VMD-L Mailing List

From: John Stone (johns_at_ks.uiuc.edu)
Date: Thu Dec 18 2014 - 15:38:11 CST

Nikolai,
  I would expect all of the high-end versions of the Kepler and Maxwell
NVIDIA GPUs would likely be workable, and I would similarly expect that
recent high-end AMD/ATI cards would work, although I haven't actually tested
any yet. I don't know what the maximum display resolutions are for Intel's
integrated graphics, but I expect that integrated GPUs would not have
sufficient graphics performance to be usable on high resolution displays like
these even if they could theoretically support the display resolution for
easier things like playing movies. I would say that in order to drive a
4K display well, one would want to obtain one of the latest GPUs in the
~$250 to $500 price range, with the GTX 980 being one such example. You
can go for higher-end cards like Quadro cards to get much more GPU memory,
but that would be for reasons other than just due to the 4K display
resolution, e.g. for rendering large/complex VMD scenes using the
interactive GPU ray tracing engine, etc. If you've got the budget for
a 4K TV, it would not make much sense to skimp on the GPU to drive it
so I wouldn't go for a low-end GPU in any case.

Cheers,
  John Stone
  vmd_at_ks.uiuc.edu

On Thu, Dec 18, 2014 at 03:13:43PM -0600, Smolin, Nikolai wrote:
> Hi,
>
> which other graphic cards besides GeForce 980 you recommend for this type
> of TVs (3D 4k)?
>
> Thanks,
>
> Nikolai
> ---------------------------------------------------------------------------------------------------------
> Nikolai Smolin, Ph.D.
> Research Associate
> Department of Cell and Molecular Physiology
> Loyola University Chicago
> 2160 South First Avenue
> Maywood, IL 60153
> Phone : 708 - 216 - 5158
> E-mail : [1]nikolai.smolin_at_gmail.com
> E-mail : [2]nsmolin_at_luc.edu
> www : [3]https://sites.google.com/site/nikolaismolin/
> On Thu, Dec 18, 2014 at 2:31 PM, John Stone <[4]johns_at_ks.uiuc.edu> wrote:
>
> Hi,
> A I wanted to send a quick update.A After updating video drivers and
> disabling the "deep color" display mode and upgrading the TV firmware
> on the LG 4K TV I've been testing, I was able to get row-interleaved
> stereo working with VMD on a GeForce 980, at full 4K resolution and
> without
> any color bleeding from chrominance downsampling.A I've been testing
> the
> LG model 49UB8500-UA, but I believe that the same results would likely
> apply to their larger 55", 60", etc models.
>
> Here are my observations about getting the TV working with VMD:
> A - If you drive the TV at native 4K resolution (3840x2160), and you
> have
> A A an up-to-date video driver and a GPU that can handle the full
> resolution
> A A without chrominance downsampling, you can simply enable
> "row-interleaved"
> A A stereo mode in VMD (nothing has to be done with the TV itself) and
> you'll
> A A get 3-D (you may have to enable VMD 3-D eye swap, particularly if
> the VMD
> A A OpenGL window gets moved around).
>
> A - If you drive the TV at any resolution _below_ the native 4K hardware
> A A resolution (e.g. 1080p or less), the TV's built-in 3-D stereo
> decoder
> A A mode MUST be activated, and then you should set VMD's stereo mode
> to
> A A match whatever the TV's built-in stereo 3-D decoder is set to.
>
> So far, we have driven the TV at full 4K resolution and stereo on Linux,
> but we have not (yet) found a Mac laptop that is able to drive the TV
> over HDMI at full 4K resolution.A All of the Macs we have tried have
> all maxed out at 1080p thus far.A We haven't yet tried Windows laptops
> yet.
> The various stereo modes worked fine with the Mac laptops, albeit at
> only 1080p resolution.
>
> One remaining issue I have observed when using stereo 3-D with the TV
> is some apparent flicker during constantly-updating VMD rotation draws,
> and an apparent lack of vertical retrace synchronization with the GPU.
> So far this effect is relatively minor, but it is worth noting.
