<div style="line-height:1.7;color:#000000;font-size:14px;font-family:arial"><DIV> </DIV>
<DIV>Thank you very much. And we will have a try on both MUMPS and the <SPAN id="tran_1_3" jQuery17108381929202886638="28" data-aligning="#src_1_2,#tran_1_3">traditional</SPAN> preconditioners provided by PETSc.</DIV>
<DIV> </DIV>
<DIV> Jim</DIV>
<DIV></DIV>
<DIV id="divNeteaseMailCard"></DIV>
<DIV><BR></DIV>>ÔÚ 2012-06-28 23:01:44£¬"Hong Zhang" <hzhang@mcs.anl.gov> дµÀ£º<BR>
<BLOCKQUOTE style="BORDER-LEFT: #ccc 1px solid; MARGIN: 0px 0px 0px 0.8ex; PADDING-LEFT: 1ex" id="isReplyContent">>Jim,<BR>
<DIV class="gmail_quote">
<BLOCKQUOTE style="BORDER-LEFT: #ccc 1px solid; MARGIN: 0px 0px 0px 0.8ex; PADDING-LEFT: 1ex" class="gmail_quote">
<DIV style="FONT-FAMILY: arial"><BR>
<DIV style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px">> In fact, I will deal with the large sparse symmetric or unsymmetric (depending on the constitutive</DIV>
<DIV style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px">>model used for the soil and the formulation of equations) linear system which usually occurs in geotechnical </DIV>
<DIV style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px">>problems. As you know, In most geotechnical problems, the coefficient matrix can be severely</DIV>
<DIV style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px">>ill-conditioned, thus calling for the development of robust and efficient preconditioners and iterative methods.</DIV>
<DIV style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px"> </DIV>
<DIV style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px">> The iterative method and parallel solver are my choices, actually are required by my professor. As I know,</DIV>
<DIV style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px">>MUMPS is just an external solver like Matlab. So is it a right way to analyse the preconditioners and iterative</DIV></DIV></BLOCKQUOTE>
<DIV><SPAN style="LINE-HEIGHT: 23px; FONT-SIZE: 14px" class="Apple-style-span">>MUMPS and Superlu_dist are parallel sparse direct solvers for ill-conditioned problems.</SPAN></DIV>
<DIV><SPAN style="LINE-HEIGHT: 23px; FONT-SIZE: 14px" class="Apple-style-span">>They can solve problems with 10k unknowns sequentially and scale up to</SPAN></DIV>
<DIV><SPAN style="LINE-HEIGHT: 23px; FONT-SIZE: 14px" class="Apple-style-span">>10+ or 100+ processors, i.e. solve problems with 100k </SPAN><SPAN style="LINE-HEIGHT: 23px; FONT-SIZE: 14px" class="Apple-style-span">unknowns.</SPAN></DIV>
<DIV><SPAN style="LINE-HEIGHT: 23px; FONT-SIZE: 14px" class="Apple-style-span">>Give them a try. Then, explore other preconditioners.</SPAN></DIV>
<DIV><SPAN style="LINE-HEIGHT: 23px; FONT-SIZE: 14px" class="Apple-style-span"><BR></SPAN></DIV>
<BLOCKQUOTE style="BORDER-LEFT: #ccc 1px solid; MARGIN: 0px 0px 0px 0.8ex; PADDING-LEFT: 1ex" class="gmail_quote">
<DIV style="FONT-FAMILY: arial">
<DIV style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px">>methods provided by PETSc like CG, GMRES, Jacobi, sor, etc. for my work? </DIV>
<DIV style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px"> </DIV>
<DIV style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px">> Also I do not quite understand the second part of what Barry said ('You will not be able to use the external</DIV>
<DIV style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px">>solvers with those mode'). So, what does 'those mode' mean. Using quad precision?</DIV></DIV></BLOCKQUOTE>
<DIV><BR></DIV>
<DIV>>Yes, high precisions. Start from small problems with robust solvers, then slowly</DIV>
<DIV>>move to large, complex problems.</DIV>
<DIV><BR></DIV>
<DIV>>Hong </DIV>
<BLOCKQUOTE style="BORDER-LEFT: #ccc 1px solid; MARGIN: 0px 0px 0px 0.8ex; PADDING-LEFT: 1ex" class="gmail_quote">
<DIV style="FONT-FAMILY: arial">
<DIV style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px"> </DIV>
<DIV style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px"><BR></DIV>
<DIV style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px"></DIV>
<DIV style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px"></DIV>
<DIV style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px"><BR></DIV><PRE style="LINE-HEIGHT: 1.7; FONT-SIZE: 14px"><BR> </PRE></DIV></BLOCKQUOTE></DIV></BLOCKQUOTE></div><br><br><span title="neteasefooter"><span id="netease_mail_footer"></span></span>