Procedures

Procedure	Location	Procedure Type	Description
add	dtfft_nvrtc_kernel_cache	Subroutine	Adds new entry to cache
add_line	dtfft_nvrtc_kernel_generator	Subroutine	Adds new line to CUDA code
aligned_alloc	dtfft_utils	Interface
alloc_and_set_aux	dtfft_transpose_plan_cuda	Function	Allocates auxiliary memory according to the backend and sets it to the plans
alloc_fft_plans	dtfft_plan	Subroutine	Allocates abstract_executor with required FFT class and populates fft_mapping with similar FFT ids
alloc_mem	dtfft_abstract_transpose_plan	Subroutine	Allocates memory based on `backend`
astring_f2c	dtfft_utils	Subroutine	Convert Fortran string to C allocatable string
autotune_grid	dtfft_transpose_plan_host	Subroutine	Creates cartesian communicator and executes various datatypes on it
autotune_grid	dtfft_transpose_plan_cuda	Subroutine	Creates cartesian grid and runs various backends on it. Can return best backend and execution time
autotune_grid_decomposition	dtfft_transpose_plan_host	Subroutine	Runs through all possible grid decompositions and selects the best one based on the lowest average execution time
autotune_grid_decomposition	dtfft_transpose_plan_cuda	Subroutine	Runs through all possible grid decompositions and selects the best one based on the lowest average execution time
autotune_mpi_datatypes	dtfft_transpose_plan_host	Subroutine
autotune_transpose_id	dtfft_transpose_plan_host	Function	Creates forward and backward transpose plans bases on source and target data distributing, executes them `DTFFT_MEASURE_ITERS` times ( 4 * `DTFFT_MEASURE_ITERS` iterations total ) Read more…
check_aux	dtfft_plan	Subroutine	Checks if aux buffer was passed by user and if not will allocate one internally
check_continuity	dtfft_pencil	Function	Check if the local pencils cover the global space without gaps
check_create_args	dtfft_plan	Function	Check arguments provided by user and sets private variables
check_device_pointers	dtfft_plan	Function	Checks if device pointers are provided by user
check_if_even	dtfft_pencil	Function	Checks if data is evenly distributed across processes
check_overlap	dtfft_pencil	Function	Check if two pencols overlap in ndims-dimensional space
cleanup	dtfft_nvrtc_kernel_cache	Subroutine	Removes unused modules from cuda context
Comm_f2c	dtfft_utils	Interface
compile_and_cache	dtfft_nvrtc_kernel	Function	Compiles kernel stored in `code` and caches pointer to CUfunction
config_constructor	dtfft_config	Function	Creates a new configuration
count_bank_conflicts	dtfft_nvrtc_block_optimizer	Function	Counts bank conflicts for a given tile size, padding, element size, and block rows.
count_unique	dtfft_utils	Function	Count the number of unique elements in the array
create	dtfft_pencil	Subroutine	Creates pencil
create	dtfft_abstract_executor	Function	Creates FFT plan
create	dtfft_executor_fftw_m	Subroutine	Creates FFT plan via FFTW3 Interface
create	dtfft_nvrtc_kernel	Subroutine	Creates kernel
create	dtfft_executor_vkfft_m	Subroutine	Creates FFT plan via vkFFT Interface
create	dtfft_abstract_transpose_plan	Function	Creates transposition plans
create	dtfft_transpose_handle_host	Subroutine	Creates `transpose_handle_host` class
create	dtfft_executor_mkl_m	Subroutine	Creates FFT plan via MKL DFTI Interface
create	dtfft_transpose_handle_cuda	Subroutine	Creates CUDA Transpose Handle
create	dtfft_abstract_backend	Subroutine	Creates Abstract GPU Backend
create	dtfft_executor_cufft_m	Subroutine	Creates FFT plan via cuFFT Interface
create	dtfft_nvrtc_kernel_cache	Subroutine	Creates cache
create	dtfft_backend_cufftmp_m	Subroutine	Creates cuFFTMp GPU Backend
create_1d_comm	dtfft_pencil	Subroutine	Creates a new 1D communicator based on the fixed dimensions of the current pencil
create_c2c	dtfft_plan	Subroutine	C2C Plan Constructor
create_c2c_core	dtfft_plan	Function	Creates plan for both C2C and R2C
create_c2c_internal	dtfft_plan	Function	Private method that combines common logic for C2C plan creation
create_c2c_pencil	dtfft_plan	Subroutine	C2C Plan Constructor
create_cart_comm	dtfft_abstract_transpose_plan	Subroutine	Creates cartesian communicator
create_cuda	dtfft_transpose_plan_cuda	Function	Creates CUDA transpose plan
create_data_handle	dtfft_transpose_handle_cuda	Subroutine	Creates handle
create_device_pointer	dtfft_nvrtc_kernel	Subroutine	Allocates memory on a device and copies `values` to it.