> I'm looking into whether this is some video driver bug or something
> else.A I'll send further updates when I get it sorted out.
>
> Cheers,
> A John Stone
> A [5]vmd_at_ks.uiuc.edu
> On Wed, Dec 17, 2014 at 10:45:24AM -0600, John Stone wrote:
> > Hi,
> >A A I think you missed my point -- row-interleaved hasn't worked for
> me
> > when running at 4K resolution, only side-by-side and over-under have
> worked
> > so far when running at the full display resolution.A Row interleaved
> is one
> > of the modes that could have issues if you've got downsampled
> > chrominance (in my case the LG TV just disables these modes entirely).
> >
> > For the LG TVs I'm testing currently, the display panel appears to be
> > row-interleaved at the hardware level, so running row-interleaved is
> the
> > best choice for VMD, but I have to find a way to get the display to
> run
> > at full 4K resolution with full resolution chrominance data.A I'm
> going to
> > try updating to a newer driver version and newer TV firmware and see
> if
> > either of those has any impact.A Another potential gotcha I'm working
> > around is that some TVs support 30-bit color (10-bit per color
> channel)
> > and that eats some of the available HDMI bandwidth.A It is possible
> that
> > row, column, and checkerboard stereo modes might be workable on the
> > shallow color depth modes but not in 30-bit color.A I'm working on
> testing
> > that next...
> >
> > Cheers,
> >A A John
> >
> > On Wed, Dec 17, 2014 at 05:34:14PM +0100, Norman Geist wrote:
> > > But isn't more comfortable using the row interleaved stereo mode
> anyway,
> > > where one hasn't to deal with all of that possible issues? It will
> have full
> > > screen stereo, windowed stereo, and the TV doesn't even know that 3D
> content
> > > is present. The only disturbing thing is the eye switch which comes
> by
> > > moving and placing the windows.
> > >
> > > Or do I miss something?
> > >
> > > Norman Geist.
> > >
> > >
> > > > -----UrsprA 1/4ngliche Nachricht-----
> > > > Von: John Stone [mailto:[6]johns_at_ks.uiuc.edu]
> > > > Gesendet: Mittwoch, 17. Dezember 2014 17:12
> > > > An: Norman Geist
> > > > Cc: 'Smolin, Nikolai'; VMD Mailing List
> > > > Betreff: Re: vmd-l: VMD with 3D TV
> > > >
> > > > Hi,
> > > >A A One interesting observation I had in testing VMD with some of
> the new
> > > > LG TVs is that although the LG 4K TV supports 6 different stereo
> modes,
> > > > only some of them are usable when the video driver is set for the
> > > > native
> > > > 4K display resolution of the TV, at least in my testing with a GTX
> 980
> > > > and driver version 343.22.A One potential catch with this is that
> video
> > > > cards using HDMI 1.x outputs may downsample the color information
> by 2x
> > > > or more, converting to so-called 4:2:2 or 4:2:0 color formats.
> > > >
> > > > I think that the downsampled color formats are inherently
> unworkable
> > > > with
> > > > row-interleaved, column-interleaved, and checkerboard stereo
> formats,
> > > > which in the case of the LG TVs I've been testing leaves only the
> > > > side-by-side and over-under video format.A The drawback of the
> > > > side-by-side and over-under formats is that they only work
> correctly
> > > > with a full-screen window, and clearly that is somewhat limiting.
> > > >
> > > > I'm currently trying to sort out a video driver version that can
> run
> > > > the TV without any downsampling of chrominance information, and
> I'm
> > > > hoping
> > > > that might allow the LG TVs I'm testing to run at full 4K
> resolution
> > > > while also supporting windowed stereo.A If you drive the LG TVs
> at
> > > > 1080p,
> > > > they do support all of the stereo modes, which is one reason I
> suspect
> > > > that
> > > > downsampled chrominance might causing the limitation in 4K modes.
> > > >
> > > > Cheers,
> > > >A A John Stone
> > > >A A [7]vmd_at_ks.uiuc.edu
> > > >
