cuda-api-wrappers
Thin C++-flavored wrappers for the CUDA Runtime API
Public Member Functions | Public Attributes | List of all members
cuda::device::compute_architecture_t Struct Reference

A numeric designator of an architectural generation of CUDA devices. More...

#include <device_properties.hpp>

Public Member Functions

const char * name () const
 
unsigned max_warp_schedulings_per_processor_cycle () const
 
unsigned max_resident_warps_per_processor () const
 
unsigned max_in_flight_threads_per_processor () const
 
memory::shared::size_t max_shared_memory_per_block () const
 
constexpr bool is_valid () const noexcept
 

Public Attributes

unsigned major
 A compute_capability_t has a "major" and a "minor" number, with "major" indicating the architecture; so this struct only has a "major" number.
 

Detailed Description

A numeric designator of an architectural generation of CUDA devices.

Note
See this listing of nVIDIA GPU microarchitectures. Also see compute_capability_t .

Member Function Documentation

◆ max_shared_memory_per_block()

memory::shared::size_t cuda::device::compute_architecture_t::max_shared_memory_per_block ( ) const
Note
On some architectures, the shared memory / L1 balance is configurable, so you might not get the maxima here without making this configuration setting

The documentation for this struct was generated from the following file: