macOS 10.14 Mojave and Xcode 10.0 support

[Felker__Kyle]

Hi, 

I recently upgraded my macOS system from 10.13.6 High Sierra and Xcode 9.4.1 to Mojave and Xcode 10.0. I had been using "Intel® Parallel Studio XE Composer Edition for C++ macOS* version 2018 update 2" for some time, but the OS upgrade of course broke some of system's functionality. Without reinstalling the Intel Compiler, my built-from-source MPI library continued to work with icc, but other libraries including HDF5 stopped working with the Intel C++ compiler. Trying a fresh build from source of HDF5 with the compiler does not fix the issue, as the linker refuses to use the shared library:

"dyld: __dyld section not supported in /Users/kfelker/.dotfiles/hdf5-1.10.3/src/.libs/libhdf5.103.dylib"

Installing the latest 2019 initial release version did not fix this, either. I see that only macOS 10.13 and Xcode 9.4 are officially supported by this version. Is there any indication of when the latest macOS environment may be supported by Intel Parallel Studio? Is there any way of knowing what partial support is currently offered?

Thanks!

[Viet_H_Intel]

We will be supporting 10.14 Mojave and Xcode 10.0 in our future version, but I don't have yet a timeframe to share with you.

Thanks,

Viet 

[marras__simone]

Hi,

What fix do you suggest while you it supported?

thanks,

S

[erling_andersen]

I wonder if there any news about this topic.

[Viet_H_Intel]

You can try Intel C++ Compiler 19.0 Update 3.

[Felker__Kyle]

I recently upgraded to Xcode 10.2.1 (released 2019-04-17), and I think it has caused compiler segfaults when using IPO on my macOS 10.14.4 system. Is this Xcode version officially supported? I reinstalled Intel C++ Compiler 19.0 Update 3 just in case, but it did not help. 

When I get to the link step after compiling with 

icpc  -O3 -std=c++11 -ipo -xhost -inline-forceinline -qopenmp-simd -qopt-prefetch=4 -diag-disable 3180 -Wall -Wextra -Werror -pedantic -pedantic-errors -diag-disable=175 -Wno-unused-variable -Wno-unused-function -Wshorten-64-to-32 ...

I get the following error:

ipo: remark #11000: performing multi-file optimizations
ipo: remark #11074: Inlining inhibited by limit max-size
ipo: remark #11074: Inlining inhibited by limit max-total-size
ipo: remark #11076: To get full report use -qopt-report=4 -qopt-report-phase ipo
ipo: remark #11006: generating object file /var/folders/ky/jhft78f938d39vnvnkxj3hdh0000gn/T/ipo_icpcCvuFIy.o
src/reconstruct/reconstruction.cpp(29): error #16219: Some OpenMP processing was skipped to constrain compile time. Consider overriding limits (-qoverride-limits).

          ": internal error: ** The compiler has encountered an unexpected problem.
** Segmentation violation signal raised. **
Access violation or stack overflow. Please contact Intel Support for assistance.

icpc: error #10014: problem during multi-file optimization compilation (code 4)

These errors do not occur at all when compiling with the same compiler version on Linux systems. 

[Felker__Kyle]

I thought that passing the following option

-qoverride-limits

might solve the issue (as indicated by the error which I missed upon first read). I guess this compiler flag behavior changed recently?

But it does not prevent the seg fault in the link step:

ipo: remark #11000: performing multi-file optimizations
ipo: remark #11074: Inlining inhibited by limit max-size
ipo: remark #11074: Inlining inhibited by limit max-total-size
ipo: remark #11076: To get full report use -qopt-report=4 -qopt-report-phase ipo
ipo: remark #11006: generating object file /var/folders/ky/jhft78f938d39vnvnkxj3hdh0000gn/T/ipo_icpcAmkIkE.o

          ": internal error: ** The compiler has encountered an unexpected problem.
** Segmentation violation signal raised. **
Access violation or stack overflow. Please contact Intel Support for assistance.

icpc: error #10014: problem during multi-file optimization compilation (code 4)

[Viet_H_Intel]

Please see this post for MacOS 10.14.4 supports https://software.intel.com/en-us/forums/intel-c-compiler/topic/807565  

[Felker__Kyle]

Thanks!