Computations on large Matrices and Vectors that contain floating-point data -- both hardware float data and arbitrary precision software float data -- can be performed very efficiently in Maple by taking advantage of a built-in library of numeric linear algebra routines. A number of these routines are provided through the alliance between Maplesoft and the Numerical Algorithms Group (NAG). In addition, parts of the CLAPACK and optimized ATLAS libraries have been integrated. In order to get the most benefit from these routines, it is important that you provide Maple with as much information about the structure of your problem as possible. Maple will almost certainly solve your problem without this additional information, and in some cases that is all you need. This Tips and Techniques explains the steps you can take when it is vital that the computations are done as efficiently as possible.