<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta content="text/html; charset=UTF-8" http-equiv="Content-Type" />
<title></title>
</head>
<body>
<p style="margin: 0px;"><span> </span></p>
<div style="margin: 5px 0px 5px 0px;">
On January 25, 2011 at 3:53 PM "Justin A. Lemkul" <jalemkul@vt.edu> wrote:<br />
<br />
><br />
><br />
> TJ Mustard wrote:<br />
><br />
> <snip><br />
><br />
> >  > 1. Do the systems in question crash immediately (i.e., step zero) or<br />
> > do they run<br />
> >  > for some time?<br />
> >  ><br />
> ><br />
> > Step 0, every time.<br />
> ><br />
> > <br />
> ><br />
> >  > 2. If they give you even a little bit of output, you can analyze<br />
> > which energy<br />
> >  > terms, etc go haywire with the tips listed here:<br />
> >  ><br />
> ><br />
> > All I have seen on these is LINCS Errors and Water molecules unable to<br />
> > be settled.<br />
> ><br />
> > <br />
> ><br />
> > But I will check this out right now, and email if I smell trouble.<br />
> ><br />
> > <br />
> ><br />
> >  ><br />
> > http://www.gromacs.org/Documentation/Terminology/Blowing_Up#Diagnosing_an_Unstable_System<br />
> >  ><br />
> >  > That would help in tracking down any potential bug or error.<br />
> >  ><br />
> >  > 3. Is it just the production runs that are crashing, or everything? <br />
> > If EM isn't<br />
> >  > even working, that smells even buggier.<br />
> ><br />
> > Awesome question here, we have seen some weird stuff. Sometimes the<br />
> > cluster will give us segmentation faults, then it will fail on our<br />
> > machines or sometimes not on our iMacs. I know weird! If EM starts on<br />
> > the cluster it will finish. Where we have issues is in positional<br />
> > restraint (PR) and MD and MD/FEP. It doesn't matter if FEP is on or off<br />
> > in a MD (although we are using SD for these MD/FEP runs).<br />
> ><br />
> > <br />
><br />
> Does "sometimes" refer to different simulations, or multiple invocations of the<br />
> same simulation system?  If you're referencing the fact that system A works<br />
> while system B doesn't, we're talking apples and oranges and it's irrelevant to<br />
> the diagnosis (and perhaps some systems simply require greater finesse or a<br />
> different protocol).  If one system continually fails on one system and works on<br />
> another, that's what we need to be discussing.  Sorry if I've missed something,<br />
> I'm just getting confused.
</div>
<p style="margin: 0px;"> </p>
<p style="margin: 0px;">What trends I see are that some jobs that "segment fault" on the cluster will also fault on our machines. But then when they fail on the cluster, via "segmentation faults", they will work on our machines. Never does a single job "sometimes" fail and sometimes not, it either does or does not, every time. We have just gotten used to rebuilding our system and starting these again. We have also gotten used to doing EM on our machines and transferring this folder to be completed on the cluster. Once there it can finish the PR, MD and FEP.</p>
<p style="margin: 0px;"> </p>
<div style="margin: 5px 0px 5px 0px;">
><br />
> ><br />
> >  ><br />
> >  > 4. Are the compilers the same on the iMac vs. AMD cluster?<br />
> ><br />
> > No I am using x86_64-apple-darwin10 GCC 4.4.4 and the cluster is using<br />
> > x86_64-redhat-linux 4.1.2 GCC.<br />
> ><br />
><br />
> Well, I know that for years weird behavior has been attributed to the gcc-4.1.x<br />
> series, including the famous warning on the downloads page:<br />
><br />
> "WARNING: do not use the gcc 4.1.x set of compilers. They are broken. These<br />
> compilers come with recent Linux distributions like Fedora 5/6 etc."<br />
><br />
> I don't know if those issues were ever resolved (some error in Gromacs that<br />
> wasn't playing nice with gcc, or vice versa).<br />
>
</div>
<p style="margin: 0px;"> </p>
<p style="margin: 0px;">Yes I am looking into this now. I hope this is our problem and not some hugely time consuming underlying problem with our system. Thank you.</p>
