Re: [gmx-users] Different Results with Different Platforms

Thank you very much, Justin. I was take note of this in subsequent runs. Thank you so, so much for your assistance. It is greatly appreciated! -- Best wishes, Kelechi -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List

Re: [gmx-users] Different Results with Different Platforms

On 10/6/16 7:13 AM, Kelechi Okoroafor wrote: Thank you for the tip, Justin. I incorrectly thought that the coordinates and box dimensions were sufficient to preserve the pressure of the system. I shall make a habit of restarting simulations with the checkpoint file rather than .gro files. My

Re: [gmx-users] Different Results with Different Platforms

On 10/5/16 9:46 PM, Kelechi Okoroafor wrote: Thank you very much, Justin. So I did not restart the simulation directly. One of my desktops had a hardware malfunction so I had to move my files between desktops. Before moving my files, I converted my final npt.gro trajectory into a .pdb file

Re: [gmx-users] Different Results with Different Platforms

On 10/5/16 8:45 PM, Kelechi Okoroafor wrote: Thank you very much, James. My apologies, I miscommunicated myself. I meant to say that I used NPT simulation to equilibrate the pressure to 1bar as desired. I had planned to do NVT simulation after the NPT simulation, expecting the pressure to

Re: [gmx-users] Different Results with Different Platforms

Thank you very much, James. My apologies, I miscommunicated myself. I meant to say that I used NPT simulation to equilibrate the pressure to 1bar as desired. I had planned to do NVT simulation after the NPT simulation, expecting the pressure to fluctuate reasonably about the mean value of 1bar.

Re: [gmx-users] Different Results with Different Platforms

As advised, I compared the systems between the different platforms and found the bug. I had initially equilibrated the pressure of my system in an NPT run before doing an NVT run on my desktop. Unbeknownst to me, editconf had adjusted the coordinates of my molecules such that the pressure during

Re: [gmx-users] Different Results with Different Platforms

Thank you so much, Szilárd. Yes, it passed the regression tests. Let me explore system characteristics and see how they differ between the platforms. -- Best wishes, Kelechi -- Gromacs Users mailing list * Please search the archive at

Re: [gmx-users] Different Results with Different Platforms

Thank you, Justin. As suggested, I re-ran the regression tests and got the result: All 16 simple tests PASSED All 45 complex tests PASSED All 142 kernel tests PASSED All 10 freeenergy tests PASSED All 12 rotation tests PASSED All 0 extra tests PASSED All 42 pdb2gmx tests PASSED I shall keep the

Re: [gmx-users] Different Results with Different Platforms

If the regressiontests pass, I'd recommend trying to identify differences measurable/reproducible between those runs on shorter and preferably known time-scales. -- Szilárd On Tue, Oct 4, 2016 at 5:30 PM, Kelechi Okoroafor wrote: > Thank you very much, Mark. > > You are

Re: [gmx-users] Different Results with Different Platforms

Thank you very much, Mark. You are absolutely right in that I do not have enough statistical data to definitely say that the structures ought to form at 200ns. However, I have repeated the experimental set in total 4 times on Stampede and twice on my school's cluster. In all cases, without

Re: [gmx-users] Different Results with Different Platforms

On 10/4/16 11:12 AM, Kelechi Okoroafor wrote: Thank you very much, Justin, for your response. It is highly appreciated. When I built Gromacs on both desktops, I recall doing the regression tests, though I do not recall the results of the test, my bad. Do you mean I should re-run the

Re: [gmx-users] Different Results with Different Platforms

Thank you very much, Justin, for your response. It is highly appreciated. When I built Gromacs on both desktops, I recall doing the regression tests, though I do not recall the results of the test, my bad. Do you mean I should re-run the regression tests? Or are there other tests I can run?

Re: [gmx-users] Different Results with Different Platforms

Hi, On Tue, Oct 4, 2016 at 4:17 PM Kelechi Okoroafor wrote: > Thank you very much, Mark, for your prompt response; it is highly > appreciated. > My sincere apologies for the miscommunication in my initial post. > > I meant to say that in performing several repetitions of

Re: [gmx-users] Different Results with Different Platforms

Thank you very much, Mark, for your prompt response; it is highly appreciated. My sincere apologies for the miscommunication in my initial post. I meant to say that in performing several repetitions of simulations, after approximately 200ns of total simulation time on (Stampede and my school's)

Re: [gmx-users] Different Results with Different Platforms

On 10/4/16 10:16 AM, Kelechi Okoroafor wrote: Thank you very much, Mark, for your prompt response; it is highly appreciated. My sincere apologies for the miscommunication in my initial post. I meant to say that in performing several repetitions of simulations, after approximately 200ns of

Re: [gmx-users] Different Results with Different Platforms

Hi, This sounds totally normal. Individual MD trajectories are only reproducible if you take great pains to make them so, and that is typically not a relevant objective (and definitely not implemented for GROMACS on GPUs). If you're trying to do an ergodic sampling, then that's the thing you want

[gmx-users] Different Results with Different Platforms

Hello, All. I had previously run the same coarse-grained MD simulations using the Martini force field with PME electrostatics on TACC's Stampede GPU cluster and my school's CPU cluster and gotten the same results. I recently built two desktop computers and the results from both my computers