Parameters field in Model for user convenience

[glwagner]

This PR adds a field parameters to Model. The field is intended to be utilized by users for setting parameters in user-defined forcing functions and boundary conditions functions.

The forcing functions now have the function signature

F.u(i, j, k, grid, time, U, C, params)

whereas boundary conditions have the function signature

boundary_condition(i, j, grid, time, iteration, U, C, params)

This PR resolves #394 and #390.

[codecov]

Codecov Report

Merging #395 into master will increase coverage by 7.06%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #395      +/-   ##
==========================================
+ Coverage   64.62%   71.69%   +7.06%     
==========================================
  Files          23       23              
  Lines        1405     1413       +8     
==========================================
+ Hits          908     1013     +105     
+ Misses        497      400      -97

Impacted Files	Coverage Δ
src/models.jl	92.85% <100%> (ø)	⬆️
src/boundary_conditions.jl	83.05% <100%> (+1.69%)	⬆️
src/time_steppers.jl	74.48% <100%> (+1.14%)	⬆️
src/output_writers.jl	77.71% <0%> (+0.54%)	⬆️
src/utils.jl	64.44% <0%> (+14.44%)	⬆️
src/fields.jl	62.02% <0%> (+15.18%)	⬆️
src/Oceananigans.jl	100% <0%> (+16.66%)	⬆️
src/poisson_solvers.jl	97.61% <0%> (+57.61%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 079cb53...bd0ce51. Read the comment docs.

[ali-ramadhan]

Performance benchmarks seem fine timing wise (model hasn't slowed down), but there seems to be some performance regression as we now perform tons of little memory allocations somewhere. I'll open an issue but this should be merged as it's not related to this PR.

---

CPU:

 ───────────────────────────────────────────────────────────────────────────────────────────────────────────
                 Forcing function benchmarks                        Time                   Allocations      
                                                            ──────────────────────   ───────────────────────
                      Tot / % measured:                          69.0s / 32.8%           14.0GiB / 55.9%    

 Section                                            ncalls     time   %tot     avg     alloc   %tot      avg
 ───────────────────────────────────────────────────────────────────────────────────────────────────────────
 128×128×128 with forcing (params) (CPU, Float64)       10    7.76s  34.2%   776ms   2.68GiB  34.2%   274MiB
 128×128×128 with forcing (consts) (CPU, Float64)       10    7.63s  33.7%   763ms   2.58GiB  32.9%   264MiB
 128×128×128   no forcing (CPU, Float64)                10    7.27s  32.1%   727ms   2.58GiB  32.9%   264MiB
 ───────────────────────────────────────────────────────────────────────────────────────────────────────────

GPU:

 ───────────────────────────────────────────────────────────────────────────────────────────────────────────
                 Forcing function benchmarks                        Time                   Allocations      
                                                            ──────────────────────   ───────────────────────
                      Tot / % measured:                          69.6s / 0.21%           9.23GiB / 0.30%    

 Section                                            ncalls     time   %tot     avg     alloc   %tot      avg
 ───────────────────────────────────────────────────────────────────────────────────────────────────────────
 128×128×128 with forcing (params) (GPU, Float64)       10   48.9ms  33.7%  4.89ms   9.62MiB  33.5%  0.96MiB
 128×128×128 with forcing (consts) (GPU, Float64)       10   48.1ms  33.2%  4.81ms   9.55MiB  33.3%  0.96MiB
 128×128×128   no forcing (GPU, Float64)                10   47.9ms  33.1%  4.79ms   9.55MiB  33.3%  0.96MiB
 ───────────────────────────────────────────────────────────────────────────────────────────────────────────

[glwagner]

we now perform tons of little memory allocations somewhere.

But only on CPU?