HPC Programming on Intel Many-Integrated-Core Hardware with MAGMA Port to Xeon Phi

<table class="figure-group"><tr class="fig-image" id="a"><td><object data="https://static.hindawi.com/articles/sp/volume-2015/502593/figures/502593.fig.004a.svgz" name="502593.fig.004a" type="image/svg+xml"></object></td></tr><tr class="fig-caption"><td><b>(a) </b>Cholesky factorization (<span class="sans-serif">magma_zpotrf_mic</span>)</td></tr><tr class="fig-image" id="b"><td><object data="https://static.hindawi.com/articles/sp/volume-2015/502593/figures/502593.fig.004b.svgz" name="502593.fig.004b" type="image/svg+xml"></object></td></tr><tr class="fig-caption"><td><b>(b) </b>LU factorization (<span class="sans-serif">magma_zgetrf_mic</span>)</td></tr><tr class="fig-image" id="c"><td><object data="https://static.hindawi.com/articles/sp/volume-2015/502593/figures/502593.fig.004c.svgz" name="502593.fig.004c" type="image/svg+xml"></object></td></tr><tr class="fig-caption"><td><b>(c) </b>QR factorization (<span class="sans-serif">magma_zgeqrf_mic</span>)</td></tr></table>

<div>Comparison of the performance versus the optimized CPU version of the MKL libraries for the three one-sided factorizations.</div>

Scientific Programming

fig4

Figure 4

Figure 4: HPC Programming on Intel Many-Integrated-Core Hardware with MAGMA Port to Xeon Phi 