Struct DeviceInfo 

pub struct DeviceInfo {

    pub name: String,
    pub ordinal: i32,
    pub compute_capability: (i32, i32),
    pub total_memory_bytes: usize,
    pub multiprocessor_count: i32,
    pub max_threads_per_block: i32,
    pub max_threads_per_sm: i32,
    pub warp_size: i32,
    pub clock_rate_mhz: f64,
    pub memory_clock_rate_mhz: f64,
    pub memory_bus_width_bits: i32,
    pub l2_cache_bytes: i32,
    pub max_shared_memory_per_block: i32,
    pub max_shared_memory_per_sm: i32,
    pub max_registers_per_block: i32,
    pub ecc_enabled: bool,
    pub tcc_driver: bool,
    pub compute_mode: i32,
    pub supports_cooperative_launch: bool,
    pub supports_managed_memory: bool,
    pub max_persisting_l2_cache_bytes: i32,
    pub async_engine_count: i32,
    pub supports_memory_pools: bool,
    pub supports_gpu_direct_rdma: bool,
    pub supports_cluster_launch: bool,
    pub supports_concurrent_kernels: bool,
    pub supports_unified_addressing: bool,
    pub max_blocks_per_sm: i32,
    pub single_to_double_perf_ratio: i32,
}

Comprehensive device information gathered in a single call.

All fields are populated via convenience methods on Device. Fields that fail to query (e.g. on an older driver) default to 0 / false.

Fields

name: String

Human-readable device name.

ordinal: i32

Zero-based device ordinal.

compute_capability: (i32, i32)

Compute capability (major, minor).

total_memory_bytes: usize

Total device memory in bytes.

multiprocessor_count: i32

Number of streaming multiprocessors.

max_threads_per_block: i32

Maximum threads per block.

max_threads_per_sm: i32

Maximum threads per streaming multiprocessor.

warp_size: i32

Warp size in threads.

clock_rate_mhz: f64

Core clock rate in MHz.

memory_clock_rate_mhz: f64

Memory clock rate in MHz.

memory_bus_width_bits: i32

Memory bus width in bits.

l2_cache_bytes: i32

L2 cache size in bytes.

max_shared_memory_per_block: i32

Maximum shared memory per block in bytes.

max_shared_memory_per_sm: i32

Maximum shared memory per SM in bytes.

max_registers_per_block: i32

Maximum 32-bit registers per block.

ecc_enabled: bool

ECC memory enabled.

tcc_driver: bool

TCC driver mode.

compute_mode: i32

Compute mode (0=default, 1=exclusive-thread, 2=prohibited, 3=exclusive-process).

supports_cooperative_launch: bool

Cooperative kernel launch support.

supports_managed_memory: bool

Managed (unified) memory support.

max_persisting_l2_cache_bytes: i32

Maximum persisting L2 cache size in bytes (Ampere+).

async_engine_count: i32

Number of async copy engines.

supports_memory_pools: bool

Supports memory pools (cudaMallocAsync).

supports_gpu_direct_rdma: bool

Supports GPU Direct RDMA.

supports_cluster_launch: bool

Supports cluster launch (Hopper+).

supports_concurrent_kernels: bool

Concurrent kernel execution supported.

supports_unified_addressing: bool

Unified addressing supported.

max_blocks_per_sm: i32

Maximum blocks per SM.

single_to_double_perf_ratio: i32

Single-to-double precision performance ratio.

Trait Implementations

impl Clone for DeviceInfo

fn clone(&self) -> DeviceInfo

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for DeviceInfo

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

Formats the value using the given formatter.

impl Display for DeviceInfo

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

Formats the value using the given formatter.