genn-team/genn

AI Summary: The task requires writing unit tests to verify the accurate timing of model simulations across different backends (likely CPU and GPU) within the GeNN (GPU-enhanced Neural Network) framework. These tests should involve creating simple models, running them for a short duration, and checking that relevant timing counters are incremented correctly. Additionally, the task includes updating the existing documentation to include information about model timing.

AI Summary: The task is to remove obsolete methods from the GeNN codebase related to the now-removed OpenCL backend. This involves identifying and deleting the methods `BackendBase::getPointerPrefix` and `BackendBase::getMergedGroupFieldHostTypeName`, verifying their removal doesn't break functionality, and updating any related documentation.

AI Summary: The task requires modifying the GeNN (GPU-enhanced Neuronal Network) codebase to add the `__restrict__` keyword to pointer members within merged group structures. This should improve optimization by informing the compiler that these pointers do not alias. The solution likely involves creating a new virtual function in the `BackendBase` class to handle the addition of `__restrict__` specifically for CUDA.