VMD-L Mailing List

From: Kevin C Chan (cchan2242-c_at_my.cityu.edu.hk)
Date: Tue Jul 05 2016 - 00:30:29 CDT

Thank you Josh and John for your great replies.

It's my fault misleading myself to type "make". Sorry :(

One quick follow-up question. As launching VMD always tells that it has
detected all the cores on the node (16 in my case), I am wondering how many
CPUs it is actually using when doing such rendering. I am planning to
submit this kind of rendering jobs into a node having some CPU-only
calculation jobs occupying maybe 8 cores of the node, is it feasible? Or I
have to specify number of cores when launching VMD? Or would you suggest
that users should never run different jobs on one single node at the same
time? Reserving one full node just for rendering is just sometimes too
expensive for us.

Thanks,
Kevin

On Fri, Jul 1, 2016 at 3:57 AM, John Stone <johns_at_ks.uiuc.edu> wrote:

> Kevin,
> Yes, if you read the "README" file that comes with the distribution or
> the installation instructions on the VMD web site, it specifies that you
> should be running "make install" to install the pre-built binaries.
> We ship VMD binaries with all of the library dependencies they need, so
> it should not be necessary to go looking for them under normal
> circumstances.
>
> Regarding rendering performance, transparent geometry always requires more
> work from a renderer than opaque geometry does. The particular performance
> impact depends on a large number of factors that are too numerous to list,
> but the 3x runtime increase you mention sounds plausible. Larger runtimes
> the one you list are usually associated with scenes that contain
> multiple overlapping transparent surfaces.
>
> Cheers,
> John Stone
> vmd_at_ks.uiuc.edu
>
> On Thu, Jun 30, 2016 at 10:16:51PM +0800, Kevin C Chan wrote:
> > Good to know the libraries were shipped with the downloaded file. Yes
> I
> > found them under lib/redistrib/lib_LINUXAMD64. After manually added
> the
> > path into LD_LIBRARY_PATH, theA vmd_LINUXAMD64 recognized it.A
> > Oh, wait, interesting. I haveA successfully installed it by typing
> "make
> > install" instead of "make" in src/. Then even if the vmd_LINUXAMD64
> still
> > show not finding any OptiX lib, it would NOT affect the launch of the
> > wrapped binary underA $install_bin_dir specified in configure.
> > Problem solved. Thanks all of you for your valuable experience shared.
> > Benchmark of my rendering of a ~100k atoms VDW representations was
> ~25s
> > per snap and ~90s if transparent presents (e.g. GlassBubble). Is it
> > normal?
> > Kevin
> > On Thu, Jun 30, 2016 at 8:57 PM, Vermaas, Joshua
> > <[1]Joshua.Vermaas_at_nrel.gov> wrote:
> >
> > Hi Kevin,
> >
> > Those two libraries are going to be the big problem, as they are the
> > optix library that you need for the super-mega-awesome GPU ray
> tracer. I
> > could have sworn that the binary download includes them. Have you
> looked
> > there? Maybe in the lib directory?
> >
> > Josh Vermaas
> > Directora**s Postdoctoral Fellow
> > National Renewable Energy Laboratory
> > [2]joshua.vermaas_at_nrel.gov<mailto:[3]joshua.vermaas_at_nrel.gov>
> >
> > On Jun 29, 2016, at 9:23 PM, Kevin C Chan
> > <[4]cchan2242-c_at_my.cityu.edu.hk<mailto:[5]
> cchan2242-c_at_my.cityu.edu.hk>>
> > wrote:
> >
> > a**Thank you John for your quick reply.
> >
> > I am using CentOS 6.6 (a**actually it is a Rocks OS). Well yes I
> could
> > get the directly downloaded Text-only binary work (which is the
> binary I
> > am using in most of the time), it seems compatible in that case.
> However
> > the directly downloaded "LINUX_64 OpenGL, CUDA, TachyonL-OptiX"
> binary
> > give two problems:
> > 1) $ ldd LINUXAMD64/vmd_LINUXAMD64
> > A A A A ...
> > liboptixu.so.1 => not found
> > liboptix.so.1 => not found
> > A A A A ...
> > A A (which I have no idea where to find them.)
> > 2) $ cd src/
> > A A $ make
