of things are the <code class="class">ParameterHandler</code>
and <code class="class">SolutionTransfer</code> classes, the
<code class="class">AssertThrow</code> macro, 1d programs,
+ curved boundaries and different mappings,
and output of more than one variable). This is certainly
something that should be improved, but is rather time
consuming.
</p>
- <li> <p>
- <strong>A better test-suite:</strong>
- We have a small suite of programs (approximately 20 programs
- at present) that can be run automatically for regression
- tests. Some of these are modified real-life applications, so
- they should cover a relevant part of the library, but it
- would be necessary to reach every corner of the library to be
- sure a certain patch does not break something. Also, a complete
- test-suite for the library would be quite valuable for ports
- to new computer systems and compilers.
- </p>
-
- <li> <p>
- <strong>Parallelization:</strong>
- We do not presently think about parallelizing the libraries
- on distributed memory machines, but the present
- implementation of multi-threading on shared memory machines is
- quite promising. We will implement multi-threading for several
- other parts of the library of which we know they are taking
- significant amounts of time in common applications.
+ <li> <p>
+ <strong>A better test-suite:</strong> We have a small
+ suite of programs (approximately 40 programs with roughly
+ 15000 lines of code at present) that can be run automatically
+ for regression tests. Some of these are modified real-life
+ applications, so they should cover a relevant part of the
+ library, but it would be necessary to reach every corner of
+ the library to be sure a certain patch does not break
+ something. Also, a complete test-suite for the library would
+ be quite valuable for ports to new computer systems and
+ compilers.
</p>
<li> <p>