create_handle	dtfft_transpose_handle_host	Subroutine	Creates transposition handle
create_helper	dtfft_abstract_backend	Subroutine	Creates helper
create_helper	dtfft_backend_mpi	Subroutine	Creates MPI helper
create_mpi	dtfft_backend_mpi	Subroutine	Creates MPI backend
create_nccl	dtfft_backend_nccl_m	Subroutine	Creates NCCL backend
create_nvtx_domain	dtfft_interface_nvtx	Subroutine	Creates a new NVTX domain
create_pencil_init	dtfft_pencil	Function	Creates and validates pencil passed by user to plan constructors
create_pencil_t	dtfft_pencil	Function	Creates pencil object, that can be used to create dtFFT plans
create_pencils_and_comm	dtfft_abstract_transpose_plan	Subroutine	Creates cartesian communicator
create_private	dtfft_transpose_plan_host	Function	Creates transposition plans
create_private	dtfft_plan	Function	Creates core
create_r2c	dtfft_plan	Subroutine	R2C Generic Plan Constructor
create_r2c_internal	dtfft_plan	Function	Private method that combines common logic for R2C plan creation
create_r2c_pencil	dtfft_plan	Subroutine	R2C Plan Constructor with pencil
create_r2r	dtfft_plan	Subroutine	R2R Plan Constructor
create_r2r_internal	dtfft_plan	Function	Creates plan for R2R plans
create_r2r_pencil	dtfft_plan	Subroutine	R2R Plan Constructor
create_transpose_2d	dtfft_transpose_handle_host	Subroutine	Creates two-dimensional transposition datatypes
create_transpose_XY	dtfft_transpose_handle_host	Subroutine	Creates three-dimensional X –> Y, Y –> X transposition datatypes
create_transpose_XZ	dtfft_transpose_handle_host	Subroutine	Creates three-dimensional X –> Z transposition datatypes Can only be used with 3D slab decomposition when slabs are distributed in Z direction
create_transpose_YZ	dtfft_transpose_handle_host	Subroutine	Creates three-dimensional Y –> Z, Z –> Y transposition datatypes
create_transpose_ZX	dtfft_transpose_handle_host	Subroutine	Creates three-dimensional Z –> X transposition datatypes Can only be used with 3D slab decomposition when slabs are distributed in Z direction
cudaDeviceSynchronize	dtfft_interface_cuda_runtime	Interface
cudaEventCreate	dtfft_interface_cuda_runtime	Interface
cudaEventCreateWithFlags	dtfft_interface_cuda_runtime	Interface
cudaEventDestroy	dtfft_interface_cuda_runtime	Interface
cudaEventElapsedTime	dtfft_interface_cuda_runtime	Interface
cudaEventRecord	dtfft_interface_cuda_runtime	Interface
cudaEventSynchronize	dtfft_interface_cuda_runtime	Interface
cudaFree	dtfft_interface_cuda_runtime	Interface
cudaGetDevice	dtfft_interface_cuda_runtime	Interface
cudaGetDeviceCount	dtfft_interface_cuda_runtime	Interface
cudaGetErrorString	dtfft_interface_cuda_runtime	Function	Helper function that returns a string describing the given nvrtcResult code If the error code is not recognized, “unrecognized error code” is returned.
cudaGetErrorString_c	dtfft_interface_cuda_runtime	Interface
cudaGetLastError	dtfft_interface_cuda_runtime	Interface
cudaMalloc	dtfft_interface_cuda_runtime	Interface
cudaMemcpy	dtfft_interface_cuda_runtime	Interface	Copies data synchronously between host and device.
cudaMemcpyAsync	dtfft_interface_cuda_runtime	Interface	Copies data asynchronously between host and device.
cudaMemGetInfo	dtfft_interface_cuda_runtime	Interface
cudaMemset	dtfft_interface_cuda_runtime	Interface
cudaSetDevice	dtfft_interface_cuda_runtime	Interface
cudaStreamCreate	dtfft_interface_cuda_runtime	Interface
cudaStreamDestroy	dtfft_interface_cuda_runtime	Interface
cudaStreamQuery	dtfft_interface_cuda_runtime	Interface
cudaStreamSynchronize	dtfft_interface_cuda_runtime	Interface
cudaStreamWaitEvent	dtfft_interface_cuda_runtime	Interface
cufftDestroy	dtfft_interface_cufft	Interface	Frees all GPU resources associated with a cuFFT plan and destroys the internal plan data structure.
cufftGetErrorString	dtfft_interface_cufft	Function	Returns a string representation of the cuFFT error code.
cufftMpAttachReshapeComm	dtfft_interface_cufft	Interface	Attaches a communication handle to a reshape. This function is not collective.
cufftMpCreateReshape	dtfft_interface_cufft	Interface	Initializes a reshape handle for future use. This function is not collective.
cufftMpDestroyReshape	dtfft_interface_cufft	Interface	Destroys a reshape and all its associated data.
cufftMpExecReshapeAsync	dtfft_interface_cufft	Interface	Executes the reshape, redistributing data_in into data_out using the workspace in workspace.
cufftMpGetReshapeSize	dtfft_interface_cufft	Interface	Returns the amount (in bytes) of workspace required to execute the handle.
cufftMpMakeReshape	dtfft_interface_cufft	Interface	Creates a reshape intended to re-distribute a global array of 3D data.
cufftPlanMany	dtfft_interface_cufft	Interface	Creates a FFT plan configuration of dimension rank, with sizes specified in the array n.
