Suite of nonlinear solvers

Edit Package sundials
https://computing.llnl.gov/projects/sundials

SUNDIALS is a SUite of Non-linear DIfferential/ALgebraic equation Solvers
for use in writing mathematical software.

SUNDIALS was implemented with the goal of providing robust time integrators
and nonlinear solvers that can easily be incorporated into existing simulation
codes. The primary design goals were to require minimal information from the
user, allow users to easily supply their own data structures underneath the
solvers, and allow for easy incorporation of user-supplied linear solvers and
preconditioners.

Refresh
Refresh
Source Files (show unmerged sources)
Filename Size Changed
_constraints 0000000106 106 Bytes
_multibuild 0000000115 115 Bytes
sundials-7.0.0.tar.gz 0091813927 87.6 MB
sundials.changes 0000018904 18.5 KB
sundials.spec 0000013283 13 KB
Latest Revision
buildservice-autocommit accepted request 1158845 from Atri Bhattacharya's avatar Atri Bhattacharya (badshah400) (revision 38)
baserev update by copy to link target
Comments 6

Klaus Kämpf's avatar

This source needs some overhaul. Sources are at https://github.com/LLNL/sundials now and 6.2.0 is the latest upstream release


Atri Bhattacharya's avatar

6.2.0 is a major update, we may want to check if it does not break dependencies down the line. In the meanwhile, I have version 5.8.0 brewing in my home branch (but if 6.2.0 does not create breakages we should push it).

The failure with mvapich2 flavour may be independent of this nonetheless, so we may want to check what is up with that (I am testing this locally), or disable this flavour entirely.


Klaus Kämpf's avatar

From what I've seen, the failure (abort() being called within snprintf()) is coming from the mvapich2 package and not from sundials. Disabling the mvapich2 flavour is certainly a reasonable short-term fix.


Atri Bhattacharya's avatar

@kwk makes sense to me. I have also checked that octave 7.1.0 (seems like the only major dependency) is compatible with sundials 6.x, so if you would consider re-submitting your update after disabling the mvapich2 flavour, I would be very grateful.

Thanks.



Klaus Kämpf's avatar

I fixed the potential buffer overflows in science:HPC/mvapich2 and re-added the mvapich2 flavour.

openSUSE Build Service is sponsored by