<span style>Hello,</span><div style><br></div><div style>i installed visit from source on my imac running mac osx lion (in order to get the parallel engine working) as well as the CarpetHDF5 plugin, and everything works...in case anyone is interested i could summarize the crucial steps i took in order to make it work for you..</div>
<div style><br></div><div style>best wishes,</div><div style><br></div><div style>Vassili</div><br><div class="gmail_quote">On Tue, May 22, 2012 at 9:57 PM, Scott Hawley <span dir="ltr"><<a href="mailto:scott.hawley@belmont.edu" target="_blank">scott.hawley@belmont.edu</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Maria,<br>
No there are only a few libopen-rte files you can unlink if you want.<br>
<br>
I noticed that these errors are coming from /usr/local/lib, which is not a<br>
location I've used.<br>
I put the links in /usr/lib.<br>
<div class="HOEnZb"><div class="h5"><br>
<br>
<br>
<br>
On 5/22/12 9:06 AM, "Babiuc-Hamilton, Maria" <<a href="mailto:babiuc@marshall.edu">babiuc@marshall.edu</a>> wrote:<br>
<br>
>Scott,<br>
><br>
>Yes, I did. The error I am getting now when configuring Cactus is:<br>
>dyld: Symbol not found: _orte_ess<br>
> Referenced from: /usr/local/lib/libmpi.0.dylib<br>
> Expected in: /usr/local/lib/libopen-rte.0.dylib<br>
> in /usr/local/lib/libmpi.0.dylib<br>
><br>
><br>
>Looking on the web, it seems that other Apple users had this problem when<br>
>trying to link libraries.<br>
>I'm afraid that I made this mess when I linked libopen-rte*?<br>
>Is there a clean and quick fix, or I need to reinstall my whole operating<br>
>system? :O<br>
><br>
>Please help!<br>
>Maria<br>
><br>
><br>
>________________________________________<br>
>From: Scott Hawley [<a href="mailto:scott.hawley@belmont.edu">scott.hawley@belmont.edu</a>]<br>
>Sent: Monday, May 21, 2012 4:05 PM<br>
>To: Babiuc-Hamilton, Maria<br>
>Cc: <a href="mailto:users@einsteintoolkit.org">users@einsteintoolkit.org</a><br>
>Subject: Re: Follow-up: [Users] How I got Visit 2.4.2 & CarpetHDF5 to<br>
>work on Mac OS X Lion<br>
><br>
>You compiled the CarpetHDF plugin like I said, and are running visit with<br>
>-noconfig?<br>
><br>
><br>
><br>
>On 5/21/12 3:03 PM, "Babiuc-Hamilton, Maria" <<a href="mailto:babiuc@marshall.edu">babiuc@marshall.edu</a>> wrote:<br>
><br>
>>Thanks. I just downgraded to Visit 2.4.2 and now it looked like I got rid<br>
>>of that error.<br>
>>However, when trying to see a h5 file, I got the error:<br>
>><br>
>>The metadata server running on host localhost has exited abnormally.<br>
>>VisIt is trying to restart it.<br>
>><br>
>>Shortly thereafter, the following occured...<br>
>><br>
>>VisIt was unable to open<br>
>>"/Users/student/Cactus/VizTools/ks-mclachlan_metric.h5". Each attempt to<br>
>>open it caused VisIt's metadata server to crash. This can occur when the<br>
>>file is corrupted, or when the underlying file format has changed and<br>
>>VisIt's readers have not been updated yet, or when the reader VisIt is<br>
>>using for your file format is not robust. Please check whether the file<br>
>>is corrupted and, if not, contact a VisIt developer.<br>
>><br>
>>Moreover, when I am trying to run Cactus, I am getting this:<br>
>><br>
>>./cactus_tov tov.rpar<br>
>>dyld: Symbol not found: __ZN3MPI3Win14Set_errhandlerERKNS_10ErrhandlerE<br>
>> Referenced from: /Users/student/Cactus/exe/./cactus_tov<br>
>> Expected in: /usr/lib/libmpi_cxx.0.dylib<br>
>><br>
>>Trace/BPT trap<br>
>><br>
