Constant Index
$#! · 0-9 · A · B · C · D · E · F · G · H · I · J · K · L · M · N · O · P · Q · R · S · T · U · V · W · X · Y · Z
A
 ALLOC_ERR, ERR_NUMBERS
 API_BIND_AUTO, API_BIND_MODE
 API_BIND_NO, API_BIND_MODE
 API_BIND_TAB, API_BIND_MODE
 API_CSC_FREE, API_ERASE_CSC
 API_CSC_PRESERVE, API_ERASE_CSC
 API_FACT_LDLT, API_FACT
 API_FACT_LLT, API_FACT
 API_FACT_LU, API_FACT
 API_IO_LOAD, API_IO
 API_IO_LOAD_CSC, API_IO
 API_IO_LOAD_GRAPH, API_IO
 API_IO_NO, API_IO
 API_IO_SAVE, API_IO
 API_IO_SAVE_CSC, API_IO
 API_IO_SAVE_GRAPH, API_IO
 API_NO, API_BOOLEAN
 API_ORDER_LOAD, API_ORDER
 API_ORDER_METIS, API_ORDER
 API_ORDER_PERSONAL, API_ORDER
 API_ORDER_SCOTCH, API_ORDER
 API_RAF_GMRES, API_RAF
 API_RAF_GRAD, API_RAF
 API_RAF_PIVOT, API_RAF
 API_RHS_1, API_RHS
 API_RHS_B, API_RHS
 API_RHS_I, API_RHS
 API_SYM_NO, API_SYM
 API_SYM_YES, API_SYM
 API_TASK_ANALYSE, API_TASK
 API_TASK_BLEND, API_TASK_OLD
 API_TASK_CLEAN, API_TASK
 API_TASK_FAX, API_TASK_OLD
 API_TASK_INIT, API_TASK
 API_TASK_NUMFACT, API_TASK
 API_TASK_ORDERING, API_TASK
 API_TASK_REFINE, API_TASK
 API_TASK_REFINEMENT, API_TASK_OLD
 API_TASK_SCOTCH, API_TASK_OLD
 API_TASK_SOLVE, API_TASK
 API_TASK_SOPALIN, API_TASK_OLD
 API_TASK_SYMBFACT, API_TASK
 API_TASK_UPDOWN, API_TASK_OLD
 API_THCOMM_DEFINED, API_COMM_THREAD_MODE
 API_THCOMM_DISABLED, API_COMM_THREAD_MODE
 API_THCOMM_NBPROC, API_COMM_THREAD_MODE
 API_THCOMM_ONE, API_COMM_THREAD_MODE
 API_TRACE_HUMREAD, API_TRACEFMT
 API_TRACE_PAJE, API_TRACEFMT
 API_TRACE_PICL, API_TRACEFMT
 API_TRACE_UNFORMATED, API_TRACEFMT
 API_VERBOSE_CHATTERBOX, API_VERBOSE
 API_VERBOSE_NO, API_VERBOSE
 API_VERBOSE_NOT, API_VERBOSE
 API_VERBOSE_UNBEARABLE, API_VERBOSE
 API_VERBOSE_YES, API_VERBOSE
 API_YES, API_BOOLEAN
 ASSERT_ERR, ERR_NUMBERS
B
 BAD_DEFINE_ERROR, ERR_NUMBERS
 BADPARAMETER_ERR, ERR_NUMBERS
C
 COMMSTEP_ALLREDUCE, COMMSTEP
 COMMSTEP_DOWN, COMMSTEP
 COMMSTEP_END, COMMSTEP
 COMMSTEP_FACTO, COMMSTEP
 COMMSTEP_FACTOEND, COMMSTEP
 COMMSTEP_INIT, COMMSTEP
 COMMSTEP_REDUCE, COMMSTEP
 COMMSTEP_UP, COMMSTEP
 COMMSTEP_UPDOEND, COMMSTEP
 CSC_DISP_CYCLIC, CSC_DISPATCH_OP
 CSC_DISP_SIMPLE, CSC_DISPATCH_OP
 CSCD_ADD, CSCD_OPERATIONS
 CSCD_KEEP, CSCD_OPERATIONS
 CSCD_MAX, CSCD_OPERATIONS
 CSCD_MIN, CSCD_OPERATIONS
 CSCD_OVW, CSCD_OPERATIONS
D
 DPARM_ANALYZE_TIME, DPARM_ACCESS
 DPARM_EPSILON_MAGN_CTRL, DPARM_ACCESS
 DPARM_EPSILON_REFINEMENT, DPARM_ACCESS
 DPARM_FACT_FLOPS, DPARM_ACCESS
 DPARM_FACT_TIME, DPARM_ACCESS
 DPARM_FILL_IN, DPARM_ACCESS
 DPARM_MEM_MAX, DPARM_ACCESS
