[m-dev.] benchmarks for finite domain solvers with --generate-trail-ops-inline

[Julien Fischer]

On Thu, 22 Dec 2005, Julien Fischer wrote:

>
> On Fri, 16 Dec 2005, Julien Fischer wrote:
>
> > On a related note, I haven't tried running the above tests with trail usage
> > analysis enabled, so it's also possible there may be some improvement there.
> >
>
> I've done this now.  With -O5 --trans-intermod-opt --analyse-trail-usage we
> get a 6.4% increase in speed and a 0.7% decrease in space in asm_fast.gc.tr on
> swordfish compared to -O5 --trans-intermod-opt --no-analyse-trail-usage.
>

Here are some more detailed benchmarks for trail usage analysis with
the finite domain solvers.

All runs were run on surprise with: -O5 --trans-intermod-opt

hlc.gc.tr (--no-analyse-trail-usage)
Time : 4514
Space: 2344776

hlc.gc.tr (--analyse-trail-usage)
Time : 4262
Space: 2279240

hl.gc.tr (--no-analyse-trail-usage)
Time : 4670
Space: 2352968

hl.gc.tr (--analyse-trail-usage)
Time : 4302
Space: 2275144

asm_fast.gc.tr (--no-analyse-trail-usage)
Time : 4992
Space: 3083660

asm_fast.gc.tr (--analyse-trail-usage)
Time : 4541
Space: 2879372
--------------------------------------------------------------------------
mercury-developers mailing list
Post messages to:       mercury-developers at cs.mu.oz.au
Administrative Queries: owner-mercury-developers at cs.mu.oz.au
Subscriptions:          mercury-developers-request at cs.mu.oz.au
--------------------------------------------------------------------------