Hi Adam,

Thanks for expressing your concerns.

This message,

>     Note: Radeon HD cards tend not to run PyMOL well.
>     Use nVidia or Intel instead, if OpenGL glitches occur.
>     Attempting to compensate for known issues...

has been in PyMOL for a long time.  It's just a fact that NVidia cards
have been less buggy and have better Linux drivers than ATI and Intel.
 The heterogeneity in (buggy) hardware makes development difficult and
expensive.  And, we're not the only ones running into this either: the
CCP4 list recently discussed this, and I've also been talking to John
Stone, the main VMD developer, who also faces such issues.


> So the new Macs only come with Radeon video cards... there are no other
> choices with a Mac. This means the MacPros no are no longer a good platform
> for running PyMOL? I just tossed $3K down the drain?

The new MacPros should be more than capable of running PyMOL.  In
fact, the above warning is more for low-end cards which are buggier,
not the high-end ones found in MacPros.  Your $3k wasn't thrown away,
you're just ahead of the curve.  I'm sure more people will buy
MacPros, forcing me to follow suit.  Also, if Apple hasn't already, I
hope they'll release a list of NVidia cards which will run on their
new MacPros, just like they have a list of Radeon cards for the
NVidia-based MacPros.


> I tried to render the surface of 3L5Q and PyMOL just crashes (6GB RAM). I
> tried to render just the top half and it still crashes.

PyMOL is known to crash when trying to render large structures.  This
is a legacy issue with the internal memory management, not your Radeon
card.  3L5Q is a very large PDB (79,452 atoms w/o hydrogens).  I agree
wholeheartedly that PyMOL should be able to render this without
crashing.  Even my machine (early 2010 Dual Quad-core MacPro w/16GB
RAM) cannot even "set surface_quality, 1" on that structure without
crashing.


> What are the OpenGL glitches that occur, are they documented somewhere?

The glitches we talk about are those like:
  * blank windows or windows not opening/closing properly (Intel GM945)
  * sticks not coloring properly (Intel945, some ATI cards, some VMs);
they're all blue when they should be mutlicolored
  * partial geometry being rendered;


> What does PyMOL do to compensate?

We detect the video cards and add extra openGL calls (usually glFlush
or glFinish) or reorder openGL calls to workaround known problems.


> I checked the PyMOL web site and under the supported operating systems and
> hardware there is no mention of the lack of Radeon support/poor Radeon
> performance. Maybe that should be added because I wouldn't have bought this
> new Mac if I would have known this. Is there anything I can do to make this
> new machine work?

We cannot list all possible systems on which PyMOL won't run.  We do
try to provide a list of machines on which it will run.  Aside from
not being able to ray trace a protein that's almost big enough to see
with the naked eye, do you see any other performance issues or
glitches or artifacts?  Could it be that PyMOL is in fact running okay
on your machine but the memory-bug combined with the Radeon warning
caught you off guard?  If PyMOL produces other bugs on your system due
to the Radeon card or not, please email me a bug report or better yet
visit http://pymol.org/contact (choose "Bug Reporting") and file it
there so it's automatically tracked internally.


> I see a 64-bit version is on the horizon, will this address Radeon
> performance or should I go back to my old MacPro with it's nVidia 7300 video
> cards?

We wanted to release a 64-bit version for a while, but the Tcl/Tk
toolkit was not ready.  I think it's still in alpha/beta.  64-bit
MacPyMOL will be released when we're sure that works.

Cheers,

-- Jason

-- 
Jason Vertrees, PhD
PyMOL Product Manager
Schrodinger, LLC

(e) jason.vertr...@schrodinger.com
(o) +1 (603) 374-7120

------------------------------------------------------------------------------
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev
_______________________________________________
PyMOL-users mailing list (PyMOL-users@lists.sourceforge.net)
Info Page: https://lists.sourceforge.net/lists/listinfo/pymol-users
Archives: http://www.mail-archive.com/pymol-users@lists.sourceforge.net

Reply via email to