> > > > On Mon, Dec 15, 2014 at 08:43:12AM +0100, Norman Geist wrote:
> > > > >A A VMD also offers the stereo mode row Interleaved, so you can
> use
> > > > any
> > > > >A A passive 3D Display. The graphics card or machine doesn't
> really
> > > > matter
> > > > >A A than, except of supporting the 4k resolution.
> > > > >
> > > > >
> > > > >
> > > > >A A Norman Geist.
> > > > >
> > > > >
> > > > >
> > > > >A A Von: [8]owner-vmd-l_at_ks.uiuc.edu
> [mailto:[9]owner-vmd-l_at_ks.uiuc.edu] Im
> > > > Auftrag
> > > > >A A von Smolin, Nikolai
> > > > >A A Gesendet: Freitag, 12. Dezember 2014 17:10
> > > > >A A An: [10]vmd-l_at_ks.uiuc.edu
> > > > >A A Betreff: vmd-l: VMD with 3D TV
> > > > >
> > > > >
> > > > >
> > > > >A A Dear All,
> > > > >
> > > > >A A I am thinking to use VMD with 4k 3D TV.
> > > > >
> > > > >A A What is the best way to do it.
> > > > >
> > > > >A A Linux or Windows? which graphic card? Which TV is best for
> this?
> > > > >
> > > > >A A Now I have a Suse 13.1 with QuadroA 2000D Workstation
> Graphics
> > > > Card -
> > > > >A A 1024MB, GDDR5, PCI Express 2.0 x16, Dual DVI, DirectX 11,
> Single-
> > > > Slot.
> > > > >
> > > > >A A Can I use this workstation with 4k 3D TV.
> > > > >
> > > > >
> > > > >
> > > > >A A Thanks,
> > > > >
> > > > >A A Nikolai
> > > > >
> > > > >A A
> ------------------------------------------------------------------
> > > > ---------------------------------------
> > > > >A A Nikolai Smolin, Ph.D.
> > > > >A A Research Associate
> > > > >
> > > > >A A Department of Cell and Molecular Physiology
> > > > >
> > > > >A A Loyola University Chicago
> > > > >
> > > > >A A 2160 South First Avenue
> > > > >
> > > > >A A Maywood, IL 60153
> > > > >A A Phone : 708 - 216 - 5158
> > > > >
> > > > >A A E-mail : [1][11]nikolai.smolin_at_gmail.com
> > > > >A A E-mail : [2][12]nsmolin_at_luc.edu
> > > > >A A www : [3][13]https://sites.google.com/site/nikolaismolin/
> > > > >
> > > > > References
> > > > >
> > > > >A A Visible links
> > > > >A A 1. mailto:[14]nikolai.smolin_at_gmail.com
> > > > >A A 2. mailto:[15]nsmolin_at_luc.edu
> > > > >A A 3. [16]https://sites.google.com/site/nikolaismolin/
> > > >
> > > > --
> > > > NIH Center for Macromolecular Modeling and Bioinformatics
> > > > Beckman Institute for Advanced Science and Technology
> > > > University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
> > > > [17]http://www.ks.uiuc.edu/~johns/A A A A A A Phone:
> [18]217-244-3349
> > > > [19]http://www.ks.uiuc.edu/Research/vmd/
> > >
> >
> > --
> > 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/A A A A A A Phone:
> [21]217-244-3349
> > [22]http://www.ks.uiuc.edu/Research/vmd/
>
> --
> NIH Center for Macromolecular Modeling and Bioinformatics
> Beckman Institute for Advanced Science and Technology
> University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
> [23]http://www.ks.uiuc.edu/~johns/A A A A A A Phone:
> [24]217-244-3349
> [25]http://www.ks.uiuc.edu/Research/vmd/
>
> References
>
> Visible links
> 1. mailto:nikolai.smolin_at_gmail.com
> 2. mailto:nsmolin_at_luc.edu
> 3. https://sites.google.com/site/nikolaismolin/
> 4. mailto:johns_at_ks.uiuc.edu
> 5. mailto:vmd_at_ks.uiuc.edu
> 6. mailto:johns_at_ks.uiuc.edu
> 7. mailto:vmd_at_ks.uiuc.edu
> 8. mailto:owner-vmd-l_at_ks.uiuc.edu
> 9. mailto:owner-vmd-l_at_ks.uiuc.edu
> 10. mailto:vmd-l_at_ks.uiuc.edu
> 11. mailto:nikolai.smolin_at_gmail.com
> 12. mailto:nsmolin_at_luc.edu
> 13. https://sites.google.com/site/nikolaismolin/
> 14. mailto:nikolai.smolin_at_gmail.com
> 15. mailto:nsmolin_at_luc.edu
> 16. https://sites.google.com/site/nikolaismolin/
> 17. http://www.ks.uiuc.edu/~johns/
> 18. tel:217-244-3349
> 19. http://www.ks.uiuc.edu/Research/vmd/
> 20. http://www.ks.uiuc.edu/~johns/
> 21. tel:217-244-3349
> 22. http://www.ks.uiuc.edu/Research/vmd/
> 23. http://www.ks.uiuc.edu/~johns/
> 24. tel:217-244-3349
> 25. http://www.ks.uiuc.edu/Research/vmd/ 

-- 
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/