>>I checked, and I do have /usr/lib/libmpi_cxx.0.dylib<br>
>>Any ideas?<br>
>><br>
>>thanks,<br>
>>Maria<br>
>><br>
>>________________________________________<br>
>>From: Scott Hawley [<a href="mailto:scott.hawley@belmont.edu">scott.hawley@belmont.edu</a>]<br>
>>Sent: Monday, May 21, 2012 3:18 PM<br>
>>To: Babiuc-Hamilton, Maria<br>
>>Cc: <a href="mailto:users@einsteintoolkit.org">users@einsteintoolkit.org</a><br>
>>Subject: Re: Follow-up: [Users] How I got Visit 2.4.2 & CarpetHDF5 to<br>
>>work on Mac OS X Lion<br>
>><br>
>>Yes, I remember seeing that error too; although it may have only been<br>
>>with<br>
>>version 2.5 that I saw that.<br>
>><br>
>>So apparently some other libraries also used to be native, and are no<br>
>>longer, and I had already created the other links.<br>
>><br>
>>So in addition to lipmpi*, you'll also want to do<br>
>><br>
>>% cd /usr/lib/; sudo ln -s /opt/local/lib/libopen-pal*<br>
>>/opt/local/lib/libopen-rte* .<br>
>><br>
>><br>
>>Let me know if that helps.<br>
>><br>
>><br>
>><br>
>><br>
>>On 5/21/12 2:11 PM, "Babiuc-Hamilton, Maria" <<a href="mailto:babiuc@marshall.edu">babiuc@marshall.edu</a>> wrote:<br>
>><br>
>>>Scott,<br>
>>><br>
>>>I finally got around to check follow your instructions, but I ran into<br>
>>>this problem:<br>
>>><br>
>>>-----------------------------------------------------------------<br>
>>> Parsed document of type Plugin<br>
>>>-----------------------------------------------------------------<br>
>>><br>
>>>Plugin: CarpetHDF5 ("Carpet AMR/curvilinear HDF5 output generated by the<br>
>>>Cactus Computational Toolkit", type=database) -- version 2.1<br>
>>> Attribute: ()<br>
>>><br>
>>>-- The C compiler identification is GNU<br>
>>>-- The CXX compiler identification is GNU<br>
>>>-- Checking whether C compiler has -isysroot<br>
>>>-- Checking whether C compiler has -isysroot - yes<br>
>>>-- Checking whether C compiler supports OSX deployment target flag<br>
>>>-- Checking whether C compiler supports OSX deployment target flag - yes<br>
>>>-- Check for working C compiler: /usr/bin/gcc<br>
>>>-- Check for working C compiler: /usr/bin/gcc -- works<br>
>>>-- Detecting C compiler ABI info<br>
>>>-- Detecting C compiler ABI info - done<br>
>>>-- Checking whether CXX compiler has -isysroot<br>
>>>-- Checking whether CXX compiler has -isysroot - yes<br>
>>>-- Checking whether CXX compiler supports OSX deployment target flag<br>
>>>-- Checking whether CXX compiler supports OSX deployment target flag -<br>
>>>yes<br>
>>>-- Check for working CXX compiler: /usr/bin/c++<br>
>>>-- Check for working CXX compiler: /usr/bin/c++ -- works<br>
>>>-- Detecting CXX compiler ABI info<br>
>>>-- Detecting CXX compiler ABI info - done<br>
>>>CMake Error at CMakeLists.txt:90 (VISIT_PLUGIN_TARGET_RTOD):<br>
>>> Unknown CMake command "VISIT_PLUGIN_TARGET_RTOD".<br>
>>><br>
>>><br>
>>>-- Configuring incomplete, errors occurred!<br>
>>>make: *** No targets specified and no makefile found. Stop.<br>
>>><br>
>>><br>
>>>The only difference was that I did a<br>
>>> ./configure --enable-cxx<br>
>>>first in my hdf5-1.8.9 directory.<br>
>>><br>
>>><br>
>>>Could you please help me out?<br>
>>><br>
>>>Christian, do you have any suggestions?<br>
>>><br>
>>>thanks,<br>
>>>Maria<br>
>>>________________________________________<br>
>>>From: Babiuc-Hamilton, Maria<br>
>>>Sent: Tuesday, May 15, 2012 8:42 AM<br>
>>>To: Scott Hawley<br>
>>>Subject: Re: [Users] How I got Visit 2.4.2 & CarpetHDF5 to work on Mac<br>
>>>OS<br>