cufftSetStream	dtfft_interface_cufft	Interface	Associates a CUDA stream with a cuFFT plan.
cufftXtExec	dtfft_interface_cufft	Interface	Executes any cuFFT transform regardless of precision and type. In case of complex-to-real and real-to-complex transforms, the direction parameter is ignored.
cuLaunchKernel	dtfft_interface_cuda	Function	Launches a CUDA kernel
destoy_helper	dtfft_backend_mpi	Subroutine	Destroys MPI helper
destroy	dtfft_pencil	Subroutine	Destroys pencil
destroy	dtfft_abstract_executor	Subroutine	Destroys plan
destroy	dtfft_executor_fftw_m	Subroutine	Destroys FFTW3 plan
destroy	dtfft_nvrtc_kernel	Subroutine	Destroys kernel
destroy	dtfft_executor_vkfft_m	Subroutine	Destroys vkFFT plan
destroy	dtfft_transpose_handle_host	Subroutine	Destroys `transpose_handle_host` class
destroy	dtfft_transpose_plan_host	Subroutine	Destroys transposition plans
destroy	dtfft_executor_mkl_m	Subroutine	Destroys MKL plan
destroy	dtfft_transpose_handle_cuda	Subroutine	Destroys CUDA Transpose Handle
destroy	dtfft_abstract_backend	Subroutine	Destroys Abstract GPU Backend
destroy	dtfft_executor_cufft_m	Subroutine	Destroys cuFFT plan
destroy	dtfft_plan	Subroutine	Destroys plan, frees all memory
destroy	dtfft_backend_cufftmp_m	Subroutine	Destroys cuFFTMp GPU Backend
destroy_code	dtfft_nvrtc_kernel_generator	Subroutine	Frees all memory
destroy_cuda	dtfft_transpose_plan_cuda	Subroutine	Destroys transposition plans
destroy_data_handle	dtfft_transpose_handle_cuda	Subroutine	Destroys handle
destroy_handle	dtfft_transpose_handle_host	Subroutine	Destroys transposition handle
destroy_helper	dtfft_abstract_backend	Subroutine	Destroys helper
destroy_mpi	dtfft_backend_mpi	Subroutine	Destroys MPI backend
destroy_nccl	dtfft_backend_nccl_m	Subroutine	Destroys NCCL backend
destroy_pencil_init	dtfft_pencil	Subroutine	Destroys pencil_init
destroy_pencil_t	dtfft_pencil	Subroutine	Destroys pencil
destroy_pencil_t_private	dtfft_pencil	Subroutine	Destroys pencil
destroy_stream	dtfft_config	Subroutine	Destroy the default stream if it was created
destroy_string	dtfft_utils	Subroutine
destroy_strings	dtfft_utils	Subroutine	Destroys array of string objects
DftiErrorMessage	dtfft_interface_mkl_m	Function	Generates an error message.
DftiErrorMessage_c	dtfft_interface_mkl_m	Interface
dl_error	dtfft_utils	Subroutine	Writes error message to the error unit
dlclose	dtfft_utils	Interface
dlerror	dtfft_utils	Interface
dlopen	dtfft_utils	Interface
dlsym	dtfft_utils	Interface
double_to_string	dtfft_utils	Function	Convert double to string
dtfft_config_t	dtfft_config	Interface	Interface to create a new configuration
dtfft_create_config	dtfft_config	Subroutine	Creates a new configuration and sets default values. Read more…
dtfft_create_plan_c2c_c	dtfft_api	Function	Creates C2C dtFFT Plan, allocates all structures and prepares FFT, C/C++ interface
dtfft_create_plan_c2c_pencil_c	dtfft_api	Function	Creates C2C dtFFT plan from Pencil, allocates all structures and prepares FFT, C/C++/Python interface
dtfft_create_plan_r2r_c	dtfft_api	Function	Creates R2R dtFFT Plan, allocates all structures and prepares FFT, C/C++/Python interface
dtfft_create_plan_r2r_pencil_c	dtfft_api	Function	Creates R2R dtFFT Plan from Pencil, allocates all structures and prepares FFT, C/C++/Python interface
dtfft_destroy_c	dtfft_api	Function	Destroys dtFFT Plan, C/C++ interface
dtfft_execute_c	dtfft_api	Function	Executes dtFFT Plan, C/C++ interface. `aux` can be NULL.
dtfft_get_alloc_bytes_c	dtfft_api	Function	Returns minimum number of bytes required to execute plan, C/C++ interface
dtfft_get_alloc_size_c	dtfft_api	Function	Returns minimum number of bytes to be allocated for `in` and `out` buffers, C/C++ interface
dtfft_get_backend_c	dtfft_api	Function	Returns selected dtfft_backend_t during autotuning
dtfft_get_backend_string	dtfft_parameters	Function	Gets the string description of a GPU backend
dtfft_get_backend_string_c	dtfft_api	Subroutine	Returns string representation of `dtfft_backend_t`
dtfft_get_cuda_stream	dtfft_parameters	Function	Returns the CUDA stream from dtfft_stream_t
dtfft_get_dims_c	dtfft_api	Function	Returns dimensions of plan, C/C++ interface
dtfft_get_element_size_c	dtfft_api	Function	Returns size of element in bytes, C/C++ interface
dtfft_get_error_string	dtfft_errors	Function	Gets the string description of an error code
dtfft_get_error_string_c	dtfft_api	Subroutine	Returns an explaination of `error_code` that could have been previously returned by one of dtFFT API calls, C/C++ interface
dtfft_get_executor_c	dtfft_api	Function	Returns executor type used in plan, C/C++ interface
dtfft_get_executor_string	dtfft_parameters	Function	Gets the string description of an executor
dtfft_get_executor_string_c	dtfft_api	Subroutine
dtfft_get_local_sizes_c	dtfft_api	Function	Returns local sizes, counts in real and Fourier spaces and number of elements to be allocated for `in` and `out` buffers, C/C++ interface.
