Choose tools with evidence.
Build a justified shortlist that fits the physics, your computer and software access.
0 of 4 checkpoints complete. Continue at step 1 ↓
STEP 1 / 4
Start with applicability
Select a family, a component and a research question. Then read what each tool models.
Foundations. Decide the problem first. Then pick the tool that fits the problem.
Explore this research question ↗Read the other explanation levels
Applied. Match particle transport, continuum flow or diffusion to the required equations and observables.
Research. Inspect closure assumptions, boundary treatments and validity limits. Cross-domain expertise does not remove the need for a new benchmark.
Checkpoint
Should a familiar code outrank a code that fits the physics?
Show explanation
Familiarity can reduce setup effort, but it cannot make an inapplicable physical model suitable.
STEP 2 / 4
Check access and hardware
Set the Available hardware and Access / setup requirement filters. Open the installation and license links for two candidates.
Foundations. Some tools run on a laptop. Others need a large computer, or permission from the people who make them.
Explore this research question ↗Read the other explanation levels
Applied. Separate a CPU teaching example from the memory and parallel computing needs of a research model.
Research. Record compiler, solver-library and parallel-runtime dependencies; test deterministic settings where possible and report stochastic seeds where relevant.
Checkpoint
Does a profile or a catalog link grant access to software?
Show explanation
No. Licenses, downloads and any access approval are controlled by the software provider.
STEP 3 / 4
Compare the evidence
Select Compare on two tools. Read the score explanation. Open one primary source for each tool.
Foundations. Use the ratings to decide where to look next. They are not test results.
Explore this research question ↗Read the other explanation levels
Applied. Compare the evidence behind each rating as well as the number. Missing documentation is an uncertainty to investigate.
Research. Require verification and validation results tied to version, observable and regime. A benchmark in a different regime may be weak evidence for your use.
Checkpoint
Are provisional scores performance measurements?
Show explanation
No. They organize discovery evidence. They are not measured solver accuracy or speed.
STEP 4 / 4
Write a selection note
Write a selection note: the question, the chosen tool, the rejected tools and why, access, hardware, and the first lab you will run.
Foundations. Start with a problem small enough that you can say what the answer should be.
Explore this research question ↗Read the other explanation levels
Applied. Use an analytical solution or trusted reference to check the numerical setup before adding complexity.
Research. Define acceptance thresholds before running. Explain which uncertainties the benchmark can and cannot constrain.
Checkpoint
What turns a shortlist into a research starting point?
Show explanation
A testable next step with explicit inputs, acceptance criteria and missing evidence.