 DPARM_PRED_FACT_TIME, DPARM_ACCESS
 DPARM_RAFF_TIME, DPARM_ACCESS
 DPARM_RELATIVE_ERROR, DPARM_ACCESS
 DPARM_SIZE, DPARM_ACCESS
 DPARM_SOLV_FLOPS, DPARM_ACCESS
 DPARM_SOLV_TIME, DPARM_ACCESS
F
 FILE_ERR, ERR_NUMBERS
I
 INTERNAL_ERR, ERR_NUMBERS
 IPARM_ABS, IPARM_ACCESS
 IPARM_ALLOCATED_TERMS, IPARM_ACCESS
 IPARM_AMALGAMATION_LEVEL, IPARM_ACCESS
 IPARM_BASEVAL, IPARM_ACCESS
 IPARM_BINDTHRD, IPARM_ACCESS
 IPARM_CPU_BY_NODE, IPARM_ACCESS
 IPARM_CSCD_CORRECT, IPARM_ACCESS
 IPARM_DEFAULT_ORDERING, IPARM_ACCESS
 IPARM_DISTRIBUTION_LEVEL, IPARM_ACCESS
 IPARM_DOF_COST, IPARM_ACCESS
 IPARM_DOF_NBR, IPARM_ACCESS
 IPARM_END_TASK, IPARM_ACCESS
 IPARM_ERROR_NUMBER, IPARM_ACCESS
 IPARM_ESP, IPARM_ACCESS
 IPARM_ESP_NBTASKS, IPARM_ACCESS
 IPARM_ESP_THRESHOLD, IPARM_ACCESS
 IPARM_FACTORIZATION, IPARM_ACCESS
 IPARM_FILL_MATRIX, IPARM_ACCESS
 IPARM_FREE_CSCPASTIX, IPARM_ACCESS
 IPARM_FREE_CSCUSER, IPARM_ACCESS
 IPARM_GMRES_IM, IPARM_ACCESS
 IPARM_GRAPHDIST, IPARM_ACCESS
 IPARM_INCOMPLETE, IPARM_ACCESS
 IPARM_INERTIA, IPARM_ACCESS
 IPARM_IO_STRATEGY, IPARM_ACCESS
 IPARM_ISOLATE_ZEROS, IPARM_ACCESS
 IPARM_ITERMAX, IPARM_ACCESS
 IPARM_LEVEL_OF_FILL, IPARM_ACCESS
 IPARM_MATRIX_VERIFICATION, IPARM_ACCESS
 IPARM_MAX_BLOCKSIZE, IPARM_ACCESS
 IPARM_MC64, IPARM_ACCESS
 IPARM_METIS_PFACTOR, IPARM_ACCESS
 IPARM_MIN_BLOCKSIZE, IPARM_ACCESS
 IPARM_MODIFY_PARAMETER, IPARM_ACCESS
 IPARM_MURGE_REFINEMENT, IPARM_ACCESS
 IPARM_NB_SMP_NODE_USED, IPARM_ACCESS
 IPARM_NB_THREAD_COMM, IPARM_ACCESS
 IPARM_NBITER, IPARM_ACCESS
 IPARM_NNZEROS, IPARM_ACCESS
 IPARM_NNZEROS_BLOCK_LOCAL, IPARM_ACCESS
 IPARM_ONLY_RAFF, IPARM_ACCESS
 IPARM_OOC_ID, IPARM_ACCESS
 IPARM_OOC_LIMIT, IPARM_ACCESS
 IPARM_OOC_THREAD, IPARM_ACCESS
 IPARM_ORDERING, IPARM_ACCESS
 IPARM_ORDERING_CMAX, IPARM_ACCESS
 IPARM_ORDERING_CMIN, IPARM_ACCESS
 IPARM_ORDERING_FRAT, IPARM_ACCESS
 IPARM_ORDERING_SWITCH_LEVEL, IPARM_ACCESS
 IPARM_PID, IPARM_ACCESS
 IPARM_REFINEMENT, IPARM_ACCESS
 IPARM_RHS_MAKING, IPARM_ACCESS
 IPARM_RHSD_CHECK, IPARM_ACCESS
 IPARM_SCHUR, IPARM_ACCESS
 IPARM_SIZE, IPARM_ACCESS
 IPARM_STARPU, IPARM_ACCESS
 IPARM_START_TASK, IPARM_ACCESS
 IPARM_STATIC_PIVOTING, IPARM_ACCESS
 IPARM_SYM, IPARM_ACCESS
 IPARM_THREAD_COMM_MODE, IPARM_ACCESS
 IPARM_THREAD_NBR, IPARM_ACCESS
 IPARM_TRACEFMT, IPARM_ACCESS
 IPARM_VERBOSE, IPARM_ACCESS
Allocation error
Default binding
Do not bind thread
Use vector given by pastix_setBind
Free the CSC when no longer needed
Do not free the CSC
$LDL^t$ Factorization
$LL^t$ Factorization
