The improved BiCG method for large and sparse linear systems on parallel distributed memory architectures | IEEE Conference Publication | IEEE Xplore