<p style="margin: 0px;"> </p>
<p style="margin: 0px;">Thanks again,</p>
<p style="margin: 0px;">TJ Mustard</p>
<p style="margin: 0px;"> </p>
<div style="margin: 5px 0px 5px 0px;">
> > I just did a quick yum search and there doesn't seem to be a newer GCC.<br />
> > We know you are going to cmake but we have yet to get it implemented on<br />
> > our cluster successfully.<br />
> ><br />
><br />
> The build system is irrelevant.  You still need a reliable C compiler, whether<br />
> using autoconf or cmake.<br />
><br />
> -Justin<br />
><br />
> > <br />
> ><br />
> > Thank you,<br />
> ><br />
> > TJ Mustard<br />
> ><br />
> > <br />
> ><br />
> >  ><br />
> >  > -Justin<br />
> >  ><br />
> >  > ><br />
> >  > ><br />
> >  > > Now I understand that my iMac works, but it only has 2 cpus and the<br />
> >  > > cluster has 320. Since we are running our jobs via a Bennet's<br />
> > Acceptance<br />
> >  > > Ratio FEP with 21 lambda windows, using just one 2 cpu machine would<br />
> >  > > take too long. Especially since we wish to start pseudo high throughput<br />
> >  > > drug testing.<br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > > In my .mdp files now, the only changes are:<br />
> >  > ><br />
> >  > > (the default setting is on the right of the ";")<br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > > define                   =     ; =<br />
> >  > ><br />
> >  > > ; RUN CONTROL PARAMETERS<br />
> >  > > integrator               = sd    ; = md<br />
> >  > > ; Start time and timestep in ps<br />
> >  > > tinit                    = 0    ; = 0<br />
> >  > > dt                       = 0.004    ; = 0.001<br />
> >  > > nsteps                   = 750000       ; = 0 (this one depends on the<br />
> >  > > window and particular part of our job)<br />
> >  > ><br />
> >  > > ; OUTPUT CONTROL OPTIONS<br />
> >  > > ; Output frequency for coords (x), velocities (v) and forces (f)<br />
> >  > > nstxout                  = 10000    ; = 100 (to save on disk space)<br />
> >  > > nstvout                  = 10000    ; = 100<br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > > ; OPTIONS FOR ELECTROSTATICS AND VDW<br />
> >  > > ; Method for doing electrostatics<br />
> >  > > coulombtype              = PME    ; = Cutoff<br />
> >  > > rcoulomb-switch          = 0    ; = 0<br />
> >  > > rcoulomb                 = 1  ; = 1<br />
> >  > > ; Relative dielectric constant for the medium and the reaction field<br />
> >  > > epsilon_r                = 1    ; = 1<br />
> >  > > epsilon_rf               = 1    ; = 1<br />
> >  > > ; Method for doing Van der Waals<br />
> >  > > vdw-type                 = Cut-off    ; = Cut-off<br />
> >  > > ; cut-off lengths       <br />
> >  > > rvdw-switch              = 0    ; = 0<br />
> >  > > rvdw                     = 1  ; = 1<br />
> >  > > ; Spacing for the PME/PPPM FFT grid<br />
> >  > > fourierspacing           = 0.12    ; = 0.12<br />
> >  > > ; EWALD/PME/PPPM parameters<br />
> >  > > pme_order                = 4    ; = 4<br />
> >  > > ewald_rtol               = 1e-05    ; = 1e-05<br />
> >  > > ewald_geometry           = 3d    ; = 3d<br />
> >  > > epsilon_surface          = 0    ; = 0<br />
> >  > > optimize_fft             = yes    ; = no<br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > > ; OPTIONS FOR WEAK COUPLING ALGORITHMS<br />
> >  > > ; Temperature coupling<br />
> >  > > tcoupl                   = v-rescale    ; = No<br />
> >  > > nsttcouple               = -1    ; = -1<br />
> >  > > nh-chain-length          = 10    ; = 10<br />
> >  > > ; Groups to couple separately<br />
> >  > > tc-grps                  = System    ; =<br />
> >  > > ; Time constant (ps) and reference temperature (K)<br />
> >  > > tau-t                    = 0.1    ; =<br />
> >  > > ref-t                    = 300    ; =<br />
> >  > > ; Pressure coupling     <br />
> >  > > Pcoupl                   = Parrinello-Rahman    ; = No<br />
> >  > > Pcoupltype               = Isotropic<br />
