chicken-hackers
[Top][All Lists]
Advanced

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

Re: [Chicken-hackers] PATCH: more efficient scheduler


From: Jörg F . Wittenberger
Subject: Re: [Chicken-hackers] PATCH: more efficient scheduler
Date: Mon, 15 Feb 2016 15:26:27 +0100
User-agent: Mozilla/5.0 (X11; Linux armv7l; rv:38.0) Gecko/20100101 Icedove/38.4.0

Hi Peter,

I completely agree on both: master essentially ready for a new release
and this change should not go into it.

Giving this change more time might be helpful.  I might want to amend it
if I find why at the moment master's performance is a crazy mix of
faster and slower than my modified 4.9.1 on my code.  Raw execution
seems often almost twice as fast while over all there is a loss of 30%
on the most usual work load.  Like 4.8 seconds for what took 3.3 before.
 By now I don't know where this comes from.

So better we prepare mentally for yet another release for chicken 4
since moving everything now to chicken 5 while I have not completed to
move to master looks like too much a change at once for me to handle.

Best

/Jörg


Am 15.02.2016 um 14:10 schrieb Peter Bex:
> On Mon, Feb 15, 2016 at 01:57:52PM +0100, Jörg F. Wittenberger wrote:
>> Hi,
>>
>> the attached patch does not fix any bugs.  It brings some performance
>> improvements under certain load conditions.  The cost: It adds the
>> overhead of another counter and a dispatch on a a fixnum modulo
>> operation - IMHO negotiable.
> 
> Hi Joerg,
> 
> This change looks interesting, but the patch flow has increased like
> crazy these last couple of weeks.  I think it's a good idea to make
> a release before we even consider making large/dangerous changes to
> essential parts like the scheduler.  There's only one remaining open
> ticket for a release right now, which is fixing compiler warnings.
> 
> But maybe we can do without that fix, as it can potentially break eggs.
> 
> The master branch is currently in a state that's acceptable for a new
> release, I think.
> 
> Cheers,
> Peter
> 




reply via email to

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