[Ipopt] MC19 not available

Stefan Vigerske stefan at math.hu-berlin.de
Mon Dec 21 11:07:33 EST 2015


Hi,

MC19 can by loaded at runtime, but it could be that the dynamic loading 
is disabled in the MSVC build.
It looks to me as if that Ipopt library was build without MC19.

But the MSVS project files are not maintained and I cannot really 
comment on what to do.

Stefan

On 12/21/2015 04:58 PM, jonathan.hogg at stfc.ac.uk wrote:
> Hi,
>
> I can't really help with this - its seems likely to be something wrong with Ipopt rather than with HSL. I do wonder if their linearsolverloader supports dynamic loading of MC19? It may be that you need to use a statically linked copy?
>
> Jonathan.
> ________________________________________
> From: 1252685 at tongji.edu.cn [1252685 at tongji.edu.cn]
> Sent: 21 December 2015 15:28
> To: ipopt at list.coin-or.org
> Cc: Hogg, Jonathan (STFC,RAL,SC); Greg Horn
> Subject: MC19 not available
>
> Thank you for your help,and I have learned a lot.
>
> And it seemed that I finally solved the problem.The initial value of variables x[1]~x[15] varies a lot,and some of them can be 1000 times larger than the others.So I decrease some of the variables which are much larger by 10 times,and the iteration count decrease from 385 to 38.
>
> But there comes another problem when I want to set the nlp_scaling_method option to equilibration-based so that it can scale automatically.The output says "Currently cannot do equilibration-based NLP scaling if MC19 is not available" .
> But I'm solving it with ma27,which was compiled into libhsl.dll using D:\Ipopt-3.12.4\Ipopt\MSVisualStudio\v8-ifort\libhsl-no-MA57. And it should have mc19ad.f compiled inside too.
> Will that make mc19 available?If not,can you give me some instructions on how to make mc19 available.
>
> Sorry for my poor english,I hope I have expressed myself clearly.Thanks in advance.
>
> Dongming Jin
>
> _______________________________________________
> Ipopt mailing list
> Ipopt at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/ipopt
>


-- 
http://www.gams.com/~stefan


More information about the Ipopt mailing list