Performance breakdown for all algorithms for the matrix raefsky3 (M2)

<p>Runtime of computational kernels for different number of processes on matrix M2 with block size <span class="math-tex">\(k=32\)</span> after 100 iterations. For each kernel, we show the maximum time taken over all processes. The power scheme dominates the runtime in RandQB_EI. QR with column tournament pivoting dominates the runtime in the deterministic methods.</p>