spyCol
Represent sparsity pattern of a matrix with column coloring
Syntax
spyCol(A,col)
Parameters
- A
a full or sparse matrix
- col
a vector of integers
Description
sp = sparse(rand(20,20) < .2); ij = spget(sp); jacobian = spCompJacobian(deff("y=f",""),sp); col = jacobian.colors; k = unique(col,"keepOrder"); gcf().color_map = jetcolormap(max(k)) subplot(1,2,1) spyCol(sp,col) title(msprintf("Colored (%d x %d) Jacobian\n%d colors",size(sp,1),size(sp,2),max(k))) subplot(1,4,3) compMat = sparse([ij(:,1) col(ij(:,2))],ones(ij(:,1))); spyCol(compMat(:,k),k) title("Compressed Jacobian") subplot(1,4,4) spyCol(jacobian.seed,1:size(jacobian.seed,2)) title("Seed matrix") gcf().children.isoview="on"
See also
- spCompJacobian — Consruct a compressed Jacobian engine from its sparsity pattern using column intersection graph coloring
- spCompHessian — Consruct a compressed Hessian engine from its sparsity pattern using column intersection graph coloring
Bibliography
A. H. Gebremedhin, D. C. Nguyen, Md. M. A. Patwary, A. Pothen}, ColPack: Software for graph coloring and related problems in scientific computing, ACM Trans. Math. Softw., N. 40(1), pp. 1-31, 2013, https://doi.org/10.1145/2513109.2513110.
Report an issue | ||
<< spCompJacobian | Sparse Jacobian computation | Sparse Decompositions >> |