>>>X Lion<br>
>>><br>
>>>Scott,<br>
>>><br>
>>>Fantastic! I'll try it right away! I had problems making Visit work with<br>
>>>MacOS last year, and had to do my own little hack, but haven't use it<br>
>>>for<br>
>>>a while, because I was afraid that if I upgrade I'll run into problems<br>
>>>again.<br>
>>>Have had any success with ygraph?<br>
>>><br>
>>>Cheers,<br>
>>>Maria<br>
>>><br>
>>>On 5/14/12 5:23 PM, "Scott Hawley" <<a href="mailto:scott.hawley@belmont.edu">scott.hawley@belmont.edu</a>> wrote:<br>
>>><br>
>>>><br>
>>>>Here's what I did:<br>
>>>><br>
>>>>1. Downloaded the 64-bit binary of Visit 2.4.2 from the Visit web site<br>
>>>>2. Downloaded the CarpetHDF directory via svn.<br>
>>>>3. Downloaded hdf5-1.8.9 source via ftp. I just put it in the<br>
>>>>CarpetHDF5<br>
>>>>directory. Did a make and make installŠ.<br>
>>>> "HDF5 Directory" is then .../hdf5-1.8.9/hdf5/ (see below)<br>
>>>>4. Made sure "Remote Login" is enabled under "Sharing" in "System<br>
>>>>Preferences". Visit uses this.<br>
>>>> Went ahead and did a "ssh localhost" just to make sure. I was<br>
>>>>actually<br>
>>>>prompted to confirm an address because I'm on LSU's network right now.<br>
>>>> (Presumably Visit needs a clean login with no pesky questions or<br>
>>>>password prompts.)<br>
>>>>5. Created symbolic links in /usr/lib to whichever MPI libraries I<br>
>>>>planned<br>
>>>>to use.<br>
>>>> Older versions of OS X had native MPI support, and this is hard<br>
>>>>coded<br>
>>>>in the binary you get from the visit site.<br>
>>>> I use OpenMPI and MacPorts, so I went to /usr/lib and did "sudo ln<br>
>>>>-s<br>
>>>>/opt/local/lib/libmpi* ."<br>
>>>>6. Deleted my ~/.visit directory entirely. This may not be necessary.<br>
>>>>7. Went in the CarpetHDF5 and ran ./install and Šbasically just did<br>
>>>>what<br>
>>>>was in the readme, that is:<br>
>>>> in the xmledit window, clicked on "Cmake" and then updated the flags,<br>
>>>>as<br>
>>>>in<br>
>>>><br>
>>>> CXXFLAGS =<br>
>>>>-I/Applications/VisIt.app/Contents/Resources/2.4.2/darwin-x86_64/includ<br>
>>>>e<br>
>>>>-I/Users/shawley/apps/VizTools/CarpetHDF5/hdf5-1.8.9/hdf5/include<br>
>>>> LDFLAGS =<br>
>>>>-L/Applications/VisIt.app/Contents/Resources/2.4.2/darwin-x86_64/lib<br>
>>>>-L/Users/shawley/apps/VizTools/CarpetHDF5/hdf5-1.8.9/hdf5/lib<br>
>>>> Saved and closed the window, let it do its thing...<br>
>>>>8. Ran visit from the command line, using the -noconfig option. The<br>
>>>>latest version of Visit has a problem with "localhost" on Mac, for some<br>
>>>>reason.<br>
>>>>so<br>
>>>>visit -noconfig<br>
>>>><br>
>>>>..and then visit came up, I opened a .h5 file and was able to do all<br>
>>>>the<br>
>>>>usual things with it!<br>
>>>><br>
>>>>Good luck!<br>
>>>>-Scott<br>
>>>><br>
>>>><br>
>>>>_______________________________________________<br>
>>>>Users mailing list<br>
>>>><a href="mailto:Users@einsteintoolkit.org">Users@einsteintoolkit.org</a><br>
>>>><a href="http://lists.einsteintoolkit.org/mailman/listinfo/users" target="_blank">http://lists.einsteintoolkit.org/mailman/listinfo/users</a><br>
>>><br>
>>><br>
>><br>
>><br>
>><br>
><br>
><br>
><br>
<br>
<br>
_______________________________________________<br>
Users mailing list<br>
<a href="mailto:Users@einsteintoolkit.org">Users@einsteintoolkit.org</a><br>
<a href="http://lists.einsteintoolkit.org/mailman/listinfo/users" target="_blank">http://lists.einsteintoolkit.org/mailman/listinfo/users</a><br>
<br>
</div></div></blockquote></div><br>