> > A A make: *** No rule to make target `y.tab.h', needed by
> > `vmd_LINUXAMD64'.A Stop.
> > A A (which did not happen for Text-only binary, although they
> seems to
> > have the same dependency.)
> >
> > Thanks,
> > Kevin
> >
> > On Wed, Jun 29, 2016 at 10:29 PM, John Stone
> > <[6]johns_at_ks.uiuc.edu<mailto:[7]johns_at_ks.uiuc.edu>> wrote:
> > Kevin,
> > A What version of CentOS are you running?A The OptiX-enabled
> binaries
> > of VMD 1.9.2 should work fine with CentOS 5.x and later.A You
> should
> > not
> > need to recompile from source, and in general I would not recommend
> it
> > unless
> > you are an expert in software development due to the need for all
> of the
> > toolkit and GPU library versions associated with CUDA and OptiX and
> so
> > on to
> > match up precisely.A What do you get when you run these commands
> > on your system?:
> > A uname -a
> > A cat /etc/issue
> >
> > Cheers,
> > A John Stone
> > A [8]vmd_at_ks.uiuc.edu<mailto:[9]vmd_at_ks.uiuc.edu>
> > On Wed, Jun 29, 2016 at 10:14:01PM +0800, Kevin C Chan wrote:
> > > Thanks so much for the reply. That's what I am asking for.
> > >
> > > However when I want to try out Fernando's script, I found that I
> could
> > not
> > > get a VMD binary with TachyonOptiX to work. I thought it was
> built-in,
> > but
> > > the RHEL binary seems not compatible with my CentOS, saying
> "make: ***
> > No
> > > rule to make target `y.tab.h', needed by `vmd_LINUXAMD64'.A
> Stop."
> > when I
> > > typed make.
> > >
> > > Also it is surprisingly difficult to find a compilation guide for
> VMD.
> > >
> > > I am wondering whether I could obtain a binary
> TachyonOptiXInternal
> > and
> > > execute it using a Text-only VMD on the cluster/nodes.
> > >
> > > Thanks,
> > > Kevin
> > >
> > > On Wed, Jun 29, 2016 at 12:16 PM, Fernando Vallejos-Burgos <
> > > [10]nandobike_at_gmail.com<mailto:[11]nandobike_at_gmail.com>> wrote:
> > >
> > > > Hi Kevin,
> > > >
> > > > I use a very simple TCL script for rendering animations
> remotely.
> > > > It goes something like this, for example, for 1000 frames
> rendering
> > in
> > > > TachyonOptiX after you set the viewpoint:
> > > >
> > > > display shadows on
> > > > display resize 1280 720
> > > > display ambientocclusion on
> > > >A A A A A display aoambient 0.80
> > > > display aodirect 0.50
> > > > for {set i 0} {$i < 1000} {incr i 1} {
> > > > display update
> > > > animate goto $i
> > > > set filename ./snap_[format "%04d" [expr $i]].tga
> > > > render aasamples TachyonLOptiXInternal 256
> > > > render TachyonLOptiXInternal $filename
> > > > }
> > > >
> > > > then I just source the TCL file and call this procedure.
> > > > For me, it runs quite fast in GPU.
> > > >
> > > > Hope this can help!
> > > >
> > > > Fernando
> > > >
> > > >
> > > > -------
> > > > Dr. Fernando Vallejos-Burgos
> > > > Researcher at Center for Energy and Environmental Science
> > > > Shinshu University, Wakasato, Nagano 380-8553, Japan
> > > > Personal web:
> > [12]http://www.vallejos.cl<[13]
> https://urldefense.proofpoint.com/v2/url?u=http-3A__www.vallejos.cl&d=CwMFaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=zFfoK61upjM5BwyoRAsX8dLq7rwWm8aw7r7dqtjgcCE&m=asS68yK4w_ytuTVQXO_fIxmAd0602DA8LBoIXQasLx8&s=Ie4131uuh8MC6GEcM3P31yG8cOxaeyLuptC_C5a4Ey8&e=
> >
> > > > Labo web:
> > [14]http://www.shinshu-u.ac.jp/faculty/engineering/kaneko-group/
> <[15]
> https://urldefense.proofpoint.com/v2/url?u=http-3A__www.shinshu-2Du.ac.jp_faculty_engineering_kaneko-2Dgroup_&d=CwMFaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=zFfoK61upjM5BwyoRAsX8dLq7rwWm8aw7r7dqtjgcCE&m=asS68yK4w_ytuTVQXO_fIxmAd0602DA8LBoIXQasLx8&s=vw2oQxW0Gmm2EbpmUYcD0J8cjALCvJecYNvNx6pNzOQ&e=
> >
> > > >
> > > >
> > > >
> > > > On Jun 29, 2016, at 12:14 PM, Kevin C Chan
> > <[16]cchan2242-c_at_my.cityu.edu.hk<mailto:[17]
> cchan2242-c_at_my.cityu.edu.hk>>
> > > > wrote:
> > > >
> > > > ???Thanks for your reply. ???I am wondering what pseudo code
