<html><body><div style="color:#000; background-color:#fff; font-family:times new roman, new york, times, serif;font-size:12pt"><div><span>Poor scaling with MPI on many-core machines can also be due uneven job distributions across cores or jobs being wastefully swapped between cores. You might be able to fix this with some esoteric configuration options of mpirun (--bind-to-core worked for me with openMPI), but the surest option is to switch to gromacs 4.5 and run using thread-level parallelisation, bypassing MPI entirely.</span></div><div><br></div><div><br></div> <div style="font-size: 12pt; font-family: 'times new roman', 'new york', times, serif; "> <div style="font-size: 12pt; font-family: 'times new roman', 'new york', times, serif; "> <div dir="ltr"> <font size="2" face="Arial"> <hr size="1"> <b><span style="font-weight:bold;">From:</span></b> Sara Campos <srrcampos@gmail.com><br> <b><span style="font-weight: bold;">To:</span></b>
gmx-users@gromacs.org <br> <b><span style="font-weight: bold;">Sent:</span></b> Monday, 20 February 2012, 17:12<br> <b><span style="font-weight: bold;">Subject:</span></b> [gmx-users] Scaling/performance on Gromacs 4<br> </font> </div> <br><div id="yiv704499279"><span style="font-family: 'courier new', monospace; ">Dear GROMACS users</span><br><br style="font-family: 'courier new', monospace; "><span style="font-family: 'courier new', monospace; ">My group has had access to a quad processor, 64 core machine (4 x Opteron 6274 @ 2.2 GHz with 16
cores)<br>and I made some</span>
<span style="font-family: 'courier new', monospace; ">performance tests, using the following specifications:</span><br style="font-family: 'courier new', monospace; "><br style="font-family: 'courier new', monospace; "><span style="font-family: 'courier new', monospace; ">System size: 299787 atoms</span><br style="font-family: 'courier new', monospace; ">
<span style="font-family: 'courier new', monospace; ">Number of MD steps: 1500</span><br style="font-family: 'courier new', monospace; "><span style="font-family: 'courier new', monospace; ">Electrostatics treatment: PME</span><br style="font-family: 'courier new', monospace; ">
<span style="font-family: 'courier new', monospace; ">Gromacs version: 4.0.4</span><br style="font-family: 'courier new', monospace; "><span style="font-family: 'courier new', monospace; ">MPI: LAM</span><br style="font-family: 'courier new', monospace; ">
<span style="font-family: 'courier new', monospace; ">Command ran: mpirun -ssi rpi tcp C mdrun_mpi ...</span><br style="font-family: 'courier new', monospace; "><br style="font-family: 'courier new', monospace; "><span style="font-family: 'courier new', monospace; ">#CPUS Time (s) Steps/s</span><br style="font-family: 'courier new', monospace; ">
<span style="font-family: 'courier new', monospace; ">64 195.000 7.69</span><br style="font-family: 'courier new', monospace; "><span style="font-family: 'courier new', monospace; ">32 192.000 7.81</span><br style="font-family: 'courier new', monospace; ">
<span style="font-family: 'courier new', monospace; ">16 275.000 5.45</span><br style="font-family: 'courier new', monospace; "><span style="font-family: 'courier new', monospace; ">8 381.000 3.94</span><br style="font-family: 'courier new', monospace; ">
<span style="font-family: 'courier new', monospace; ">4 751.000 2.00</span><br style="font-family: 'courier new', monospace; "><span style="font-family: 'courier new', monospace; ">2 1001.000 1.50</span><br style="font-family: 'courier new', monospace; ">
<span style="font-family: 'courier new', monospace; ">1 2352.000 0.64</span><br style="font-family: 'courier new', monospace; "><br style="font-family: 'courier new', monospace; "><span style="font-family: 'courier new', monospace; ">The scaling is not good. But the weirdest is the 64 processors performing</span><br style="font-family: 'courier new', monospace; ">
<span style="font-family: 'courier new', monospace; ">the same as 32. I see the plots from Dr. Hess on the GROMACS 4 paper on JCTC</span><br style="font-family: 'courier new', monospace; "><span style="font-family: 'courier new', monospace; ">and I do not understand why this is happening. Can anyone help?</span><br style="font-family: 'courier new', monospace; ">
<br style="font-family: 'courier new', monospace; "><span style="font-family: 'courier new', monospace; ">Thanks in advance,</span><br style="font-family: 'courier new', monospace; "><span style="font-family: 'courier new', monospace; ">Sara</span><br>
</div><br>-- <br>gmx-users mailing list <a ymailto="mailto:gmx-users@gromacs.org" href="mailto:gmx-users@gromacs.org">gmx-users@gromacs.org</a><br><a href="http://lists.gromacs.org/mailman/listinfo/gmx-users" target="_blank">http://lists.gromacs.org/mailman/listinfo/gmx-users</a><br>Please search the archive at <a href="http://www.gromacs.org/Support/Mailing_Lists/Search" target="_blank">http://www.gromacs.org/Support/Mailing_Lists/Search</a> before posting!<br>Please don't post (un)subscribe requests to the list. Use the <br>www interface or send it to <a ymailto="mailto:gmx-users-request@gromacs.org" href="mailto:gmx-users-request@gromacs.org">gmx-users-request@gromacs.org</a>.<br>Can't post? Read <a href="http://www.gromacs.org/Support/Mailing_Lists" target="_blank">http://www.gromacs.org/Support/Mailing_Lists</a><br><br> </div> </div> </div></body></html>