dtfft_get_pencil_c	dtfft_api	Function	Returns pencil decomposition info, C/C++ interface
dtfft_get_platform_c	dtfft_api	Function	Returns selected dtfft_platform_t during autotuning
dtfft_get_precision_c	dtfft_api	Function	Returns precision used in plan, C/C++ interface
dtfft_get_precision_string	dtfft_parameters	Function	Gets the string description of a precision
dtfft_get_precision_string_c	dtfft_api	Subroutine
dtfft_get_stream_c	dtfft_api	Function	Returns Stream associated with plan
dtfft_get_version	dtfft_parameters	Interface	Get dtFFT version
dtfft_get_version_current	dtfft_parameters	Function	Returns the current version code
dtfft_get_version_required	dtfft_parameters	Function	Returns the version code required by the user
dtfft_get_z_slab_enabled_c	dtfft_api	Function	Checks if dtFFT Plan is using Z-slab optimization
dtfft_mem_alloc_c	dtfft_api	Function	Allocates memory for dtFFT Plan, C/C++ interface
dtfft_mem_free_c	dtfft_api	Function	Frees memory for dtFFT Plan, C/C++ interface
dtfft_pencil_t	dtfft_pencil	Interface	Type bound constuctor for dtfft_pencil_t
dtfft_report_c	dtfft_api	Function	Reports dtFFT Plan, C/C++ interface
dtfft_set_config	dtfft_config	Subroutine	Sets configuration parameters
dtfft_set_config_c	dtfft_api	Function	Sets dtFFT configuration, C/C++ interface
dtfft_stream_t	dtfft_parameters	Interface	Creates dtfft_stream_t from integer(cuda_stream_kind)
dtfft_transpose_c	dtfft_api	Function	Executes single transposition, C/C++ interface.
dynamic_load	dtfft_utils	Function	Dynamically loads library and its symbols
effort_eq	dtfft_parameters	Function
effort_ne	dtfft_parameters	Function
estimate_bank_conflict_ratio	dtfft_nvrtc_block_optimizer	Function	Estimates the bank conflict ratio for a given kernel configuration
estimate_coalescing	dtfft_nvrtc_block_optimizer	Function	Estimate memory coalescing efficiency for a given kernel configuration and transpose type
estimate_memory_pressure	dtfft_nvrtc_block_optimizer	Function	Analytical estimation of memory pressure based on GPU architecture
estimate_occupancy	dtfft_nvrtc_block_optimizer	Function	Calculates theoretical occupancy for a given kernel configuration
estimate_optimal_padding	dtfft_nvrtc_block_optimizer	Function	Estimates the optimal padding for a given tile size and element size
evaluate_analytical_performance	dtfft_nvrtc_block_optimizer	Function	This function evaluates the performance of a kernel configuration based on various architectural and problem-specific parameters.
execute	dtfft_abstract_executor	Subroutine	Executes plan
execute	dtfft_executor_fftw_m	Subroutine	Executes FFTW3 plan
execute	dtfft_nvrtc_kernel	Subroutine	Executes kernel on stream
execute	dtfft_executor_vkfft_m	Subroutine	Executes vkFFT plan
execute	dtfft_abstract_transpose_plan	Subroutine	Executes single transposition
execute	dtfft_transpose_handle_host	Subroutine	Executes transposition
execute	dtfft_executor_mkl_m	Subroutine	Executes MKL plan
execute	dtfft_transpose_handle_cuda	Subroutine	Executes transpose - exchange - unpack
execute	dtfft_abstract_backend	Subroutine	Executes GPU Backend
execute	dtfft_executor_cufft_m	Subroutine	Executes cuFFT plan
execute	dtfft_plan	Subroutine	Executes plan
execute	dtfft_backend_cufftmp_m	Subroutine	Executes cuFFTMp GPU Backend
execute_cuda	dtfft_transpose_plan_cuda	Subroutine	Executes single transposition
execute_mpi	dtfft_backend_mpi	Subroutine	Executes MPI backend
execute_nccl	dtfft_backend_nccl_m	Subroutine	Executes NCCL backend
execute_private	dtfft_transpose_plan_host	Subroutine	Executes single transposition
execute_private	dtfft_plan	Subroutine	Executes plan with specified auxiliary buffer
execute_ptr	dtfft_plan	Subroutine	Executes plan using type(c_ptr) pointers instead of buffers
execute_type_eq	dtfft_parameters	Function
execute_type_ne	dtfft_parameters	Function
executor_eq	dtfft_parameters	Function
executor_ne	dtfft_parameters	Function
fftw_execute_dft	dtfft_interface_fftw_m	Interface
fftw_execute_dft_c2r	dtfft_interface_fftw_m	Interface
fftw_execute_dft_r2c	dtfft_interface_fftw_m	Interface
fftw_execute_r2r	dtfft_interface_fftw_m	Interface
fftw_plan_many_dft	dtfft_interface_fftw_m	Interface
fftw_plan_many_dft_c2r	dtfft_interface_fftw_m	Interface
fftw_plan_many_dft_r2c	dtfft_interface_fftw_m	Interface
fftw_plan_many_r2r	dtfft_interface_fftw_m	Interface
fftwf_execute_dft	dtfft_interface_fftw_m	Interface
fftwf_execute_dft_c2r	dtfft_interface_fftw_m	Interface
fftwf_execute_dft_r2c	dtfft_interface_fftw_m	Interface
fftwf_execute_r2r	dtfft_interface_fftw_m	Interface
fftwf_plan_many_dft	dtfft_interface_fftw_m	Interface
fftwf_plan_many_dft_c2r	dtfft_interface_fftw_m	Interface
fftwf_plan_many_dft_r2c	dtfft_interface_fftw_m	Interface
fftwf_plan_many_r2r	dtfft_interface_fftw_m	Interface
find_valid_combination	dtfft_nvrtc_block_optimizer	Subroutine	This subroutine optimizes the tile size and number of rows for narrow matrices by adjusting them to be compatible with the warp size.
