ffc team mailing list archive

[ffc@xxxxxxxxxx (FFC)]

One or more new changesets pushed to the primary ffc repository.
A short summary of the last three changesets is included below.

changeset:   1136:eed380a7e3084a0721e9165bd770575e28952d18
tag:         tip
user:        Anders Logg <logg@xxxxxxxxx>
date:        Sat May 31 23:04:08 2008 +0200
files:       ChangeLog src/ffc/jit/jit.py
description:
Make JIT optimization (-O0 / -O2) optional


changeset:   1135:5716834d07960cd01647d557e6718b410b202713
user:        Anders Logg <logg@xxxxxxxxx>
date:        Thu May 29 13:20:29 2008 +0200
files:       ChangeLog src/ffc/jit/jit.py
description:
Add in-memory cache to speed up JIT compiler for repeated assembly.
This may speed up JIT assembly by seconds for complex forms.
FFC now has two levels of cache: a disk cache and an in-memory cache.


changeset:   1134:b63e7d27e38dd2485f4adf8a26e8e714020ffc61
user:        Anders Logg <logg@xxxxxxxxx>
date:        Wed May 28 22:44:42 2008 +0200
files:       ChangeLog src/demo/SubDomain.form src/demo/SubDomains.form src/ffc/compiler/analysis/formdata.py src/ffc/compiler/codegeneration/tensor/tensorgenerator.py src/ffc/compiler/format/ufcformat.py src/ffc/compiler/language/integral.py src/ffc/compiler/language/reassignment.py src/test/regression/reference/SubDomain.h
description:
Several updates for handling subdomain integrals:
 - Allow simple specification of subdomain integrals.
   Now dx(0), dx(1), ds(0), ds(1) etc work which is equivalent
   to dx0 = Integral("cell", 0) etc.
 - Allow specification of integrals on subdomains without
   needing to define all integrals in the range, which makes
   the following form perfectly fine: M = f*dx(3). Integrals
   for dx(0), dx(1) and dx(2) are automatically generated but
   they are zero.

----------------------------------------------------------------------
For more details, visit http://www.fenics.org/hg/ffc