paper: minor cosmentic suggestions #52
Conversation
| Such changes in the stability of the fluid flow can be studied by performing numerical simulations with a model and observing its transient behavior after a certain time. | ||
| This is, however, computationally very expensive, and in many cases infeasible. | ||
| Instead, so called continuation methods exist, that are able to trace stable and unstable steady states in parameter space, without having to perform the expensive transient simulations [@dijkstra:05]. | ||
| Instead, so called continuation methods are able to trace stable and unstable steady states in parameter space, obviating expensive transient simulations [@dijkstra:05]. | ||
| The `TransiFlow` Python package implements a continuation framework in which fluid flow problems can be studied with the help of several computational back-ends that can, based on the needs of the user, easily be switched between. |
There was a problem hiding this comment.
I'd add here a paragraph on what the model can actually do (like the first paragraph in the docs) and what its
constraints are. To me, it's not really clear which "fluid flow problems" can be
studied (only later "canonical" problems are mentioned), and what said "needs of the user" can be, which the code caters for. I.e., I'd appreciate an answer to the possible question: "can I use this code for my project?"
See also related comment below
| Therefore, ease of developing and using the parallel software is crucial. | ||
|
|
||
| By abstracting away the computational back-end from the user, the user can develop their model on their own machine (e.g. a laptop) in Python using the SciPy back-end, and once the model works, run a large scale simulation on a supercomputer e.g. using the `Trilinos` back-end, which can use a combination of OpenMP, MPI and potentially GPUs, without requiring any changes to the code. |
There was a problem hiding this comment.
"the user can develop their model" -- what does this mean? what are the building blocks that transiflow provides for developing models that are not already included?
I think this is also something that could be added to the docs.
No description provided.