Struct DeviceOccupancyInfo 

pub struct DeviceOccupancyInfo {
    pub sm_count: u32,
    pub max_threads_per_sm: u32,
    pub max_blocks_per_sm: u32,
    pub max_registers_per_sm: u32,
    pub max_shared_memory_per_sm: u32,
    pub warp_size: u32,
}

Hardware parameters needed for CPU-side occupancy estimation.

Fields

sm_count: u32

Number of streaming multiprocessors on the device.

max_threads_per_sm: u32

Maximum resident threads per SM.

max_blocks_per_sm: u32

Maximum concurrent blocks per SM.

max_registers_per_sm: u32

Total 32-bit registers available per SM.

max_shared_memory_per_sm: u32

Shared memory capacity per SM in bytes.

warp_size: u32

Threads per warp (typically 32).

Implementations

impl DeviceOccupancyInfo

pub fn for_compute_capability(sm_major: u32, sm_minor: u32) -> Self

Return synthetic DeviceOccupancyInfo for a given SM compute capability, enabling CPU-side occupancy analysis without a live GPU.

Covers all major NVIDIA GPU architectures from Turing through Blackwell. Unknown architectures fall back to Ampere SM 8.6 defaults.

SM capability table

Architecture	sm_major	sm_minor	SMs	Threads/SM	Smem/SM
Turing	7	5	68	1024	65536
Ampere A100	8	0	108	2048	167936
Ampere GA10x	8	6	84	1536	102400
Ada Lovelace	8	9	76	1536	101376
Hopper H100	9	0	132	2048	232448
Blackwell B100	10	0	132	2048	262144
Blackwell B200	12	0	148	2048	262144

Trait Implementations

impl Clone for DeviceOccupancyInfo

fn clone(&self) -> DeviceOccupancyInfo

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for DeviceOccupancyInfo

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

Formats the value using the given formatter.

impl Hash for DeviceOccupancyInfo

fn hash<__H: Hasher>(&self, state: &mut __H)

Feeds this value into the given Hasher.

fn hash_slice<H>(data: &[Self], state: &mut H)

where H: Hasher, Self: Sized,

Feeds a slice of this type into the given Hasher.

impl PartialEq for DeviceOccupancyInfo

fn eq(&self, other: &DeviceOccupancyInfo) -> 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 Copy for DeviceOccupancyInfo

impl Eq for DeviceOccupancyInfo

impl StructuralPartialEq for DeviceOccupancyInfo