[petsc-dev] TSComputeRHSJacobianConstant issue

[Jed Brown]

Barry Smith <bsmith at mcs.anl.gov> writes:

>> On Jun 28, 2016, at 9:30 AM, Jed Brown <jed at jedbrown.org> wrote:
>> 
>> Thanks, Barry.  My blunder with the tests.  Should we still try to do the optimization?
>
>    I don't know, but it is creating a whole bunch of matrices which I didn't understand since with only RHS Jacobian it should be able to just use the one matrix. 
>
>    So is it ending up with 3 matrices in memory requiring 3 times the memory needs? Definitely a bad thing!  
>
>    Perhaps you can debug why my optimization broke the reuse ability and fix the reuse issue (if it is fixable) with my optimization? I don't have time to do it now.

I'm not confident enough to do it now and I'm about to go offline in
India for a few weeks.  Hopefully this will help me get back to it.

https://bitbucket.org/petsc/petsc/issues/135/memory-optimization-for-ts
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 818 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20160628/47e73485/attachment.sig>