float_to_string	dtfft_utils	Function	Convert double to string
free_datatypes	dtfft_transpose_handle_host	Subroutine	Frees temporary datatypes
free_mem	dtfft_abstract_transpose_plan	Subroutine	Frees memory based on `backend`
generate_candidates	dtfft_nvrtc_block_optimizer	Subroutine	Generate kernel configuration candidates for given problem
get	dtfft_nvrtc_kernel_cache	Function	Returns cached kernel if it exists. If not returns null pointer.
get_alloc_bytes	dtfft_plan	Function	Returns minimum number of bytes required to execute plan
get_alloc_size	dtfft_plan	Function	Wrapper around `get_local_sizes` to obtain number of elements only
get_ampere_architecture	dtfft_nvrtc_block_optimizer	Function	Ampere architecture (Compute Capability 8.0)
get_aux_size	dtfft_abstract_transpose_plan	Function
get_aux_size	dtfft_transpose_plan_cuda	Function
get_aux_size	dtfft_transpose_handle_cuda	Function	Returns number of bytes required by aux buffer
get_aux_size	dtfft_abstract_backend	Function	Returns number of bytes required by aux buffer
get_backend	dtfft_abstract_transpose_plan	Function	Returns plan GPU backend
get_backend	dtfft_plan	Function	Returns selected GPU backend during autotuning
get_code_init	dtfft_nvrtc_kernel_generator	Subroutine	Generates basic code that is used in all other kernels
get_comm	dtfft_api	Function
get_conf_backend	dtfft_config	Function	Returns GPU backend set by the user or default one
get_conf_configs_to_test	dtfft_config	Function	Returns the number of configurations to test
get_conf_forced_kernel_optimization	dtfft_config	Function	Whether forced kernel optimization is enabled or not
get_conf_internal	dtfft_config	Interface	Returns value from configuration unless environment variable is set
get_conf_internal_int32	dtfft_config	Function	Returns value from configuration unless environment variable is set
get_conf_internal_logical	dtfft_config	Function	Returns value from configuration unless environment variable is set
get_conf_kernel_optimization_enabled	dtfft_config	Function	Whether kernel optimization is enabled or not
get_conf_log_enabled	dtfft_config	Function	Whether logging is enabled or not
get_conf_measure_iters	dtfft_config	Function	Returns the number of measurement iterations
get_conf_measure_warmup_iters	dtfft_config	Function	Returns the number of warmup iterations
get_conf_mpi_enabled	dtfft_config	Function	Whether MPI backends are enabled or not
get_conf_nccl_enabled	dtfft_config	Function	Whether NCCL backends are enabled or not
get_conf_nvshmem_enabled	dtfft_config	Function	Whether nvshmem backends are enabled or not
get_conf_pipelined_enabled	dtfft_config	Function	Whether pipelined backends are enabled or not
get_conf_platform	dtfft_config	Function	Returns platform set by the user or default one
get_conf_stream	dtfft_config	Function	Returns either the custom provided by user or creates a new one
get_conf_z_slab_enabled	dtfft_config	Function	Whether Z-slab optimization is enabled or not
get_contiguous_execution_blocks	dtfft_nvrtc_kernel	Subroutine	Gets the number of blocks and threads for a contiguous execution
get_datatype_from_env	dtfft_config	Function	Obtains datatype id from environment variable
get_device_props	dtfft_interface_cuda_runtime	Interface
get_dims	dtfft_plan	Subroutine	Returns global dimensions
get_element_size	dtfft_plan	Function	Returns number of bytes required to store single element.
