Polyester.jl VS CrossHair

This is a new stiff ODE solver which uses multi-threading within the method to make the solve of single ODEs faster in cases where it's too small to benefit from implicit threading in BLAS. Uses Polyester mixed with a very specific adaptivity method to optimize the work calculation to hit this performance. And of note, it's the type of manual parallel algorithm that cannot use array-based parallelism offered by machine learning libraries, so it's an algorithm you kind of need to implement in C++ or... Julia! Was a fun multi-year project and am really glad to see it completed.

The Folds.jl [1] package mentioned in the article is very nicely written.

For another alternative to Julia's built-in `Threads.@threads` macro, folks may also be interested in checking out `@batch` from Polyester.jl [2] (formerly CheapThreads.jl), which features particularly low-overhead threading.

[1] https://github.com/JuliaFolds/Folds.jl

[2] https://github.com/JuliaSIMD/Polyester.jl

Writing some Python for Hacktoberfest? Try out CrossHair while you do that and get credit for a blog post too! https://github.com/pschanely/CrossHair/issues/173

CrossHair, Hypothesis, and Mutmut for advanced testing.

When you say "formal verification methods", what kind of techniques are you interested in? While using interactive theorem provers will most likely not become very widespread, there are plenty of tools that use formal techniques to give more correctness guarantees. These tools might give some guarantees, but do not guarantee complete functional correctness. WireGuard (VPN tunnel) is I think a very interesting application where they verified the protocol. There are also some tools in use, e.g. Mythril and CrossHair, that focus on detecting bugs using symbolic execution. There's also INFER from Facebook/Meta which tries to verify memory safety automatically. The following GitHub repo might also interest you, it lists some companies that use formal methods: practical-fm

The main difference that Klara bring to the table, compared to similar tool like pynguin and Crosshair is that the analysis is entirely static, meaning that no user code will be executed, and you can easily extend the test generation strategy via plugin loading (e.g. the options arg to the Component object returned from function above is not needed for test coverage).

Just in case you are looking for an alternative approach: if you write contracts in your code, you might also consider crosshair [1] or icontract-hypothesis [2]. If your function/method does not need any pre-conditions then the the type annotations can be directly used.

(I'm one of the authors of icontract-hypothesis.)

[1] https://github.com/pschanely/CrossHair

[2] https://github.com/mristin/icontract-hypothesis

There's a tool for verification of Python programs based on contracts which uses Z3: https://github.com/pschanely/CrossHair

You can use it as part of your CI or during the development (there's even a neat "watch" mode, akin to auto-correct).

Looking at some of your SMT-based projects, I'd love to compare your SMT solver notes with my mine from working on https://github.com/pschanely/CrossHair

Sadly, there aren't a lot of resources on how to use SMT solvers well.