Re: [AMBER-Developers] Is anyone seeing bad behavior in trying to download Miniconda?

From: Scott Brozell via AMBER-Developers <amber-developers.ambermd.org>
Date: Mon, 20 Feb 2023 20:14:56 -0500

Hi,

Starting at 14:09, 15:03, 16:37, and 19:37 today, i ran some trials of
./configure_cmake.py -bw -v > & out &

Two and a quarter minutes was the median wall time:

label & trials & min & max & mean & std_dev & median & mode \\\
wall,sec & 13 & 130.18 & 154.3 & 138.79 & 7.46 & 135.23 & None \\\


50.802u 18.386s 2:22.20 48.6% 0k shared+0k 5656+3111104io 0pf+0w 794860k max mem
49.068u 19.442s 2:19.44 49.1% 0k shared+0k 1544+3093272io 0pf+0w 780256k max mem
47.990u 17.798s 2:10.18 50.5% 0k shared+0k 1544+3093296io 0pf+0w 783216k max mem
48.409u 18.971s 2:14.84 49.9% 0k shared+0k 1544+3093312io 0pf+0w 782284k max mem

48.290u 17.193s 2:30.70 43.4% 0k shared+0k 1544+3093224io 0pf+0w 780584k max mem
49.236u 19.828s 2:21.47 48.8% 0k shared+0k 1544+3093248io 0pf+0w 778348k max mem
48.237u 19.003s 2:10.88 51.3% 0k shared+0k 1544+3093240io 0pf+0w 781136k max mem

48.386u 19.077s 2:13.84 50.3% 0k shared+0k 1544+3093240io 0pf+0w 786392k max mem
48.435u 16.687s 2:11.66 49.4% 0k shared+0k 1544+3093264io 0pf+0w 781716k max mem
50.107u 19.642s 2:26.08 47.7% 0k shared+0k 2072+3093256io 3pf+0w 781764k max mem

48.780u 20.179s 2:34.30 44.6% 0k shared+0k 1544+3093272io 0pf+0w 776920k max mem
48.714u 20.464s 2:13.49 51.8% 0k shared+0k 1544+3093256io 0pf+0w 781996k max mem
50.281u 17.317s 2:15.23 49.9% 0k shared+0k 1544+3093208io 0pf+0w 781516k max mem

I use the super awesome tcsh interactively, which i have set to
automagically time every command that's more than 2 seconds:
time (2 %Uu %Ss %E %P %Xk shared+%Dk %I+%Oio %Fpf+%Ww %Mk max mem)

scott

On Mon, Feb 20, 2023 at 02:03:13PM -0500, Scott Brozell via AMBER-Developers wrote:
> Hi,
>
> I have been building routinely and have not noticed any slow downs.
> The configure step takes 3 minutes and change on the clusters i use
> mostly. On our newest cluster it is occasionally about 2 minutes.
> I don't recall a cmake error termination in 2023.
>
> Although I never use run_cmake unless i am beta testing, i am always
> using -DDOWNLOAD_MINICONDA=TRUE:
>
> Starting configure_cmake.py on Sat Feb 18 21:40:09 2023 via:
> ./configure_cmake.py -bw -v --mpi --cuda
>
> ...
> Running cmake command:
> cmake /tmp/amber -DCMAKE_INSTALL_PREFIX=/tmp/amber22 -DCOMPILER=GNU -DMPI=TRUE -DCUDA=TRUE -DOPENMP=FALSE -DBUILD_GUI=TRUE -DINSTALL_TESTS=TRUE -DBUILD_PYTHON=TRUE -DDOWNLOAD_MINICONDA=TRUE -DCHECK_UPDATES=TRUE -DAPPLY_UPDATES=FALSE -DCMAKE_BUILD_TYPE=Release -DCOLOR_CMAKE_MESSAGES=FALSE -DCMAKE_VERBOSE_MAKEFILE=TRUE
>
>
> Red Hat Enterprise Linux Server release 7.9 (Maipo)
> /usr/bin/python
> Python 2.7.5
>
> Red Hat Enterprise Linux release 8.4 (Ootpa)
> /usr/bin/python
> Python 3.6.8
>
> scott
>
> On Mon, Feb 20, 2023 at 10:18:07AM -0500, Daniel Roe via AMBER-Developers wrote:
> > Cmake eventually succeeded. I didnt try the build yet though.
> >
> > On Mon, Feb 20, 2023 at 10:00 AM David A Case via AMBER-Developers <
> > amber-developers.ambermd.org> wrote:
> >
> > > On Mon, Feb 20, 2023, Dan Roe wrote:
> > > >
> > > >Just to check, I ran cmake (similar invocation except for the install
> > > >dirs) for Amber (GIT master) on LoBoS (Centos 7.9).
> > > >
> > > >[droe.tesla2 ~]$ which python
> > > >/usr/bin/python
> > > >[droe.tesla2 ~]$ python --version
> > > >Python 2.7.5
> > > >
> > > >This time it took much longer (5m53.667s), hanging on the
> > > >miniconda/python setup steps.
> > >
> > > Thanks for the info. Did things eventually succeed, or in the end (after 6
> > > minutes), get a cmake error?

_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Mon Feb 20 2023 - 17:30:02 PST
Custom Search