Charm++ parallel programming system

[Nikhil Jain <nikhil AT illinois.edu>]

Hello Everyone, 

I have tagged the second release candidate for the next stable release, version 6.7.0, of Charm++. The git tag is v6.7.0-rc2; this version has passed all of the usual testing in autobuild, and has fixed the issues reported with the first release candidate.

Please take out some time to test this version for your code, and report any bugs/regressions. 

Here is a list of significant changes that have been made since RC1.

1) Bug fix to handle CUDA threads correctly at exit.

2) Bug fix in the recovery code on a node failure.

3) Missing MPI-2.0 API additions to AMPI.

4) Add flags to enable compilation of Charm++ on newer Cray compilers with C++11 support.

5) Scalable startup is now the default behavior when launching a job using chamrun.

6) Bug fixes in AMPI functions - MPI_Comm_create, MPI_Testall.

7) Added support for ++mpiexec with poe.

8) Disable ASLR on Darwin builds.

9) Add new API related to migration in AMPI.

Here is a concise list of changes that were included in the first release candidate (excluding the numerous bug fixes):

1) New API for efficient formula-based distributed spare array creation.

2) CkLoop is now built by default.

3) CommLib has been deleted.

4) PXSHM auto detects the node size. +nodesize is deprecated.

5) Out-of-tree build is now supported.

6) New target: multicore-linux-arm7

7) CmiBool has been dropped in favor of C++'s bool.

8) CBase_Foo::pup need not be called from Foo::pup.

Thanks

On behalf of PPL,

Nikhil

Nikhil Jain, nikhil.jain AT acm.org, http://charm.cs.uiuc.edu/people/nikhil
Doctoral Candidate @ CS, UIUC