[yt-users] install error

Cameron Hummels chummels at astro.columbia.edu
Fri Jul 13 08:48:19 PDT 2012 

I'm +1 on making things more explicit in the install_script message to 
users (particularly of mac).  I think most people treat it like a black 
box and hope for the best.  And if it breaks they come here. If 
something popped up and said, *make sure you have xcode gcc installed*, 
it might be better for all involved.

Cameron


On 07/13/2012 11:38 AM, Matthew Turk wrote:
> Hi Casey,
>
> Thanks for the summary of the options -- perhaps we should make the
> warning more explicit in the install_script?
>
> With respect to things being far behind, I think the situation will
> only get worse.  Apple stopped updating gcc when they moved to GPLv3,
> and have started shipping their own stack based on LLVM.  Perhaps a
> brainstorming session of options -- the Xcode gcc, AT&T research
> gfortran, Macports, other binary installers, etc etc -- would be
> helpful over on yt-dev for addressing how to keep up to date with GCC
> as well as ensure compatibility with as much as possible.
>
> -Matt
>
> On Fri, Jul 13, 2012 at 8:17 AM, Casey W. Stark <caseywstark at gmail.com> wrote:
>> That's a good point Sam. I've seen a lot of people reccomend the hpc for Mac
>> compilers, but they don't support many Darwin specific flags, causing common
>> build systems to break. This took way too long to figure out as a new Mac
>> user. Much better to stick with Xcode gcc and get the AT&T research
>> gfortran.
>>
>> If you really want a new version of gnu compilers (kind of amazing how far
>> behind they are), I think macports has newer versions built to have the
>> apple flags.
>>
>> - Casey
>>
>>
>> On Friday, July 13, 2012, Sam Skillman wrote:
>>> Beware -- updating gcc to anything past 4.2 will likely break the compile
>>> on a mac.  You want to make sure to use the Xcode gcc, which you should be
>>> able to do by setting:
>>>
>>> export CC=gcc-4.2
>>> export CXX=g++-4.2
>>>
>>> or the equivaltent command in csh if you aren't a bash person.  If you had
>>> done that, then I'm not sure what the problem would be.  You'll likely have
>>> to wipe the existing install for those exports to do you any good.
>>>
>>> Sam
>>>
>>> On Fri, Jul 13, 2012 at 1:33 AM, Elizabeth Tasker
>>> <tasker at astro1.sci.hokudai.ac.jp> wrote:
>>>
>>> Hi Nathan,
>>>
>>> This is also a mac - but I'll suggest we update gcc. Thanks!
>>>
>>> Elizabeth
>>>
>>>
>>> On Jul 13, 2012, at 4:21 PM, Nathan Goldbaum wrote:
>>>
>>>> Hi Elizabeth,
>>>>
>>>> I tried running the install script on my mac laptop after your e-mail.
>>>> I was able to build yt from the install script with no problems.
>>>>
>>>> It would help if you could let us know a bit more about the system
>>>> you're trying to build yt on.  In particular, I think you'll run into issues
>>>> if you're using an old version of gcc that doesn't support OpenMP.  Version
>>>> 4.2 or newer should work.
>>>>
>>>> -Nathan
>>>>
>>>>
>>>> On 7/13/12 12:16 AM, Elizabeth Tasker wrote:
>>>>> Hi Andrew,
>>>>>
>>>>> We ran the install script from scratch, so I assume it would have
>>>>> installed the version of cython it needed?
>>>>>
>>>>> Elizabeth
>>>>>
>>>>>
>>>>> On Jul 13, 2012, at 1:59 PM, Andrew Myers wrote:
>>>>>
>>>>>> Hi Elizabeth,
>>>>>>
>>>>>> I saw errors like this when I tried to use an older version of Cython.
>>>>>> Are you sure that you're using version 0.16?
>>>>>>
>>>>>> -Andrew M
>>>>>>
>>>>>> On Thu, Jul 12, 2012 at 9:55 PM, Elizabeth Tasker
>>>>>> <tasker at astro1.sci.hokudai.ac.jp <mailto:tasker at astro1.sci.hokudai.ac.jp>>
>>>>>> wrote:
>>>>>>
>>>>>>     Hi,
>>>>>>
>>>>>>     We've had some problems installing yt via the install script.
>>>>>>
>>>>>>     Initially it did not download distribute, which we did by hand.
>>>>>>     Then, it couldn't get pip, so I commented that line out and now
>>>>>>     it's died on cython with:
>>>>>>
>>>>>>
>>>>>>
>>>>>>     cythonc:> yt/utilities/lib/PointsInVolume.c
>>>>>>     building extension "yt.utilities.lib.QuadTree" sources
>>>>>>     cythonc:> yt/utilities/lib/QuadTree.c
>>>>>>     building extension "yt.utilities.lib.RayIntegrators" sources
>>>>>>     cythonc:> yt/utilities/lib/RayIntegrators.c
>>>>>>     building extension "yt.utilities.lib.VolumeIntegrator" sources
>>>>>>     cythonc:> yt/utilities/lib/VolumeIntegrator.c
>>>>>>     building extension "yt.utilities.lib.grid_traversal" sources
>>>>>>     cythonc:> yt/utilities/lib/grid_traversal.c
>>>>>>
>>>>>>     Error compiling Cython file:
>>>>>>     ------------------------------------------------------------
>>>>>>     ...
>>>>>>             if im.vd_strides[0] == -1:
>>>>>>                 with nogil, parallel(num_threads = num_threads):
>>>>>>                     idata = <ImageAccumulator *>
>>>>>>     malloc(sizeof(ImageAccumulator))
>>>>>>                     idata.supp_data = self.supp_data
>>>>>>                     v_pos = <np.float64_t *> malloc(3 *
>>>>>>     sizeof(np.float64_t))
>>>>>>                     for j in prange(size,
>>>>>> schedule="static",chunksize=1):
>>>>>>                                   ^
>>>>>>     ------------------------------------------------------------
>>>>>>
>>>>>>     yt/utilities/lib/grid_traversal.pyx:368:31: Invalid keyword
>>>>>>     argument: chunksize
>>>>>>
>>>>>>     Error compiling Cython file:
>>>>>>     ------------------------------------------------------------
>>>>>>     ...
>>>>>>                     idata.supp_data = self.supp_data
>>>>>>                     v_pos = <np.float64_t *> malloc(3 *
>>>>>>     sizeof(np.float64_t))
>>>>>>                     v_dir = <np.float64_t *> malloc(3 *
>>>>>>     sizeof(np.float64_t))
>>>>>>                     # If we do not have a simple image plane, we have
>>>>>>     to cast all
>>>>>>                     # our rays
>>>>>>                     for j in prange(size, schedule="dynamic",
>>>>>>     chunksize=100):
>>>>>>                                   ^
>>
>> _______________________________________________
>> yt-users mailing list
>> yt-users at lists.spacepope.org
>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>>
> _______________________________________________
> yt-users mailing list
> yt-users at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>

-- 
Cameron Hummels
PhD Candidate, Astronomy Department of Columbia University
Public Outreach Director, Astronomy Department of Columbia University
NASA IYA New York State Student Ambassador
http://outreach.astro.columbia.edu
PGP: 0x06F886E3

More information about the yt-users mailing list