warp.fem.cache.dynamic_kernel

warp.fem.cache.dynamic_kernel(suffix, kernel_options=None, allow_overloads=False)

Return a decorator that caches a specialized warp.Kernel.

Parameters:

• suffix (Any) – Hashable key used to specialize and cache the kernel.

• kernel_options (dict[str, Any] | None) – Optional kernel options merged into the generated module.

• allow_overloads – Whether to include argument types in the cache key.

Returns:

A decorator that registers and returns a cached warp.Kernel.