Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Highlighted
53 Views

C++ compiler update causing template specialization error

I am using the Intel C++ compiler to compile the Stan Math library (https://github.com/stan-dev/math).

I recently updated my Intel C++ compiler from 2018.1.156 to 2018.2.185 through Parallel Studio XE 2018 Update 2.  Following this update, some code that previously worked fine, now produces the following error,

error: more than one partial specialization matches the template argument list of class

I posted my error on the Stan Math library forum, and one of the library developers thinks that it is a compiler bug.  Here is a link to my post of the Stan Math library forum, which has some more details of the problem,

http://discourse.mc-stan.org/t/function-multi-normal-log-causing-intel-compiler-issues-again/4762

0 Kudos
5 Replies
Highlighted
Moderator
53 Views

Hi Philip,

Can you create a preprocess file for us to investigate? (keep the same command line, only change -c to -E).

 

Thanks,

Viet 

0 Kudos
Highlighted
53 Views

Sorry for not getting back to you.  I have still not resolved the issue in my original post.  I am coming back to it now and would like to sort it out.  Following the suggestion in the reply I generated a (pretty large) preprocess file - see attached.

I am now using a more recent version of the Intel tools - the install directory is called "compilers_and_libraries_2019.0.117"

0 Kudos
Highlighted
Moderator
53 Views

Thanks for the .i file. Let me do some investigation and get back to you.

 

 

0 Kudos
Highlighted
Moderator
53 Views

I've reported this to our Front End team. The internal bug number is CMPLRIL0-31126

Thanks,

 

0 Kudos
Highlighted
Beginner
53 Views

We are also seeing this problem, currently using the Intel C++ compiler version 19.0.1.144 .

Are there any updates on getting this fixed? Is there a possible workaround?

Also discussing this with the Stan developers via https://github.com/stan-dev/cmdstan/issues/625

0 Kudos