Jin Rhee

Installing and using GCC on Mac Silicon

2024-10-27T00:00:00-07:00

*[master][~/repositories/hepcpp/cpluspluscourse/exercises]$ type gcc
gcc is /usr/bin/gcc
*[master][~/repositories/hepcpp/cpluspluscourse/exercises]$ gcc --version
Apple clang version 15.0.0 (clang-1500.3.9.4)
Target: arm64-apple-darwin23.5.0
Thread model: posix
InstalledDir: /Library/Developer/CommandLineTools/usr/bin

The gcc shipped with Mac Silicon is Apple clang → what?

I want to use the real gcc:

$ brew install gcc

Brew gcc installed in /opt/homebrew/bin/:

$ ls | grep gcc
aarch64-apple-darwin23-gcc-14
aarch64-apple-darwin23-gcc-ar-14
aarch64-apple-darwin23-gcc-nm-14
aarch64-apple-darwin23-gcc-ranlib-14
gcc-14
gcc-ar-14
gcc-nm-14
gcc-ranlib-14
$ ./gcc-14 --version
gcc-14 (Homebrew GCC 14.2.0_1) 14.2.0
Copyright (C) 2024 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Now to use the brew installed gcc, I simply need to add installation directory to PATH:

*[master][~/repositories/hepcpp/cpluspluscourse/exercises]$ echo $PATH | grep bin
... :/opt/homebrew/bin:/opt/homebrew/sbin: ... :/usr/bin:/bin:/usr/sbin ...

Both the shipped and brew installed gccs are on PATH…

$ type gcc
gcc is /usr/bin/gcc

aaaand the shipped gcc is still being used.

I can either:

alias; or
create new gcc symlink → need to have PATH specify /usr/local/bin before /usr/bin

References: https://stackoverflow.com/questions/64992467/mac-clang-installation-seems-to-override-gcc-install

https://apple.stackexchange.com/questions/245891/installed-gcc-with-homebrew-now-how-to-use-that-gcc-instead-of-clang

Use LLVM/Clang as CMake compiler

2024-09-30T00:00:00-07:00

Switch to LLVM/Clang compiler for CMake (Ubuntu 22.04 aarch64 on Apple Silicon via Multipass)

A package I work with compiles fine on Linux AMD, but crashes in a firey hell on my Mac.

find_package(OpenMP)
if (OPENMP_FOUND
    set (CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${OpenMP_C_FLAGS}")
    set (CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${OpenMP_CXX_FLAGS}")
endif()

Results in:

CMake Error at CMakeLists.txt:161 (add_executable):
  Target "s_graphs_prefiltering_node" links to target "OpenMP::OpenMP_CXX"
  but the target was not found.  Perhaps a find_package() call is missing for
  an IMPORTED target, or an ALIAS target is missing?

Check libomp-dev is installed:

$ apt show libomp-dev
Package: libomp-dev
Version: 1:14.0-55~exp2
Priority: extra
Section: universe/libdevel
Source: llvm-defaults (0.55~exp2)
Origin: Ubuntu
Maintainer: Ubuntu Developers 
Original-Maintainer: LLVM Packaging Team 
Bugs: https://bugs.launchpad.net/ubuntu/+filebug
Installed-Size: 16.4 kB
Depends: libomp-14-dev (>= 14~)
Download-Size: 3074 B
APT-Manual-Installed: yes
APT-Sources: http://ports.ubuntu.com/ubuntu-ports jammy/universe arm64 Packages
Description: LLVM OpenMP runtime - dev package
 The runtime is the part of the OpenMP implementation that your code is
 linked against, and that manages the multiple threads in an OpenMP program
 while it is executing.
 This is a dependency package providing the default LLVM OpenMP dev
 package.

This was solved by using LLVM/Clang as the compiler for CMake rather than gcc:

Use NoMachine to access VM

2024-09-30T00:00:00-07:00

A GUI is often needed when using VMs and ssh’ing into shell is insufficient. Windows Remote Desktop offers a ssh-based remote access method, but has a really slow GUI. I’ve found NoMachine to have a much faster response (which saves a lot of headaches).

You need NoMachine on both the client and host machines: Install NoMachine

Connecting to a machine over an internet connection has a few more steps, but simply accessing a VM on your host computer is really easy!