get_env	dtfft_config	Interface	Obtains environment variable
get_env_base	dtfft_config	Function	Base function of obtaining dtFFT environment variable
get_env_int32	dtfft_config	Function	Base Integer function of obtaining dtFFT environment variable
get_env_int8	dtfft_config	Function	Obtains int8 environment variable
get_env_logical	dtfft_config	Function	Obtains logical environment variable
get_env_string	dtfft_config	Function	Obtains string environment variable
get_executor	dtfft_plan	Function	Returns FFT Executor associated with plan
get_inverse_kind	dtfft_utils	Function	Get the inverse R2R kind of transform for the given R2R kind
get_kernel	dtfft_nvrtc_kernel	Subroutine	Compiles kernel and caches it. Returns compiled kernel.
get_kernel_args	dtfft_nvrtc_kernel	Subroutine	Populates kernel arguments based on kernel type
get_local_size	dtfft_pencil	Subroutine	Computes local portions of data based on global count and position inside grid communicator
get_local_sizes	dtfft_pencil	Subroutine	Obtain local starts and counts in `real` and `fourier` spaces
get_local_sizes	dtfft_plan	Subroutine	Obtain local starts and counts in `real` and `fourier` spaces
get_neighbor_function_code	dtfft_nvrtc_kernel_generator	Subroutine	Generated device function that is used to determite id of process that to which data is being sent or from which data has been recieved based on local element coordinate
get_pencil	dtfft_plan	Function	Returns pencil decomposition
get_plan_execution_time	dtfft_transpose_plan_host	Function	Creates transpose plan and executes it `DTFFT_MEASURE_WARMUP_ITERS` + `DTFFT_MEASURE_ITERS` times Read more…
get_platform	dtfft_plan	Function	Returns execution platform of the plan (HOST or CUDA)
get_precision	dtfft_plan	Function	Returns precision of the plan
get_stream_int64	dtfft_plan	Subroutine	Returns CUDA stream associated with plan
get_stream_ptr	dtfft_plan	Subroutine	Returns CUDA stream associated with plan
get_transpose_kernel	dtfft_nvrtc_kernel	Subroutine
get_transpose_kernel_code	dtfft_nvrtc_kernel_generator	Function	Generates code that will be used to locally tranpose data and prepares to send it to other processes ndims == 2
get_transpose_type	dtfft_pencil	Function	Determines transpose ID based on pencils
get_true_transpose_type	dtfft_nvrtc_kernel_cache	Function	Returns generic transpose id. Since X-Y and Y-Z transpositions are symmectric, it returns only one of them. X-Z and Z-X are not symmetric
get_unpack_kernel_code	dtfft_nvrtc_kernel_generator	Function	Generates code that will be used to unpack data when it is recieved
get_unpack_pipelined_kernel_code	dtfft_nvrtc_kernel_generator	Function	Generates code that will be used to partially unpack data when it is recieved from other process
get_varying_dim	dtfft_pencil	Function
get_volta_architecture	dtfft_nvrtc_block_optimizer	Function	Volta architecture (Compute Capability 7.0)
get_z_slab_enabled	dtfft_plan	Function	Returns logical value is Z-slab optimization enabled internally
gpu_backend_eq	dtfft_parameters	Function
gpu_backend_ne	dtfft_parameters	Function
init_environment	dtfft_config	Subroutine
init_internal	dtfft_config	Function	Checks if MPI is initialized and loads environment variables
int32_to_string	dtfft_utils	Function	Convert 32-bit integer to string
int64_to_string	dtfft_utils	Function	Convert 64-bit integer to string
int8_to_string	dtfft_utils	Function	Convert 8-bit integer to string
is_backend_cufftmp	dtfft_parameters	Function
is_backend_mpi	dtfft_parameters	Function
is_backend_nccl	dtfft_parameters	Function
is_backend_nvshmem	dtfft_parameters	Function
is_backend_pipelined	dtfft_parameters	Function
is_cuda_executor	dtfft_parameters	Function
is_device_ptr	dtfft_utils	Interface
is_host_executor	dtfft_parameters	Function
is_null_funptr	dtfft_utils	Function	Checks if pointer is NULL
is_null_ptr	dtfft_utils	Function	Checks if pointer is NULL
is_null_ptr	dtfft_utils	Interface	Checks if pointer is NULL
is_nvshmem_ptr	dtfft_interface_nvshmem	Function	Checks if pointer is a symmetric nvshmem allocated pointer
is_same_ptr	dtfft_utils	Function	Checks if two pointer are the same
is_transpose_kernel	dtfft_parameters	Function
is_unpack_kernel	dtfft_parameters	Function
is_valid_comm_type	dtfft_parameters	Function
is_valid_dimension	dtfft_parameters	Function
is_valid_effort	dtfft_parameters	Function
is_valid_execute_type	dtfft_parameters	Function
is_valid_executor	dtfft_parameters	Function
is_valid_gpu_backend	dtfft_parameters	Function
is_valid_platform	dtfft_parameters	Function
