Behaviour discrepancy between the REPL and the testing environment round 2: second reduction boogaloo

This is a sequel to #339 that now transcends the boundaries of loop unrolling, depending only on the configuration of the kernel launch. The setup is more or less the same and can be found in [this updated repository](https://github.com/Hamiltonian-Action/mysterious_magic). Unlike last time around, this issue is observed even in the latest development branch (b725007 at the time of writing).

## Screenshots

### REPL

### Testing

Use the updated repository linked in the issue to reproduce the discrepancy on development commit b725007. Compare the REPL and testing environments, focusing on how kernel-launch configuration affects the reduction behavior. Done means the differing behavior is reproduced and resolved so both environments agree, with a regression test if the repository provides a suitable testing entry point.

Written by the indexing model from the issue text.