Computational Comparison

[Msimm007]

Computational Comparison: Sequential Implicit vs Stability

We want to perform a computational comparison between the two existing methods

Preliminary parameters to fix:

1. initial time = 0.01
2. No second order extrapolation
3. no implict time der in pl
4. incompressible is false
5. everything is 2d ( for now)
6. don't touch penalty parameters
7. No RT0 and no darcy with gravity
8. always compute errors

Tests to run:

• q1-rt1-bdf1; h/2; iterative
• q1-rt1-bdf1; h/2; direct
• q1-rt1-bdf1; h/4; iterative
• q1-rt1-bdf1; h/4 direct
• q2-rt2-bdf1; h/2; iterative
• q2-rt2-bdf1; h/2; direct
• q2-rt2-bdf1; h/4 ; iterative
• q2-rt2-bdf1; h/4; direct

Values to save:

• average assembling time for pl
• average solving time for pl
• average assembling time for Sa
• average solving time for Sa
• average assembling time for Sv
• average solving time for Sv
• total time simulation (for 100 iterations).

We will be doing this on manufactured tests. To do this, we need to:

• Create vectors to save the assemble and solve time for each primary unknown
• Decide on max iterations because Cappanera just said 100 but not sure on how to do this
• save times to output to a text file to easily read and understand