| File | What | When |
|---|---|---|
W15_lecture.html
(.qmd source) |
Workshop slides — the session plan, the progress talk (four slides
in five minutes), verification at mid-project (predicted vs observed
rates, the wrong yardstick, Richardson and the GCI, -np 1 =
-np 2, a run-time budget), the code review, what to cut if
you are behind |
in class |
W15_lab.ipynb
(Colab) |
Verification kit: projkit.py (NumPy-only helpers to
copy into your project) and a worked example (nonlinear
reaction–diffusion, Newton) — predict and measure rates, errors against
the interpolant vs the exact expression, a QoI without an exact value,
-np 1 vs -np 2, run-time budget; §8 applies it
to your project |
in class while the others present (≈ 45 min), then on your project |
talk_template/ |
Optional Quarto template for the progress talk
(progress_talk.qmd, four slides;
quarto render progress_talk.qmd) |
before the session |
../projects/index.html |
The project topics A–J — required tasks, topic-specific verification, milestones | always |
../dolfinx_0.11_api_reference.md |
DOLFINx 0.11 API reference sheet — §16x errors
against the exact expression, SNES/KSP reasons, the projkit
helpers — paste it into your LLM |
always |
The session starts with the 10-minute quiz on Week 14
(closed-AI). Then the workshop slides, then the progress talks
in the announced order: 10 minutes per student — a 5-minute
talk, then a 5-minute code review with your
project.ipynb open on the screen (two questions: show
where … and what happens if …).
Upload by the evening before the session: your talk
slides (PDF or HTML) and your current project.ipynb
with outputs (Runtime → Restart session and run
all first).
Weight: progress talk and code review 5 % of the course grade (term project 40 % = proposal 5 + progress 5 + final report and code 20 + final presentation 10).
There is no homework this week. Week 16: final presentations; the final report (Quarto) and code are due one week after the Week 16 session.
Where code can run: projkit.py is NumPy
only — it works in ordinary cells and in %%fenicsx cells.
rdx.py (the example) uses DOLFINx and runs only in
%%fenicsx cells. The lab runs in about one minute
locally.