Struct CacheAwareConfig 

pub struct CacheAwareConfig {
    pub l1_cache_size: usize,
    pub l2_cache_size: usize,
    pub l3_cache_size: usize,
    pub element_size: usize,
}

Cache topology description used to derive blocking parameters.

All sizes are in bytes. The defaults (L1 = 32 KiB, L2 = 256 KiB, L3 = 8 MiB) are representative of modern x86-64 server CPUs and are used whenever hardware detection is not available.

Fields

l1_cache_size: usize

L1 data-cache size in bytes (default 32 KiB)

l2_cache_size: usize

L2 unified-cache size in bytes (default 256 KiB)

l3_cache_size: usize

L3 shared-cache size in bytes (default 8 MiB)

element_size: usize

Size of a single element in bytes (default 8 for f64)

Implementations

impl CacheAwareConfig

pub fn new() -> Self

Construct with well-known default cache sizes and element_size = 8.

pub fn detect() -> Self

Attempt to detect L1/L2/L3 sizes from the host hardware.

On Linux the kernel exposes per-cache information under /sys/devices/system/cpu/cpu0/cache/index*/size. On macOS the same data is available through sysctl. If detection fails for any reason the function silently returns the same defaults as CacheAwareConfig::new.

pub fn tile_size_for_matmul(&self, n: usize) -> usize

Compute the optimal square tile edge length for a matrix-multiply blocking scheme so that three tiles fit simultaneously in the L2 cache.

n is the largest matrix dimension; the returned tile size is clamped to [4, n].

pub fn block_size_for_scan(&self) -> usize

Compute the block size for a sequential scan so that one block fits comfortably in the L1 data cache.

Trait Implementations

impl Clone for CacheAwareConfig

fn clone(&self) -> CacheAwareConfig

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for CacheAwareConfig

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for CacheAwareConfig

fn default() -> Self

Returns the “default value” for a type.

impl PartialEq for CacheAwareConfig

fn eq(&self, other: &CacheAwareConfig) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Eq for CacheAwareConfig

impl StructuralPartialEq for CacheAwareConfig