dot com
- Organization: Networks and Infrastructure Lab (IL02/2240), Motorola Labs
In article <140D21516EC2D3119EE7009027876644049B5F8A@hplex1.hpl.hp.com>
you write:
> I noticed what apppears to be a regression in the synchronization cost with
> the the trunk. [...]
>
> Does it make sense that this would have changed?
>
> What's the proper syntax for
>
> __inline_and_I_really_really_mean_it ?
Does the observed change in speed correspond to around July 20?
2001-07-20 Daniel Berlin <dan@cgsoftware.com>
* params.def: Change default max inline insns to 100.
That parameter had been set at 10000. To see if it was that change in
default settings, try: g++ --param max-inline-insns\=10000
I'm sure Daniel would like to hear what setting is best for your
example code. There was some discussion that 100 was too low. It has
been seen to hurt STL code badly.
Regards,
Loren
--
Loren J. Rittle
Senior Staff Software Engineer, Distributed Object Technology Lab
Networks and Infrastructure Research Lab (IL02/2240), Motorola Labs
rittle@rsch.comm.mot.com, KeyID: 2048/ADCE34A5, FDC0292446937F2A240BC07D42763672