> >  > > nstpcouple               = -1    ; = -1<br />
> >  > > ; Time constant (ps), compressibility (1/bar) and reference P (bar)<br />
> >  > > tau-p                    = 1    ; = 1<br />
> >  > > compressibility          = 4.5e-5    ; =<br />
> >  > > ref-p                    = 1.0    ; =<br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > > ; OPTIONS FOR BONDS   <br />
> >  > > constraints              = all-bonds    ; = none<br />
> >  > > ; Type of constraint algorithm<br />
> >  > > constraint-algorithm     = Lincs    ; = Lincs<br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > > ; Free energy control stuff<br />
> >  > > free-energy              = yes    ; = no<br />
> >  > > init-lambda              = 0.00       ; = 0<br />
> >  > > delta-lambda             = 0    ; = 0<br />
> >  > > foreign_lambda           =        0.05 ; =<br />
> >  > > sc-alpha                 = 0.5    ; = 0<br />
> >  > > sc-power                 = 1.0    ; = 0<br />
> >  > > sc-sigma                 = 0.3    ; = 0.3<br />
> >  > > nstdhdl                  = 1    ; = 10<br />
> >  > > separate-dhdl-file       = yes    ; = yes<br />
> >  > > dhdl-derivatives         = yes    ; = yes<br />
> >  > > dh_hist_size             = 0    ; = 0<br />
> >  > > dh_hist_spacing          = 0.1    ; = 0.1<br />
> >  > > couple-moltype           = LGD    ; =<br />
> >  > > couple-lambda0           = vdw-q    ; = vdw-q<br />
> >  > > couple-lambda1           = none    ; = vdw-q<br />
> >  > > couple-intramol          = no     ;    = no<br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > > Some of these change due to positional restraint md and energy<br />
> > minimization.<br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > > All of these settings have come from either tutorials, papers or<br />
> > peoples<br />
> >  > > advice.<br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > > If it would be advantageous I can post my entire energy minimization,<br />
> >  > > positional restraint, md, and FEP mdp files.<br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > > Thank you,<br />
> >  > ><br />
> >  > > TJ Mustard<br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > ><br />
> >  > >>><br />
> >  > >>><br />
> >  > >>> Below is my command sequence:<br />
> >  > >>><br />
> >  > >>><br />
> >  > >>><br />
> >  > >>> echo<br />
> >  > >>><br />
> > ==============================================================================================================================<br />
> >  > >>> date >>RNAP-C.joblog<br />
> >  > >>> echo g453s-grompp -f em.mdp -c RNAP-C_b4em.gro -p RNAP-C.top -o<br />
> >  > >>> RNAP-C_em.tpr<br />
> >  > >>> /share/apps/gromacs-4.5.3-single/bin/g453s-grompp -f em.mdp -c<br />
> >  > >>> RNAP-C_b4em.gro -p RNAP-C.top -o RNAP-C_em.tpr<br />
> >  > >>> date >>RNAP-C.joblog<br />
> >  > >>> echo g453s-mdrun -v -s RNAP-C_em.tpr -c RNAP-C_after_em.gro -g<br />
> >  > >>> emlog.log -cpo state_em.cpt -nt 2<br />
> >  > >>> /share/apps/gromacs-4.5.3-single/bin/g453s-mdrun -v -s RNAP-C_em.tpr<br />
> >  > >>> -c RNAP-C_after_em.gro -g emlog.log -cpo stat_em.cpt -nt 2<br />
> >  > >>> date >>RNAP-C.joblog<br />
> >  > >>> echo g453s-grompp -f pr.mdp -c RNAP-C_after_em.gro -p RNAP-C.top -o<br />
> >  > >>> RNAP-C_pr.tpr<br />
> >  > >>> /share/apps/gromacs-4.5.3-single/bin/g453s-grompp -f pr.mdp -c<br />
> >  > >>> RNAP-C_after_em.gro -p RNAP-C.top -o RNAP-C_pr.tpr<br />
> >  > >>> echo g453s-mdrun -v -s RNAP-C_pr.tpr -e pr.edr -c RNAP-C_after_pr.gro<br />
> >  > >>> -g prlog.log -cpo state_pr.cpt -nt 2 -dhdl dhdl-pr.xvg<br />
> >  > >>> /share/apps/gromacs-4.5.3-single/bin/g453s-mdrun -v -s RNAP-C_pr.tpr<br />
> >  > >>> -e pr.edr -c RNAP-C_after_pr.gro -g prlog.log -cpo state_pr.cpt -nt 2<br />
> >  > >>> -dhdl dhdl-pr.xvg<br />
> >  > >>> date >>RNAP-C.joblog<br />
> >  > >>> echo g453s-grompp -f md.mdp -c RNAP-C_after_pr.gro -p RNAP-C.top -o<br />
> >  > >>> RNAP-C_md.tpr<br />
