[vtkusers] is Vtk 5.2.1 64 bit clean?

[David Cole]

Just a small clarification. The default value for VTK_USE_64BIT_IDS depends
on the type of binary you are building. For 32-bit binaries, it is OFF, for
64-bit binaries it is ON by default.

from VTK's CMakeLists.txt:
=================================
IF("${CMAKE_SIZEOF_VOID_P}" GREATER 4)
 SET(VTK_USE_64BIT_IDS_DEFAULT ON)
ELSE("${CMAKE_SIZEOF_VOID_P}" GREATER 4)
 SET(VTK_USE_64BIT_IDS_DEFAULT OFF)
ENDIF("${CMAKE_SIZEOF_VOID_P}" GREATER 4)


On Tue, Feb 17, 2009 at 9:10 AM, Berk Geveci <berk.geveci at kitware.com>wrote:

> >  Thorough testing of your extremely large data sets is well warranted.
> I'm
> > not trying to take anything away from VTK. I think it is a wonderful
> toolkit
> > that has allowed me to help people solve real problems. Just a word of
> > caution that if your data sets start getting large (into 64 bit
> territory)
> > be ready to possibly hunt down some bugs in the VTK code.
>
> I would strongly encourage developers and users to report any issues
> they find related to 64 bit issues when processing large datasets. We
> are actively hunting such bugs. I want to clarify a few things:
>
> * You will be in the "64 bit territory" when the number of points or
> cells in your datasets is bigger than 2^31 (we use signed ints). It is
> totally possible to have datasets much larger than 4 GiB without
> exceeding 2^31 points/cells.
>
> * Make sure to turn on VTK_USE_64BIT_IDS before compiling. This is OFF
> by default. Without setting VTK_USE_64BIT_IDS to on, you cannot
> process datasets that have more than 2^31 points/cells even on a 64
> bit machine. When VTK_USE_64BIT_IDS is on, VTK will use a 64 bit
> integer for indexing. Otherwise, it will use a 32 bit integer.
>
> -berk
>
>
> On Sat, Feb 14, 2009 at 11:16 PM, Michael Jackson
> <mike.jackson at bluequartz.net> wrote:
> > There are those issues and possible issues with 32 bit math results being
> > placed into a 64 bit value such as the following:
> >
> > int x, y, z;
> >
> > int64 total = x * y * z;
> >
> > on the surface that looks fine, but what happens when x * y * z goes
> beyond
> > the max signed int value? It rolls over and you get all sorts of weird
> > errors.
> >
> >  I found and fixed a few of these last year and others have gone on the
> "64
> > bit hunt" before but no one really sure just where the rest of the bugs
> are
> > at.
> >
> >  Thorough testing of your extremely large data sets is well warranted.
> I'm
> > not trying to take anything away from VTK. I think it is a wonderful
> toolkit
> > that has allowed me to help people solve real problems. Just a word of
> > caution that if your data sets start getting large (into 64 bit
> territory)
> > be ready to possibly hunt down some bugs in the VTK code.
> >
> > ---
> > Mike Jackson                 www.bluequartz.net
> >
> >
> >
> > On Feb 14, 2009, at 8:37 PM, shekharc wrote:
> >
> >> No, not atm. Perhaps that's why it works for me so far.
> >>
> >> Is it fair to say that 64-bit issues would crop up when say the array
> >> lengths are like 2^32 or bigger (because of ints)? Other libraries (like
> >> Blitz++) tend to have this issue too I believe.
> >>
> >> Cheers
> >> Shakes
> >>
> >> Sean McBride wrote:
> >>>
> >>> On 2/13/09 12:58 PM, shekharc said:
> >>>>
> >>>> Works fine on Windows Vista/XP x64 and Linux x64 fine for me so far.
> >>>
> >>> Out of curiosity, are you using datasets > 4 GiB?
> >
> > _______________________________________________
> > Powered by www.kitware.com
> >
> > Visit other Kitware open-source projects at
> > http://www.kitware.com/opensource/opensource.html
> >
> > Please keep messages on-topic and check the VTK FAQ at:
> > http://www.vtk.org/Wiki/VTK_FAQ
> >
> > Follow this link to subscribe/unsubscribe:
> > http://www.vtk.org/mailman/listinfo/vtkusers
> >
> _______________________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Please keep messages on-topic and check the VTK FAQ at:
> http://www.vtk.org/Wiki/VTK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://www.vtk.org/mailman/listinfo/vtkusers
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20090217/a7f2730f/attachment.htm>