BLUPF90

Module Gibbs is a collection of data manipulation subroutines useful for implementing the Gibbs sampler predominantly in the BLUPF90 environment. To understand the module fully, please read the documentation on modules SPARSEM and PROB, and on BLUPF90.

Subroutine link_hash was inspired by Urs Schnyder, ETH, Switzerland.

Module Gibbs

LINK_HASH

call link_hash_ija(xx,xx_ija)

xx - matrix in sparse_hashm format xx_ija - matrix in sparse_ija format

Matrix in sparse_hashm form can be rapidly constructed but cannot easily be used. Matrix in sparse_ija form can easily be used but cannot be set up directly. Conversion from sparse_hashm to sparse_ija as provided in module SPARSEM is relatively slow. Subroutine link_hash provides a fast conversion from sparse_hashm to sparse_ija format when xx with the same nonzero structure needs to be converted repeatedly. When this routine is called the first time, a link is created that points to equivalent locations in xx and xx_ija. During subsequent calls, it is assumed that the nonzero structure of xx and xx_ija remains the same, and the conversion is done rapidly by using the link to update the numerical values only.

Solve_iterm_block

call solve_iterm_block(xx_ija,xy,sol,i,j,diag,op)

xx_ija - Matrix in sparse_ija format xy - (r8) vector of right hand sides sol - (r8) vector of solutions i - integer value of first equation to solve j - integer value of last equation to solve diag - (r8) matrix of dimension j-i+1 x j-i+1 op - a character variable containing either 'solve' or “update'

When op='solve', the subroutine solves a block of equations from i to j, and puts the diagonal part of the matrix xx_ija into a dense matrix diag. When op='update', the right hand side is updated for the current block of solutions. The “update' option should always follow the “solve” option or the solutions will be incorrect. The need for the 'update' option arises from xx_ija being half-stored.

Other subroutines

Extra subroutines (undocumented) have been added to enable multiple-trait Gibbs-sampling with single-trait matrices only. See programs gibbs.f90 and gibbs1f90.f90 for details.