Struct monument::Config

pub struct Config {
    pub num_threads: Option<usize>,
    pub optimisation_passes: Vec<Mutex<Pass>>,
    pub graph_size_limit: usize,
    pub split_by_start_chunk: bool,
    pub queue_limit: usize,
    pub mem_forget_search_data: bool,
}

Configuration parameters for Monument which don’t change which compositions are emitted.

Fields

num_threads: Option<usize>

Number of threads used to generate compositions. If None, this uses the number of physical CPU cores (i.e. ignoring hyper-threading).

optimisation_passes: Vec<Mutex<Pass>>

graph_size_limit: usize

The maximum graph size, in nodes. If a search would produce a graph bigger than this, it is aborted.

split_by_start_chunk: bool

queue_limit: usize

mem_forget_search_data: bool

If true, the data structures used by searches will be leaked using std::mem::forget. This massively improves the termination speed (because all individual allocations don’t need to be freed), but only makes sense for the CLI, where Monument will do exactly one search run before terminating (thus returning the memory to the OS anyway).

Trait Implementations

impl Default for Config

fn default() -> Self

Returns the “default value” for a type.