search for: dpotri

Hello all, In a piece of c++ code (part of a future R package), I'm inverting a relatively large covariance matrix many times using: F77_CALL(dpotrf)(&upper, &xnrow, R, &xnrow, &info); F77_CALL(dpotri)(&upper, &xnrow, R, &xnrow, &info); I am sure that the input matrix is PD and inverts just fine. However, I have found that dpotri gives one or more NaNs from time-to-time in the upper (or lower depending on the specification) triangle. When I say, from time-to-time, I mean with...

I give up. Maybe it is my search (Windows) but I cannot seem to find the definition of the F77_CALL or F77_NAME macros. Either there are too many matches or the search just doesn't find it. For example where is the source for: F77_CALL(dpotri) ? Thank you. Kevin

...0.4, 0.4, 1.7}; double id[4] = {1.0, 0.0, 0.0, 1.0}; double lmat[3]; F77_CALL(dpotrf)("L", &d, mat, &d, &info); lmat[0] = mat[0]; lmat[1] = mat[1]; lmat[2] = mat[3]; F77_CALL(dppsv)("L", &d, &d, lmat, id, &d, &info); // id now contains L^(-T) F77_CALL(dpotri)("L", &d, mat, &d, &info); // mat contains mat^(-1) Rprintf("%f\n", id[0] * id[0]); // owing to that id is lower triangular Rprintf("%f\n", mat[0]); return(R_NilValue); } I expected both printed values to be identical, or almost so. But issuing .Call(&q...

..., 0.0, 1.0}; > double lmat[3]; > F77_CALL(dpotrf)("L", &d, mat, &d, &info); > lmat[0] = mat[0]; > lmat[1] = mat[1]; > lmat[2] = mat[3]; > F77_CALL(dppsv)("L", &d, &d, lmat, id, &d, &info); > // id now contains L^(-T) > F77_CALL(dpotri)("L", &d, mat, &d, &info); > // mat contains mat^(-1) > > Rprintf("%f\n", id[0] * id[0]); > // owing to that id is lower triangular > Rprintf("%f\n", mat[0]); > > return(R_NilValue); > > } > > I expected both printed val...

...ecision back. I have a few questions: 1. Does this affect the portability to other OSs or processors? 2. I'm returning the results in a matrix. Will a matrix of REALs be sufficient for holding long doubles, or will it be cast back to doubles? 3. Will calls to FORTRAN BLAS (like dsymv, dpotrf, dpotri) still work with long doubles? Thanks for any help you can provide. -- Richard D. Morey Assistant Professor Psychometrics and Statistics Rijksuniversiteit Groningen / University of Groningen http://drsmorey.org/research/rdmorey

...,999)))) [1] 30.29 0.21 30.59 NA NA This was the first time that windows R outperformed linux R even PIII Xeon 1GHz is slightly faster than PIII 1.125GHz. Notice the abnormal result 9.4s from PIII-M 1.2GHz (cannot be three times faster). After using native fortran lapack (call to dpotrf and dpotri returns 9s for PIII Xeon), I realized that R may treat the matrix as a general matrix rather than symmetric except the abnormal case. To confirm this, I modified the R_LapackPP.cc in Matrix package and it turned out that is.MMatrix() failed to return true. Hence if (checkClass(classes...

...&n_j, &info); if(info != 0) { error("c++ error: Cholesky failed in sp.lm (N1)\n"); } // adesso C contiene la parte superiore dellafattorizzazione di cholesky logdet = 0.0; for(i = 0; i < n_j; i++) logdet += 2*log(C_P[i*n_j+i]); F77_NAME(dpotri)(upper, &n_j, C_P, &n_j, &info); if(info != 0) { error("c++ error: Cholesky inverse failed in sp.lm (N2)\n"); } for(j=0;j<n_j-1;j++) { for(h=j+1;h<n_j;h++) { Sum_Sigma_inv += C_P[j*n_j+h] ; } } Su...