> >  > >>> /share/apps/gromacs-4.5.3-single/bin/g453s-grompp -f md.mdp -c<br />
> >  > >>> RNAP-C_after_pr.gro -p RNAP-C.top -o RNAP-C_md.tpr<br />
> >  > >>> date >>RNAP-C.joblog<br />
> >  > >>> echo g453s-mdrun -v -s RNAP-C_md.tpr -o RNAP-C_md.trr -c<br />
> >  > >>> RNAP-C_after_md.gro -g md.log -e md.edr -cpo state_md.cpt -nt 2 -dhdl<br />
> >  > >>> dhdl-md.xvg<br />
> >  > >>> /share/apps/gromacs-4.5.3-single/bin/g453s-mdrun -v -s RNAP-C_md.tpr<br />
> >  > >>> -o RNAP-C_md.trr -c RNAP-C_after_md.gro -g md.log -e md.edr -cpo<br />
> >  > >>> state_md.cpt -nt 2 -dhdl dhdl-md.xvg<br />
> >  > >>> date >>RNAP-C.joblog<br />
> >  > >>> echo g453s-grompp -f FEP.mdp -c RNAP-C_after_md.gro -p RNAP-C.top -o<br />
> >  > >>> RNAP-C_fep.tpr<br />
> >  > >>> /share/apps/gromacs-4.5.3-single/bin/g453s-grompp -f FEP.mdp -c<br />
> >  > >>> RNAP-C_after_md.gro -p RNAP-C.top -o RNAP-C_fep.tpr<br />
> >  > >>> date >>RNAP-C.joblog<br />
> >  > >>> echo g453s-mdrun -v -s RNAP-C_fep.tpr -o RNAP-C_fep.trr -c<br />
> >  > >>> RNAP-C_after_fep.gro -g fep.log -e fep.edr -cpo state_fep.cpt -nt 2<br />
> >  > >>> -dhdl dhdl-fep.xvg<br />
> >  > >>> /share/apps/gromacs-4.5.3-single/bin/g453s-mdrun -v -s RNAP-C_fep.tpr<br />
> >  > >>> -o RNAP-C_fep.trr -c RNAP-C_after_fep.gro -g fep.log -e fep.edr -cpo<br />
> >  > >>> state_fep.cpt -nt 2 -dhdl dhdl-fep.xvg<br />
> >  > >>><br />
> >  > >>><br />
> >  > >>><br />
> >  > >>><br />
> >  > >>><br />
> >  > >>> I can add my .mdps but I do not think they are the problem since I<br />
> >  > >>> know it works on my personal iMac.<br />
> >  > >>><br />
> >  > >>><br />
> >  > >>><br />
> >  > >>> Thank you,<br />
> >  > >>><br />
> >  > >>> TJ Mustard<br />
> >  > >>> Email: mustardt@onid.orst.edu <mailto:mustardt@onid.orst.edu><br />
> >  > >>><br />
> >  > >><br />
> >  > ><br />
> >  > ><br />
> >  > > TJ Mustard<br />
> >  > > Email: mustardt@onid.orst.edu<br />
> >  > ><br />
> >  ><br />
> >  > --<br />
> >  > ========================================<br />
> >  ><br />
> >  > Justin A. Lemkul<br />
> >  > Ph.D. Candidate<br />
> >  > ICTAS Doctoral Scholar<br />
> >  > MILES-IGERT Trainee<br />
> >  > Department of Biochemistry<br />
> >  > Virginia Tech<br />
> >  > Blacksburg, VA<br />
> >  > jalemkul[at]vt.edu | (540) 231-9080<br />
> >  > http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin<br />
> >  ><br />
> >  > ========================================<br />
> >  > --<br />
> >  > gmx-users mailing list    gmx-users@gromacs.org<br />
> >  > http://lists.gromacs.org/mailman/listinfo/gmx-users<br />
> >  > Please search the archive at<br />
> > http://www.gromacs.org/Support/Mailing_Lists/Search before posting!<br />
> >  > Please don't post (un)subscribe requests to the list. Use the<br />
> >  > www interface or send it to gmx-users-request@gromacs.org.<br />
> >  > Can't post? Read http://www.gromacs.org/Support/Mailing_Lists<br />
> >  ><br />
> ><br />
> > <br />
> ><br />
> > TJ Mustard<br />
> > Email: mustardt@onid.orst.edu<br />
> ><br />
><br />
> --<br />
> ========================================<br />
><br />
> Justin A. Lemkul<br />
> Ph.D. Candidate<br />
> ICTAS Doctoral Scholar<br />
> MILES-IGERT Trainee<br />
> Department of Biochemistry<br />
> Virginia Tech<br />
> Blacksburg, VA<br />
> jalemkul[at]vt.edu | (540) 231-9080<br />
> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin<br />
><br />
> ========================================<br />
> --<br />
> gmx-users mailing list    gmx-users@gromacs.org<br />
> http://lists.gromacs.org/mailman/listinfo/gmx-users<br />
> Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting!<br />
> Please don't post (un)subscribe requests to the list. Use the<br />
> www interface or send it to gmx-users-request@gromacs.org.<br />
> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists<br />
>
</div>
<p style="margin: 0px;"> </p>
<p style="font-family: monospace; white-space: nowrap; margin: 5px 0px 5px 0px;">TJ Mustard<br />
Email: mustardt@onid.orst.edu</p>
</body>
</html>