$LU$ Factorization
Load ordering during ordering step and symbol matrix instead of symbolic factorisation.
Load CSC(d) during ordering step.
Load graph during ordering step.
No output or input
Save ordering during ordering step and symbol matrix instead of symbolic factorisation.
Save CSC(d) during ordering step.
Save graph during ordering step.
No
Load ordering from disk
Use \metis{} ordering
Apply user’s permutation
Use \scotch{} ordering
GMRES
Conjugate Gradient ($LL^t$ or $LDL^t$ factorization)
Iterative Refinement (only for $LU$ factorization)
$ \forall i, X_i = 1 $
User’s right hand side
$ \forall i, X_i = i $
Nonsymmetric matrix
Symmetric matrix
Tasks mapping and scheduling
Tasks mapping and scheduling
Clean
Symbolic factorization
Set default parameters
Numerical factorization
Ordering
Numerical refinement
Numerical refinement
Ordering
Numerical solve
Numerical factorization
Symbolic factorization
Numerical solve
Given by IPARM_NB_THREAD_COMM
No dedicated communication thread
One communication thread per computation thread
One dedicated communication thread
Use human-readable text trace format
Use Paje trace format
Use PICL trace format
Unformated trace format
Like a gossip
Some messages
Silent mode, no messages
Really talking too much...
Many messages
Yes
Error in one assertion
Error with defines during compilation
Bad parameters given
Ask for an MPI_ALLREDUCE in Funneled mode
Ask to begin the down step
All communications are ended
Ask to begin the factorization step
Signal the end of communications in factorization step
Signal the return of communication thread in initial state
Ask for an MPI_REDUCE in Funneled mode
Ask to begin the up step
Signal the end of different steps during backward/forward substitution
Reparts cyclicly the columns over the proc.
Reparts linearly the columns over the proc.
Add coefficient values.
Keep value from first CSCD.
Keep maximum of first and second CSCD.
Keep minimum of first and second CSCD.
Overwrite with second CSCD value.
Time for Analyse step (wallclock) Default: - OUT
Epsilon for magnitude control Default: 1e^{-31} IN
Epsilon for refinement Default: 1e^{-12} IN
Numerical Factorization flops (rate!)
Time for Numerical Factorization step (wallclock) Default: - OUT
Fill-in Default: - OUT
Maximum memory (-DMEMORY_USAGE) Default: - OUT
Predicted factorization time Default: - OUT
Time for Refinement step (wallclock) Default: - OUT
Relative backward error Default: - OUT
Dparm Size IGNORE Default: - IN
Solve flops (rate!)
