On Mon, 14 Oct 2013, Vikram Garg wrote:
Perhaps a simple fourth option is to just require that R(u,phi) +
S(u) be separate user-defined QoIs. It seems about as intrusive as
requiring assemble_qoi_remainder function definition.
This is as unintrusive as (3), no new APIs at all. It's never as
inefficient as (3)... it's not as good as (1) or (2) in the mixed
case but it's actually better in the pure-heterogeneous case.
It's more programming work for user code, though, and I like making
life as simple as possible in the user code... but then again, if we
switch to (1), app code written for (4) will be forwards-compatible
with the change.
I'm now leaning towards starting with (4) for testing purposes,
merging the adjoint_dirichlet branch when we're getting correct
results, then probably moving to (1) before we add an example code.
---
Roy
------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60134071&iu=/4140/ostg.clktrk
_______________________________________________
Libmesh-devel mailing list
Libmesh-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libmesh-devel
