<div dir="ltr"><div class="gmail_quote"><div dir="ltr">On Fri, Jun 15, 2018 at 6:30 PM Storm Weiner <<a href="mailto:stormweiner@berkeley.edu">stormweiner@berkeley.edu</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div><div><div>Hey there,<br><br></div>I'm trying to use the benchmark example script but I don't understand the usage.<br></div></div></div></blockquote><div><br></div><div>I have not updated the script. We now use -log_view, and so the option would be</div><div><br></div><div> -log_view :summary.py:ascii_info_detail</div><div><br></div><div>Eventually, I will update it. We are working on it now.</div><div><br></div><div> Thanks,</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 dir="ltr"><div><div>For reference, I am talking about the script located at <b>petsc/src/benchmarks/benchmarkExample.py</b><br><br>I execute using the suggested example parameters at the bottom of the script <br><br><b>>> ./src/benchmarks/benchmarkExample.py --events IntegBatchCPU IntegBatchGPU IntegGPUOnly --num 52 DMComplex --refine 0.0625 0.00625 0.000625 0.0000625 --blockExp 4 --order=1 CPU='dm_view show_residual=0 compute_function batch' GPU='dm_view show_residual=0 compute_function batch gpu gpu_batches=8'<br></b></div><div><b> <br></b></div><b>Namespace(batch=False, comp=4, daemon=False, events=['VecMDot', 'VecMAXPY', 'KSPGMRESOrthog', 'MatMult', 'VecCUSPCopyTo', 'VecCUSPCopyFrom', 'MatCUSPCopyTo'], gpulang='OpenCL', library='SNES', module='summary', num=50, runs=["CPU='pc_type=none", 'mat_no_inode', 'dm_vec_type=seq', "dm_mat_type=seqaij'", "GPU='pc_type=none", 'mat_no_inode', 'dm_vec_type=seqcusp', 'dm_mat_type=seqaijcusp', "cusp_synchronize'"], size=['10', '20', '50', '100'], stage='Main_Stage')<br>Executing: /g/g19/weiner6/petsc/arch-mvapich-fftw3/lib/snes/ex50 -preload off -log_summary summary.dat -log_summary_python summary.py -'pc_type none -da_grid_x 10 -da_grid_y 10<br><br>/bin/sh: -c: line 0: unexpected EOF while looking for matching `''<br>/bin/sh: -c: line 1: syntax error: unexpected end of file<br><br><br>Traceback (most recent call last):<br> File "./benchmarkExample.py", line 453, in <module><br> run_DMDA(ex, name, opts, args, sizes, times, events, log=log)<br> File "./benchmarkExample.py", line 356, in run_DMDA<br> processSummary('summary', args.stage, args.events, times[name], events[name])<br> File "./benchmarkExample.py", line 123, in processSummary<br> m = __import__(moduleName)<br>ImportError: No module named summary<br><br></b><br><br><br><br></div><div>There are two issues here. <br><br></div><div>1) the script attempts to run <b>petsc/arch-mvapich-fftw3/lib/snes/ex50</b> when it should attempt to run <b>petsc/src/<library>/examples/tutorials/ex<num></b> as indicated in the help menu. In fact, there is no <b>petsc/<arch>/lib/snes</b> directory.<br><br></div><div>2) The <b>"No module named summary"</b> error causes the python script to crash.<br><br></div><div>What is going on here? How should I be using this script? Is there any documentation to describe the usage? I find the help menu to be a bit too sparse and the source code too dense to follow closely.<br><br><br></div><div>More specifically, I want to use this profiling script for more than just the standard petsc examples. If I understand the source correctly, the simplest way to accomplish that is to create a new directory tree <b>petsc/src/dummy_lib/examples/tutorials </b>and populate it with <b>ex<num> </b>compiled executables. I would then access these by running benchmarkExample.py with the <b>--library dummy_lib</b> argument. Is this a reasonable way to benchmark arbitrary petsc programs?<br></div><div><br><br></div><div>Thanks for your time,<br></div><div>Storm<br></div></div>
</blockquote></div><br clear="all"><div><br></div>-- <br><div dir="ltr" class="gmail_signature" data-smartmail="gmail_signature"><div dir="ltr"><div><div dir="ltr"><div>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</div><div><br></div><div><a href="http://www.caam.rice.edu/~mk51/" target="_blank">https://www.cse.buffalo.edu/~knepley/</a><br></div></div></div></div></div></div>