is_valid_precision	dtfft_parameters	Function
is_valid_r2r_kind	dtfft_parameters	Function
is_valid_transpose_type	dtfft_parameters	Function
kernel_type_eq	dtfft_parameters	Function
kernel_type_ne	dtfft_parameters	Function
load	dtfft_interface_vkfft_m	Function	Loads VkFFT library
load_cuda	dtfft_interface_cuda	Function	Loads the CUDA Driver library and needed symbols
load_library	dtfft_utils	Function	Dynamically loads library
load_nvrtc	dtfft_interface_nvrtc	Function	Dynamically loads nvRTC library and its functions
load_symbol	dtfft_utils	Function	Dynamically loads symbol from library
load_vkfft	dtfft_interface_vkfft_m	Function	Loads VkFFT library based on the platform
make_plan	dtfft_executor_mkl_m	Subroutine	Creates general MKL plan
make_public	dtfft_pencil	Function	Creates public object that users can use to create own FFT backends
mem_alloc	dtfft_executor_fftw_m	Subroutine	Allocates FFTW3 memory
mem_alloc	dtfft_executor_vkfft_m	Subroutine	Dummy method. Raises `error stop`
mem_alloc	dtfft_abstract_transpose_plan	Subroutine	Allocates memory based on selected backend
mem_alloc	dtfft_executor_mkl_m	Subroutine	Allocates MKL memory
mem_alloc	dtfft_executor_cufft_m	Subroutine	Dummy method. Raises `error stop`
mem_alloc_c32_1d	dtfft_plan	Subroutine	Allocates pointer of rank 1
mem_alloc_c32_2d	dtfft_plan	Subroutine	Allocates pointer of rank 2
mem_alloc_c32_3d	dtfft_plan	Subroutine	Allocates pointer of rank 3
mem_alloc_c64_1d	dtfft_plan	Subroutine	Allocates pointer of rank 1
mem_alloc_c64_2d	dtfft_plan	Subroutine	Allocates pointer of rank 2
mem_alloc_c64_3d	dtfft_plan	Subroutine	Allocates pointer of rank 3
mem_alloc_host	dtfft_utils	Function	Allocates memory using C11 Standard alloc_align with 16 bytes alignment
mem_alloc_ptr	dtfft_plan	Subroutine	Allocates memory specific for this plan
mem_alloc_r32_1d	dtfft_plan	Subroutine	Allocates pointer of rank 1
mem_alloc_r32_2d	dtfft_plan	Subroutine	Allocates pointer of rank 2
mem_alloc_r32_3d	dtfft_plan	Subroutine	Allocates pointer of rank 3
mem_alloc_r64_1d	dtfft_plan	Subroutine	Allocates pointer of rank 1
mem_alloc_r64_2d	dtfft_plan	Subroutine	Allocates pointer of rank 2
mem_alloc_r64_3d	dtfft_plan	Subroutine	Allocates pointer of rank 3
mem_free	dtfft_executor_fftw_m	Subroutine	Frees FFTW3 aligned memory
mem_free	dtfft_executor_vkfft_m	Subroutine	Dummy method. Raises `error stop`
mem_free	dtfft_abstract_transpose_plan	Subroutine	Frees memory allocated with mem_alloc
mem_free	dtfft_executor_mkl_m	Subroutine	Frees MKL aligned memory
mem_free	dtfft_executor_cufft_m	Subroutine	Dummy method. Raises `error stop`
mem_free_c32_1d	dtfft_plan	Subroutine	Frees previously allocated memory specific for this plan
mem_free_c32_2d	dtfft_plan	Subroutine	Frees previously allocated memory specific for this plan
mem_free_c32_3d	dtfft_plan	Subroutine	Frees previously allocated memory specific for this plan
mem_free_c64_1d	dtfft_plan	Subroutine	Frees previously allocated memory specific for this plan
mem_free_c64_2d	dtfft_plan	Subroutine	Frees previously allocated memory specific for this plan
mem_free_c64_3d	dtfft_plan	Subroutine	Frees previously allocated memory specific for this plan
mem_free_host	dtfft_utils	Interface
mem_free_ptr	dtfft_plan	Subroutine	Frees previously allocated memory specific for this plan
mem_free_r32_1d	dtfft_plan	Subroutine	Frees previously allocated memory specific for this plan
mem_free_r32_2d	dtfft_plan	Subroutine	Frees previously allocated memory specific for this plan
mem_free_r32_3d	dtfft_plan	Subroutine	Frees previously allocated memory specific for this plan
mem_free_r64_1d	dtfft_plan	Subroutine	Frees previously allocated memory specific for this plan
mem_free_r64_2d	dtfft_plan	Subroutine	Frees previously allocated memory specific for this plan
mem_free_r64_3d	dtfft_plan	Subroutine	Frees previously allocated memory specific for this plan
mkl_dfti_commit_desc	dtfft_interface_mkl_m	Interface
mkl_dfti_create_desc	dtfft_interface_mkl_m	Interface
mkl_dfti_execute	dtfft_interface_mkl_m	Interface
mkl_dfti_free_desc	dtfft_interface_mkl_m	Interface
mkl_dfti_mem_alloc	dtfft_interface_mkl_m	Interface
mkl_dfti_mem_free	dtfft_interface_mkl_m	Interface
mkl_dfti_set_value	dtfft_interface_mkl_m	Interface	Sets one particular configuration parameter with the specified configuration value.
ncclCommDeregister	dtfft_interface_nccl	Interface	Deregister a buffer for collective communication.
ncclCommDestroy	dtfft_interface_nccl	Interface	Destroy a communicator object comm.
