VMD-L Mailing List
From: Gianluca Interlandi (gianluca_at_u.washington.edu)
Date: Fri Oct 31 2014 - 15:59:46 CDT
- Next message: Josh Vermaas: "Re: psfgen N-terminal patches"
- Previous message: Ashar Malik: "Re: psfgen N-terminal patches"
- In reply to: Josh Vermaas: "Re: Graphics card suitable for VMD"
- Next in thread: Douglas Houston: "Interactive Display of Molecular Orbitals"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
> The quadro you listed is more than enough for running VMD. Personally
> I've been quite happy going with the "gaming" cards. A GTX 980 or 970
> will be cheaper with the same amount of RAM
Just to add my 2c. One thing to consider is also how much power the card
draws and whether the power supply of your desktop offers the needed
power. The other consideration is also heat and noise. I don't know the
GTX 980 or 970, but often the higher end graphics cards get hot and have
big cooling fans. Otherwise, I agree that the Quadro K4200 is more than
what is necessary to run VMD (I assume here just for visualization).
Maybe, Afroditi could benefit from selecting a cheaper card and spending
the difference in more CPU/RAM and/or a SSD drive? Please, take this with
a grain of salt, since we don't exactly know what else you are going to
use your desktop for.
Gianluca
(which is the figure of merit that determines the size of the system
> you can do the shiny new things John was advertising yesterday with). For the ~100k
> atom systems I work with frequently, I've noticed no problems with either my old GTX
> 580 or my newer 980 in driving a modest multi-monitor setup. In fact, I might suggest
> you better invest the price difference in some combination of more disk/RAM/CPU. This
> is mostly explained the features you are actually buying when you get a quadro:
>
> 1.) A single slot form factor, so you can use the adjacent PCI slot on the
> motherboard. Usually ignoreable, because most of the motherboards I've seen have
> their fast PCI express lanes separated in such a way that you are only blocking a
> "slow" PCI slot you wouldn't put a GPU on anyway.
> 2.) Lower power requirements. The quadros tend to have lower power draws relative to
> their gaming card counterparts, which means you can get away with a smaller power
> supply. You are buying/building the system from scratch, and can match the
> power-supply size accordingly.
> 3.) Error correcting memory. Useless for VMD purposes as far as I can tell. For
> running MD simulations themselves, useless as well, and may actually be harmful. See
> http://dx.doi.org/10.1145/2484762.2484774 for a comparison study that some of the
> Amber guys did in collaboration with NVIDIA.
> 4.) The Quadros/Teslas usually have better double-precision performance relative to
> the gaming counterparts (often by a substantial multiplicative factor). This isn't
> helpful to you though, since the developers of the MD packages with the largest
> "marketshare" for biomolecular simulation (NAMD, Amber, Gromacs) use single-precision
> in their GPU versions for the calculations.
> 5.) Better support contracts (ie, they exist). This is a big reason why large
> corporations/government agencies buy them. Not sure if that is something that is a
> factor at a single-user level though.
>
> Or for those that prefer a short video explanation (aimed at gamers):
> https://www.youtube.com/watch?v=U5Npt1BSF04
>
> Basically you shouldn't have a problem with the card you selected, but there may be a
> better fit for you in the product line.
>
> -Josh Vermaas
>
> On 10/31/2014 10:38 AM, Dudo wrote:
> well, the card costs only some thousand bucks so I am afraid in order to
> take full advantage of it you'll have to buy like 100 inch double full HD
> monitor or so, which does not come with the price :-(
>
> best,
> Dudo
>
> On Fri, Oct 31, 2014 at 3:46 PM, Afroditi Maria Zaki
> <afroditimaria.zaki_at_postgrad.manchester.ac.uk> wrote:
> Hi!
> I am working on molecular dynamics and I am planning to buy a new
> workstation . The graphics card I am thinking of buying is Nvidia Quadro
> K4200 4GB, so I would appreciate your opinion whether this one is
> suitable for running VMD.
> Thanks in advance for your consideration!
>
> Regards,
> Afroditi
>
>
>
>
> --
> ____________________
> Ing. Dusan Racko, PhD https://www.researchgate.net/profile/Dusan_Racko
> Polymer Institute of the Slovak Academy of Sciences
> Dubravska cesta 3
> 845 41 Bratislava, Slovak Republic
> tel: +421 2 3229 4321
>
>
>
>
-----------------------------------------------------
Gianluca Interlandi, PhD gianluca_at_u.washington.edu
+1 (206) 685 4435
http://artemide.bioeng.washington.edu/
Research Assistant Professor at the Department of Bioengineering
at the University of Washington, Seattle WA U.S.A.
-----------------------------------------------------
- Next message: Josh Vermaas: "Re: psfgen N-terminal patches"
- Previous message: Ashar Malik: "Re: psfgen N-terminal patches"
- In reply to: Josh Vermaas: "Re: Graphics card suitable for VMD"
- Next in thread: Douglas Houston: "Interactive Display of Molecular Orbitals"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]