syten::EnvVars Namespace Reference

Namespace for generic static variables initialised from environment variables. More...

Functions
void	init_env_bool (char const *name, bool &target, std::string const &msg="")
	Sets target to true if name is a set environment variables, if msg is non-zero, then also prints msg. More...
template<typename Function >
void	init_env_fun (char const *name, Function f)
	Calls the specified function f is an environment variable name exists. More...
template<typename Target , typename Function >
void	init_env_var (char const *name, Target &target, Function f, std::string const &msg="")
	Sets target to f(str) with str initialised from the environment variable name, if set. More...
template<typename Target >
void	init_env_var (char const *name, Target &target, std::string const &msg="")
	Sets target to str, with str initialised from the environment variable name, if set. More...
void	init_env_vars ()
	Initialises generic environment variables, called from SYTEN_BPO_INIT. More...

Variables
std::size_t	cuda_num_handles = 10
	Number of handles to initialise for each GPU, set via SYTEN_CUDA_NUM_HANDLES. More...
std::size_t	cuda_size_to_gpu = 1024 * 1024
	When deciding in an algorithm whether to move a tensor to the GPU or from the GPU, use this value (in bytes) as the threshold. More...
bool	dbg_ad = false
	Set via SYTEN_DBG_AD. More...
TruncationType	default_tpo_truncate = TruncationType::Delinearise
	Set via SYTEN_DEFAULT_TPO_TRUNCATE, one of the non-default truncation types. More...
bool	dense_tts = false
	Set via SYTEN_DENSE_TTS, enabled temporary transpose storage. More...
bool	disable_backtraces = false
	Set via SYTEN_DISABLE_BACKTRACES, does what it says. More...
bool	disable_history = false
	Set via SYTEN_NO_HISTORY, disables adding new elements to History() objects. More...
bool	dynarray_allocs = false
	Set via SYTEN_INFO_DYNARRAY_ALLOC, causes DynArray-allocators to print the size of the allocated memory region. More...
bool	info_reordering = false
	Set via SYTEN_INFO_REORDERING, causes info about relevant reordering in dense products to be printed. More...
int	minimise_ranks = 0
	Set via SYTEN_MINIMISE_RANKS, enables less memory-intensive algorithms by reducing tensor ranks at the cost of computational efficiency. More...
bool	mpo_parse_notrunc = false
	Set via SYTEN_MPO_PARSE_NOTRUNC, disables truncation of MPO during parsing. More...
bool	no_lz4_compression = false
	If this flag is set, no compression is done during saving. More...
bool	prod_check_early = false
	Set via SYTEN_PROD_CHECK_EARLY and from the Pyten toolkit. More...
HandleFermions	stensor_handle_ferm = HandleFermions::y()
	Disables all fermionic signs in STensors. More...
bool	stensor_register = false
	Set to true to register every syten::STensor object into the STensor registry. More...
int	t3n_dmrg_energy_precision = 15
	precision of energy output in T3N-DMRG. More...
bool	tensor_time = false
	Set via SYTEN_TENSOR_TIME, enables timing of tensor products and decompositions. More...
bool	use_external_debugger = false
	Set via SYTEN_USE_EXTERN_DEBUG, disables termination handling and backtrace printing. More...
bool	varapplyortho_timing = false
	Set via SYTEN_VARAPPLYORTHO_TIMING, enables detailed print-outs of timing information in apply_op_orthogonalise_fit(). More...

Detailed Description

Namespace for generic static variables initialised from environment variables.