DSDP
File List
Here is a list of all documented files with brief descriptions:
[detail level 123]
o-examples
|o*color.cSecond Basic Example: Read graph from file, formulate the SDP relaxation of k-coloring problem, solve using DSDP, and apply randomized algorithm to generate approximate solutions
|o*dsdp.cCall DSDP from the Matlab environment
|o*maxcut.cMost Basic Example: read graph from file, formulate the SDP relaxation of maximum cut problem, solve using DSDP, and apply randomized algorithm to generate approximate solutions
|o*readsdpa.cRead SDPA data files, pass data into DSDP solver, and print solution
|o*stable.cRead graph from file, formulate the Maximum Stable Set problem, and solve using DSDP
|\*theta.cRead graph complement from file, formulate the Lovasz theta problem, and solve using DSDP
o-include
|o*dsdp5.hThe API to DSDP for those applications using DSDP as a subroutine library
|o*dsdpbasictypes.hSolver, solution types, termination codes,
|o*dsdplapack.hDSDP uses BLAS and LAPACK for many of its operations
|o*dsdpmem.hMemory allocation in DSDP
|\*dsdpsys.hError handling, printing, and profiling
\-src
 o-bounds
 |o*allbounds.cBound all the variables y in (D) and implement DSDPCone operations
 |\*dbounds.cIndividually bound variables y
 o-lp
 |\*dsdplp.cSet linear inequalities in (D) and implement the DSDPCone operations
 o-sdp
 |o*dsdpadddata.cSet block sizes, sparsity, format, and data matrices
 |o*dsdpadddatamat.cSet sparse or dense matrices into the cone
 |o*dsdpblock.cOperations on a single SDP block
 |o*dsdpdatamat.cCall an implementation of the data matrix operations
 |o*dsdpdatamat.hThe interface between the SDPCone and the data matrices
 |o*dsdpdatamat_impl.hStructure of function pointers that each SDP data matrix type (sparse, dense, constant, identity, ...) must implement
 |o*dsdpdsmat.cCall an implmentation of the Delta S matrix operation
 |o*dsdpdsmat.hThe interface between the SDPCone and the Delta S matrix
 |o*dsdpdsmat_impl.hStructure of function pointers that each SDP Delta S matrix type (sparse, dense, diagonal, ...) must implement
 |o*dsdpdualmat.cCall an implementation of the S matrix operations
 |o*dsdpdualmat.hThe interface between the SDPCone and the matrix S
 |o*dsdpdualmat_impl.hStructure of function pointers that each symmetric positive definite matrix type (dense, sparse) must implement
 |o*dsdplanczos.hLanczos procedure determines the maximum step length
 |o*dsdpsdp.hInternal SDPCone data structures and routines
 |o*dsdpstep.cLanczos procedure determines the maximum step length
 |o*dsdpxmat.cCall an implementation of the basic dense matrix array operations
 |o*dsdpxmat.hThe interface between the SDPCone and the dense matrix array
 |o*dsdpxmat_impl.hStructure of function pointers that each dense matrix array type (upper full, packed symmetric, ...) must implement
 |o*printsdpa.cPrint data or solution in SDPA format
 |o*sdpcompute.cCompute the gradient vector and Hessian matrix. Also compute X matrices
 |o*sdpcone.cImplement operations on the SDPCone object
 |o*sdpconesetup.cSetup the internal data structures needed by the SDPCone object
 |o*sdpconevec.cImplementation of the SDPCone vector operations
 |o*sdpconevec.hEach block of the SDPCone has two vectors of appropriate size
 |o*sdpkcone.cImplement the DSDPCone operations using the SDPCone subroutines
 |\*sdpsss.cDetermine the sparsity of each block, and choose suitable dual, Delta S, and X matrix structures
 o-solver
 |o*dsdp.hInternal data structure for the DSDP solver
 |o*dsdpcg.cApply Conjugate Gradient method to solve the Schur matrix
 |o*dsdpcg.hInternal data structure for CG method
 |o*dsdpcone.cMethods of a cone
 |o*dsdpcone.hThe public interface between the cones and the solver
 |o*dsdpcone_impl.hImplementations of a cone (SDP,LP,...) must provide a structure of function pointers
 |o*dsdpconverge.cMonitor convergence
 |o*dsdpconverge.hDetect convergence of the solver from the duality gap and step sizes
 |o*dsdpcops.cApplies conic operations to each cone in the solver
 |o*dsdpobjcone.cApply a constraint that the objective solution (D) is greater than some lower bound
 |o*dsdpprintout.cPrint iteration statistics
 |o*dsdprescone.cVariable r must be nonnegative
 |o*dsdpschurmat.cSolve the Schur matrix for a step direction
 |o*dsdpschurmat.hMethods of a Schur Matrix
 |o*dsdpschurmat_impl.hFunction pointers that a Schur complement matrix (dense, sparse, parallel dense) must provide
 |o*dsdpschurmatadd.cCones to assemble the Schur matrix with these routines
 |o*dsdpsetdata.cSet parameters for the solver and retrieve statistics about the current solution
 |o*dsdpsetoptions.cSet DSDP options from file or command line arguments
 |o*dsdpsetup.cCreate DSDP solver and its data strucutures
 |o*dsdpvec.hVector operations used by the solver
 |o*dsdpx.cX variables, tolerances, errors, and feasibility
 |o*dualalg.cImplements the dual-scaling algorithm
 |\*dualimpl.cDual-scaling operations needed in the solver
 o-sys
 |o*dsdperror.cError codes returned for most subroutines
 |o*dsdplog.cProfile the performance of DSDP
 |o*dsdploginfo.cPrint the progress of the DSDP solver
 |\*dsdptime.cTiming routines for GNU and Microsoft compilers
 \-vecmat
  o*cholmat.cSparse Cholesky for Schur complement matrix
  o*cholmat2.cSparse Cholesky for dual S matrix
  o*diag.cDSDPDualMat, DSDPDSMat, and the DSDPSchurMat implentations for diagonal matrices
  o*dlpack.cDSDPDataMat, DSDPDualMat, DSDPDSMat, DSDPSchurMat, DSDPXMat, objects implemented in dense upper packed symmetric format
  o*drowcol.cDSDPDataMat object such that A(i,j) is nonzero only if i or j equals some integer k. Not completed
  o*dufull.cDSDPDataMat, DSDPDualMat, DSDPDSMat, DSDPSchurMat, DSDPXMat objects implemented in symmetric upper full symmetric format
  o*identity.cDSDPDataMat object representing a multiple of the identity matrix
  o*onemat.cDSDPDataMat object such that all elements are the same value
  o*rmmat.cDSDPDataMat object of rank one outer product
  o*sdpvec.cDSDPVec operations
  o*spds.cDSDPDualMat object with sparse data structures
  o*vech.cDSDPDataMat for sparse matrices in upper packed symmetric format
  o*vechu.cDSDPDataMat for sparse matrices in upper full symmetric format
  \*zeromat.cDSDPDataMat object with all zero entries