ncclCommInitRank	dtfft_interface_nccl	Interface	Creates a new communicator (multi thread/process version). Read more…
ncclCommRegister	dtfft_interface_nccl	Interface	Register a buffer for collective communication.
ncclGetErrorString	dtfft_interface_nccl	Function	Generates an error message.
ncclGetErrorString_c	dtfft_interface_nccl	Interface	Returns a human-readable string corresponding to the passed error code.
ncclGetUniqueId	dtfft_interface_nccl	Interface	Generates an Id to be used in ncclCommInitRank. ncclGetUniqueId should be called once when creating a communicator and the Id should be distributed to all ranks in the communicator before calling ncclCommInitRank. uniqueId should point to a ncclUniqueId object allocated by the user.
ncclGroupEnd	dtfft_interface_nccl	Interface	End a group call. Read more…
ncclGroupStart	dtfft_interface_nccl	Interface	Start a group call. Read more…
ncclMemAlloc	dtfft_interface_nccl	Interface	Allocate a GPU buffer with size. Allocated buffer head address will be returned by ptr, and the actual allocated size can be larger than requested because of the buffer granularity requirements from all types of NCCL optimizations.
ncclMemFree	dtfft_interface_nccl	Interface	Free memory allocated by ncclMemAlloc().
ncclRecv	dtfft_interface_nccl	Interface	Receive data from rank peer into recvbuff. Read more…
ncclSend	dtfft_interface_nccl	Interface	Send data from sendbuff to rank peer. Read more…
nvrtcGetErrorString	dtfft_interface_nvrtc	Function	Helper function that returns a string describing the given nvrtcResult code For unrecognized enumeration values, it returns “NVRTC_ERROR unknown”
nvshmem_free	dtfft_interface_nvshmem	Interface
nvshmem_malloc	dtfft_interface_nvshmem	Interface
nvshmem_my_pe	dtfft_interface_nvshmem	Interface
nvshmem_ptr	dtfft_interface_nvshmem	Interface
nvshmemx_float_alltoall_on_stream	dtfft_interface_nvshmem	Interface
nvshmemx_init_status	dtfft_interface_nvshmem	Interface
nvshmemx_sync_all_on_stream	dtfft_interface_nvshmem	Interface
nvtxDomainCreate_c	dtfft_interface_nvtx	Interface
nvtxDomainRangePop_c	dtfft_interface_nvtx	Interface
nvtxDomainRangePushEx_c	dtfft_interface_nvtx	Interface
operator(/=)	dtfft_parameters	Interface
operator(==)	dtfft_parameters	Interface
pencil_c2f	dtfft_pencil	Subroutine	Converts C pencil to Fortran pencil
pencil_f2c	dtfft_pencil	Subroutine	Converts Fortran pencil to C pencil
platform_eq	dtfft_parameters	Function
platform_ne	dtfft_parameters	Function
pop_nvtx_domain_range	dtfft_interface_nvtx	Subroutine	Pops a range from the NVTX domain
precision_eq	dtfft_parameters	Function
precision_ne	dtfft_parameters	Function
push_nvtx_domain_range	dtfft_interface_nvtx	Subroutine	Pushes a range to the NVTX domain
r2r_kind_eq	dtfft_parameters	Function
r2r_kind_ne	dtfft_parameters	Function
remove	dtfft_nvrtc_kernel_cache	Subroutine	Takes CUDA kernel as an argument and searches for it in cache If kernel is found than reduces `ref_count` of such entry and kernel becomes a null pointer
report	dtfft_plan	Subroutine	Prints plan-related information to stdout
run_autotune_backend	dtfft_transpose_plan_cuda	Subroutine	Runs autotune for all backends
run_mpi_a2a	dtfft_backend_mpi	Subroutine	Executes MPI all-to-all communication
run_mpi_p2p	dtfft_backend_mpi	Subroutine	Executes MPI point-to-point communication
set_unpack_kernel	dtfft_abstract_backend	Subroutine	Sets unpack kernel for pipelined backend
sort_by_varying_dim	dtfft_pencil	Subroutine
sort_candidates_by_score	dtfft_nvrtc_block_optimizer	Subroutine	Sorting candidates by their performance scores
stream_from_int64	dtfft_parameters	Function	Creates dtfft_stream_t from integer(cuda_stream_kind)
string	dtfft_utils	Interface	Creates string object
string_c2f	dtfft_utils	Subroutine	Convert C string to Fortran string
string_constructor	dtfft_utils	Function	Creates string object
string_f2c	dtfft_utils	Subroutine	Convert Fortran string to C string
to_cstr	dtfft_nvrtc_kernel_generator	Subroutine	Converts Fortran CUDA code to C pointer
to_str	dtfft_utils	Interface	Convert various types to string
transpose	dtfft_plan	Subroutine	Performs single transposition Read more…
transpose_ptr	dtfft_plan	Subroutine	Performs single transposition using type(c_ptr) pointers instead of buffers Read more…
transpose_type_eq	dtfft_parameters	Function
transpose_type_ne	dtfft_parameters	Function
unload_library	dtfft_utils	Subroutine	Unloads library
write_message	dtfft_utils	Subroutine	Write message to the specified unit

Help