cuda-api-wrappers
Thin C++-flavored wrappers for the CUDA Runtime API
Public Attributes | Static Public Attributes | List of all members
cuda::link::options_t Struct Reference
Collaboration diagram for cuda::link::options_t:
Collaboration graph
[legend]

Public Attributes

register_index_t max_num_registers_per_thread { no_max_registers_limit }
 Limit the number of registers which a kernel thread may use. More...
 
grid::block_dimension_t min_num_threads_per_block { no_min_num_threads_per_block }
 The minimum number of threads per block which the compiler should target. More...
 
span< char > info_log
 
span< char > error_log
 
optimization_level_t optimization_level { dont_set_optimization_level }
 Compilation optimization level (as in -O1, -O2 etc.) More...
 
struct {
   bool   obtain_from_cuda_context { true }
 
   bool   use_specific { true }
 
   device::compute_capability_t   specific
 
target
 
bool specify_fallback_strategy { false }
 
fallback_strategy_t fallback_strategy { prefer_ptx }
 
bool generate_debug_information { false }
 Whether or not to generate indications of which PTX/SASS instructions correspond to which lines of the source code, within the compiled output (-lineinfo)
 
bool generate_source_line_number_information { false }
 Whether or not to generate debug information within the compiled output (-g)
 
bool verbose_log
 
bool specify_default_load_caching_mode { false }
 
caching_mode_t default_load_caching_mode
 Specifies which of the PTX load caching modes use by default, when no caching mode is specified in a PTX instruction (-dlcm)
 

Static Public Attributes

static constexpr const register_index_t no_max_registers_limit { 0 }
 
static constexpr const register_index_t no_min_num_threads_per_block { 0 }
 
static constexpr const optimization_level_t dont_set_optimization_level { maximum_optimization_level + 1 }
 

Member Data Documentation

◆ fallback_strategy

fallback_strategy_t cuda::link::options_t::fallback_strategy { prefer_ptx }
Todo:
Use an optional.

◆ max_num_registers_per_thread

register_index_t cuda::link::options_t::max_num_registers_per_thread { no_max_registers_limit }

Limit the number of registers which a kernel thread may use.

Todo:
Use an optional.

◆ min_num_threads_per_block

grid::block_dimension_t cuda::link::options_t::min_num_threads_per_block { no_min_num_threads_per_block }

The minimum number of threads per block which the compiler should target.

Note
can't be combined with a value for the target property.
Todo:
Use an optional.

◆ optimization_level

optimization_level_t cuda::link::options_t::optimization_level { dont_set_optimization_level }

Compilation optimization level (as in -O1, -O2 etc.)

Todo:
Use an optional.

◆ target

struct { ... } cuda::link::options_t::target
Todo:
Use a variant or optional+variant.

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