The results of a computational evaluation of several Newton's and quas
i-Newton's method solvers are discussed and analyzed. Computer time an
d memory requirements for iterating a solution to the steady state are
recorded for each method. Roe's flux-difference splitting together wi
th the Spekreijse/Van Albada continuous limiter is used for the spatia
l discretization. Sparse matrix inversions are performed using a modif
ied version of the Boeing real sparse library routines and the conjuga
te gradient squared algorithm. The methods are applied to exact and ap
proximate Newton's method Jacobian systems for flat plate and flat-pla
te/wedge-type geometries. Results indicate that the quasi-Newton's met
hod solvers do not exhibit quadratic convergence, but can be more effi
cient than the exact Newton's method in select cases.