include/libint2/diis.h - Githubissues

swillow commented 5 years ago

Please check the following line.

line number 250: A.block(1, 1, rank-1, rank-1) = B.block(nskip, nskip, rank-1, rank-1)norm; --> A.block(1, 1, rank-1, rank-1) = B.block(nskip, nskip, nvec-1, nvec-1)norm;

evaleev commented 5 years ago

3rd and 4th args to block are block sizes, hence they must match on both sides. see https://eigen.tuxfamily.org/dox/group__TutorialBlockOperations.html

swillow commented 5 years ago

I did not check whether there is a difference between Eigen and armadillo (in the definition of submatrix or block).

Thanks,

Soohaeng

On Thu, Jan 17, 2019 at 2:10 PM Eduard Valeyev notifications@github.com wrote:

Closed #117 https://github.com/evaleev/libint/issues/117.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/evaleev/libint/issues/117#event-2081271983, or mute the thread https://github.com/notifications/unsubscribe-auth/ARgFDbrjDbEhU0G60s8EQSYcR5bhhibdks5vENi8gaJpZM4aGl2y .

evaleev / libint

include/libint2/diis.h #117

Please check the following line.

line number 250: A.block(1, 1, rank-1, rank-1) = B.block(nskip, nskip, rank-1, rank-1)norm; --> A.block(1, 1, rank-1, rank-1) = B.block(nskip, nskip, nvec-1, nvec-1)norm;

evaleev / libint

include/libint2/diis.h #117

Please check the following line.

line number 250: A.block(1, 1, rank-1, rank-1) = B.block(nskip, nskip, rank-1, rank-1)*norm; --> A.block(1, 1, rank-1, rank-1) = B.block(nskip, nskip, nvec-1, nvec-1)*norm;

line number 250: A.block(1, 1, rank-1, rank-1) = B.block(nskip, nskip, rank-1, rank-1)norm; --> A.block(1, 1, rank-1, rank-1) = B.block(nskip, nskip, nvec-1, nvec-1)norm;