When doing jacobians of nonlinear systems, the inames used for
a given sum factorization kernel output depend on the accumulation
context.

Otherwise debugging of nonlinear problems becomes hard.

We havent got a nonlinear sumfact test case so far, so this one did not break.
test coming soon.

It spares us some hassle with function manglers.

Feature/stokes application cherrypick

See merge request !182

By using loopys group mechanism. Each sum factorization kernel
defines a group that conflicts with all other sum factorization
groups.

Conflicts:
	python/dune/perftool/sumfact/realization.py
	python/dune/perftool/sumfact/vectorization.py

Feature/fastdg systems

Closes #93

See merge request !173

Do not generate code for stage 1 sumfact kernels that don't get used

Closes #100

See merge request !181

Save all stage 1 sum factorization kernels that are used in
accumulation expression in the cache during the dry run. Discard all
inactive sum factorization kernels in decide_vetorization_strategy.

Fix sympy simplification and do it by default

See merge request !180

Ignore hard-to-fix warnings in local operators

Closes #95

See merge request !179

We recently learned that usage of sub_elements in UFL is ambiguous...

otherwise my paraview throws me segfault.

It will be necessary to correctly do fastdg for systems