On Sat, Feb 19, 2011 at 7:43 AM, Gong Ding <span dir="ltr"><<a href="mailto:gdiso@ustc.edu">gdiso@ustc.edu</a>></span> wrote:<br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div>So dynamic array is suitable for acting as a plus to the preallocation. <br>Only a few extra matrix entries not considered in the preallocation are needed to be processed. <br></div><div>For example, an integral boundary condition with dynamic integration range may have different nonzero entry in a row,<br>
which can be hold by the dynamic array.</div></blockquote><div><br></div><div>I did benchmark of the STL dynamic data structures, and the memory overhead is quite large. The problem</div><div>here is not necessarily what could be done, but what community expectations are. People are not going to</div>
<div>ditch their old Fortran code for something that allocations 3-4 times the memory.</div><div><br></div><div>As Barry points out, you could easily make a new subclass.</div><div><br></div><div> Matt</div><div> </div>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;"><div> <br> </div><div><div></div><div class="h5"><blockquote style="padding-left:5px;margin-left:5px;border-left:2px solid rgb(160, 198, 229);margin-right:0px">
<br><p>Values in each location are often set many times. Once per element in FEM, so about 20 times for P1 tets. That uses a lot more memory and you need to sort that beast to count correctly. Using a separate dynamic data structure for each row would be a lot more mallocs, but you could keep the rows sorted and avoid storing 20 copies, however insertion is still expensive. A heap is nice for insertion, but not for searching.</p>
<p>So dynamic data structures could help, but they still cost quite a bit. The preallocation problem is trivial for finite difference methods so any useful solution needs to handle many insertions to the same location.</p>
<p></p><blockquote type="cite">On Feb 19, 2011 9:35 AM, "Gong Ding" <<a href="mailto:gdiso@ustc.edu" target="_blank">gdiso@ustc.edu</a>> wrote:<br><br>Hi,<br>
After reading the source code of aij.c, I think the MatSetValues function can be more flexible when preallocation is not correct.<br>
<br>
Why not use a dynamic array such as c++ vector of triple(a, i, j) to buffer the operation?<br>
And flush the buffer to real a,i,j array when MatAssemblyEnd is called?<br>
<font color="#888888"><br>
Gong Ding<br>
</font></blockquote>
</blockquote><span></span></div></div></blockquote></div><br><br clear="all"><br>-- <br>What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>
-- Norbert Wiener<br>