#[repr(C)]pub struct OrtCUDAProviderOptions {
pub device_id: c_int,
pub cudnn_conv_algo_search: OrtCudnnConvAlgoSearch,
pub gpu_mem_limit: usize,
pub arena_extend_strategy: c_int,
pub do_copy_in_default_stream: c_int,
pub has_user_compute_stream: c_int,
pub user_compute_stream: *mut c_void,
pub default_memory_arena_cfg: *mut OrtArenaCfg,
pub tunable_op_enable: c_int,
pub tunable_op_tuning_enable: c_int,
}
Expand description
\brief CUDA Provider Options
\see OrtApi::SessionOptionsAppendExecutionProvider_CUDA
Fields§
§device_id: c_int
\brief CUDA device Id Defaults to 0.
cudnn_conv_algo_search: OrtCudnnConvAlgoSearch
\brief CUDA Convolution algorithm search configuration. See enum OrtCudnnConvAlgoSearch for more details. Defaults to OrtCudnnConvAlgoSearchExhaustive.
gpu_mem_limit: usize
\brief CUDA memory limit (To use all possible memory pass in maximum size_t) Defaults to SIZE_MAX. \note If a ::OrtArenaCfg has been applied, it will override this field
arena_extend_strategy: c_int
\brief Strategy used to grow the memory arena
0 = kNextPowerOfTwo
1 = kSameAsRequested
Defaults to 0.
\note If a ::OrtArenaCfg has been applied, it will override this field
do_copy_in_default_stream: c_int
\brief Flag indicating if copying needs to take place on the same stream as the compute stream in the CUDA EP 0 = Use separate streams for copying and compute. 1 = Use the same stream for copying and compute. Defaults to 1. WARNING: Setting this to 0 may result in data races for some models. Please see issue #4829 for more details.
has_user_compute_stream: c_int
\brief Flag indicating if there is a user provided compute stream Defaults to 0.
user_compute_stream: *mut c_void
\brief User provided compute stream.
If provided, please set has_user_compute_stream
to 1.
default_memory_arena_cfg: *mut OrtArenaCfg
\brief CUDA memory arena configuration parameters
tunable_op_enable: c_int
\brief Enable TunableOp for using. Set it to 1/0 to enable/disable TunableOp. Otherwise, it is disabled by default. This option can be overriden by environment variable ORT_CUDA_TUNABLE_OP_ENABLE.
tunable_op_tuning_enable: c_int
\brief Enable TunableOp for tuning. Set it to 1/0 to enable/disable TunableOp tuning. Otherwise, it is disabled by default. This option can be overriden by environment variable ORT_CUDA_TUNABLE_OP_TUNING_ENABLE.
Trait Implementations§
source§impl Clone for OrtCUDAProviderOptions
impl Clone for OrtCUDAProviderOptions
source§fn clone(&self) -> OrtCUDAProviderOptions
fn clone(&self) -> OrtCUDAProviderOptions
1.0.0 · source§fn clone_from(&mut self, source: &Self)
fn clone_from(&mut self, source: &Self)
source
. Read more