Sunil,<br><br>I tried the mapped_matrix, and it performed as bad as compressed_matrix. I can&#39;t understand why because everyone says it is faster and I could see it myself in a simple app I wrote, like that sparse fill samples. I can&#39;t understand what is wrong with my code. Guys recommended me to use a generalized_vector_of_vector to build the global stiffness matrix and then copy it to the compressed_matrix (which is a good choice for multiplications later), then I got something like this <a href="http://codepad.org/1Jwx1Lgb">http://codepad.org/1Jwx1Lgb</a>.<br>
<br>Just to explain that code better, <span class="n">t</span><span class="o">-&gt;</span><span class="n">getGlobalIndex</span><span class="p">(</span><span class="n">j</span><span class="p">) and </span><span class="n">t</span><span class="o">-&gt;</span><span class="n">getCorotatedStiffness0</span><span class="p">(</span><span class="n">j</span><span class="p">,</span> <span class="n">k</span><span class="p">) are just regular gets, nothing special happens inside them. </span><span class="n">t</span><span class="o">-&gt;</span><span class="n">computeCorotatedStiffness</span><span class="p">() is not something computationally expensive,</span> it performs one Gram-Schimdt orthonormalization on a 3x3 matrix and then 32 3x3 matrix multiplies (not using ublas matrices there). No matter what matrix I use for RKR_1 and RK I get the same poor performance. One proof that makes the code slow is that if I comment those two lines where I perform the element wise sum, one of my samples runs 15-20 times faster (in this specific sample, m_tetrahedrons.size()==617). Then, I&#39;m really lost there, I wouldn&#39;t like to just throw all my work with ublas away because of this. There must be a solution to this problem.<br>
<br>Rui,<br><br>Hey that eigen looks awesome, very promising. If I don&#39;t get to solve this problem with ublas soon, I may give it a try. Thanks for the suggestion.<br><br>Thanks guys,<br><br>x<br><pre><br></pre>