> would
> > it be like.
> > > > Simply
> > > >
> > > > source visualization state
> > > > tachyon_binary ...
> > > >
> > > > ? As I used to do it through GUI, I have to
> > > > 1) tune the visual angle
> > > > 2) turn on AO
> > > > 3) render > tachyon/tachyon_internal > type BMP as extension
> > > >
> > > > How could all these be done using a script?
> > > >
> > > > Thanks,
> > > > Kevin
> > > >
> > > > On Wed, Jun 29, 2016 at 12:34 AM, John Stone
> > <[18]johns_at_ks.uiuc.edu<mailto:[19]johns_at_ks.uiuc.edu>> wrote:
> > > >
> > > > Hi,
> > > >A Hi, sorry for the slow reply, I just returned from
> international
> > travel
> > > >
> > > >
> > > >
> >
> > --
> > NIH Center for Macromolecular Modeling and Bioinformatics
> > Beckman Institute for Advanced Science and Technology
> > University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
> > [20]http://www.ks.uiuc.edu/~johns/<[21]
> https://urldefense.proofpoint.com/v2/url?u=http-3A__www.ks.uiuc.edu_-7Ejohns_&d=CwMFaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=zFfoK61upjM5BwyoRAsX8dLq7rwWm8aw7r7dqtjgcCE&m=asS68yK4w_ytuTVQXO_fIxmAd0602DA8LBoIXQasLx8&s=s0vyfiUCPXCkWuw7RWRv4H4_IIMQOyfDdIbdl_MdCHo&e=
> >A
> > A A A A A Phone: [22]217-244-3349<tel:[23]217-244-3349>
> > [24]http://www.ks.uiuc.edu/Research/vmd/<[25]
> https://urldefense.proofpoint.com/v2/url?u=http-3A__www.ks.uiuc.edu_Research_vmd_&d=CwMFaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=zFfoK61upjM5BwyoRAsX8dLq7rwWm8aw7r7dqtjgcCE&m=asS68yK4w_ytuTVQXO_fIxmAd0602DA8LBoIXQasLx8&s=w--C1hB0JcC3BKg2uI7x1tlH3hTN2sE3lmfpyrYKU2U&e=
> >
> >
> > References
> >
> > Visible links
> > 1. mailto:Joshua.Vermaas_at_nrel.gov
> > 2. mailto:joshua.vermaas_at_nrel.gov
> > 3. mailto:joshua.vermaas_at_nrel.gov
> > 4. mailto:cchan2242-c_at_my.cityu.edu.hk
> > 5. mailto:cchan2242-c_at_my.cityu.edu.hk
> > 6. mailto:johns_at_ks.uiuc.edu
> > 7. mailto:johns_at_ks.uiuc.edu
> > 8. mailto:vmd_at_ks.uiuc.edu
> > 9. mailto:vmd_at_ks.uiuc.edu
> > 10. mailto:nandobike_at_gmail.com
> > 11. mailto:nandobike_at_gmail.com
> > 12. http://www.vallejos.cl/
> > 13.
> https://urldefense.proofpoint.com/v2/url?u=http-3A__www.vallejos.cl&d=CwMFaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=zFfoK61upjM5BwyoRAsX8dLq7rwWm8aw7r7dqtjgcCE&m=asS68yK4w_ytuTVQXO_fIxmAd0602DA8LBoIXQasLx8&s=Ie4131uuh8MC6GEcM3P31yG8cOxaeyLuptC_C5a4Ey8&e=
> > 14. http://www.shinshu-u.ac.jp/faculty/engineering/kaneko-group/
> > 15.
> https://urldefense.proofpoint.com/v2/url?u=http-3A__www.shinshu-2Du.ac.jp_faculty_engineering_kaneko-2Dgroup_&d=CwMFaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=zFfoK61upjM5BwyoRAsX8dLq7rwWm8aw7r7dqtjgcCE&m=asS68yK4w_ytuTVQXO_fIxmAd0602DA8LBoIXQasLx8&s=vw2oQxW0Gmm2EbpmUYcD0J8cjALCvJecYNvNx6pNzOQ&e=
> > 16. mailto:cchan2242-c_at_my.cityu.edu.hk
> > 17. mailto:cchan2242-c_at_my.cityu.edu.hk
> > 18. mailto:johns_at_ks.uiuc.edu
> > 19. mailto:johns_at_ks.uiuc.edu
> > 20. http://www.ks.uiuc.edu/~johns/
> > 21.
> https://urldefense.proofpoint.com/v2/url?u=http-3A__www.ks.uiuc.edu_-7Ejohns_&d=CwMFaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=zFfoK61upjM5BwyoRAsX8dLq7rwWm8aw7r7dqtjgcCE&m=asS68yK4w_ytuTVQXO_fIxmAd0602DA8LBoIXQasLx8&s=s0vyfiUCPXCkWuw7RWRv4H4_IIMQOyfDdIbdl_MdCHo&e=
> > 22. tel:217-244-3349
> > 23. tel:217-244-3349
> > 24. http://www.ks.uiuc.edu/Research/vmd/
> > 25.
> https://urldefense.proofpoint.com/v2/url?u=http-3A__www.ks.uiuc.edu_Research_vmd_&d=CwMFaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=zFfoK61upjM5BwyoRAsX8dLq7rwWm8aw7r7dqtjgcCE&m=asS68yK4w_ytuTVQXO_fIxmAd0602DA8LBoIXQasLx8&s=w--C1hB0JcC3BKg2uI7x1tlH3hTN2sE3lmfpyrYKU2U&e=
>
> --
> NIH Center for Macromolecular Modeling and Bioinformatics
> Beckman Institute for Advanced Science and Technology
> University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
> http://www.ks.uiuc.edu/~johns/ Phone: 217-244-3349
> http://www.ks.uiuc.edu/Research/vmd/
>