Time for Solve step (wallclock) Default: - OUT
Error in In/Out operations
Internal error
ABS (Automatic Blocksize Splitting) Default: API_NO IN
Maximum memory allocated for matrix terms Default: - OUT
Amalgamation level Default: 5 IN
Baseval used for the matrix Default: 0 IN
Thread binding mode (see Thread binding modes) Default: API_BIND_AUTO IN
Number of CPUs per SMP node Default: 0 IN
Indicate if the cscd has been redistributed after blend Default: API_NO IN
Use default ordering parameters with \scotch{} or \metis{} Default: API_YES IN
Distribution level IGNORE Default: IN
Degree of freedom for cost computation (If different from IPARM_DOF_NBR) Default: 0 IN
Degree of freedom per node Default: 1 IN
Indicate the last step to execute (see PaStiX steps) Default: API_TASK_CLEAN IN
Return value Default: - OUT
ESP (Enhanced Sparse Parallelism) Default: API_NO IN
Return the number of tasks generated by ESP Default: - OUT
Minimal block sizee to switch in ESP mode (128 * 128) Default: 16384 IN
Factorization mode (see Factorization modes) Default: API_FACT_LDLT IN
Initialize matrix coefficients (for test only) IGNORE Default: IN
Free internal CSC (Use only without call to Refin.
Free user CSC Default: API_CSC_PRESERVE IN
GMRES restart parameter Default: 25 IN
Specify if the given graph is distributed or not Default: API_YES IN
Incomplete factorization Default: API_NO IN
Return the inertia (symmetric matrix without pivoting) Default: - OUT
IO strategy (see Checkpoints modes) Default: API_IO_NO IN
Isolate null diagonal terms at the end of the matrix Default: API_NO IN
Maximum iteration number for refinement Default: 250 IN
Level of fill for incomplete factorization Default: 1 IN
Check the input matrix Default: API_NO IN
Maximum block size Default: 120 IN
MC64 operation pastix.h IGNORE Default: 0 IN
\metis{} pfactor Default: 0 IN
Minimum block size Default: 60 IN
Indicate if parameters have been set by user Default: API_YES IN
Enable refinement in MURGE Default: API_YES IN
Number of SMP node used IGNORE Default: IN
Number of thread(s) for communication Default: 1 IN
Number of iterations performed in refinement Default: - OUT
Number of nonzero entries in the factorized matrix Default: - OUT
Number of nonzero entries in the local block factorized matrix Default: - OUT
Refinement only Default: API_NO IN
Out of core run ID IGNORE Default: - OUT
Out of core memory limit (Mo) Default: 2000 IN
Out of core thread number IGNORE Default: 1 IN
Choose ordering Default: API_ORDER_SCOTCH IN
Ordering cmax parameter (see \scotch{} User’s Guide) Default: 100000 IN
Ordering cmin parameter (see \scotch{} User’s Guide) Default: 0 IN
Ordering frat parameter (see \scotch{} User’s Guide) Default: 8 IN
Ordering switch level (see \scotch{} User’s Guide) Default: 120 IN
Pid of the first process (used for naming the log directory) Default: -1 OUT
Refinement type (see Refinement modes) Default: API_RAF_GMRES IN
Right-hand-side making (see Right-hand-side modes) Default: API_RHS_B IN
Set to API_NO to avoid RHS redistribution Default: API_YES IN
Schur mode Default: API_NO IN
Iparm Size IGNORE Default: IN
Use StarPU runtime Default: API_NO IN
Indicate the first step to execute (see PaStiX steps) Default: API_TASK_ORDERING IN
Static pivoting Default: - OUT
Symmetric matrix mode (see Symmetric modes) Default: API_SYM_YES IN
Threaded communication mode (see Communication modes) Default: API_THCOMM_DISABLED IN
Number of threads per MPI process Default: 1 IN
Trace format (see Trace modes) Default: API_TRACE_PICL IN
Verbose mode (see Verbose modes) Default: API_VERBOSE_NO IN
Close