Advanced options

If you are using advanced features of the Arm Performance Libraries, you can use the following additional options:

OpenMP

To use OpenMP for linking to multi-threaded Arm performance libraries, use the following commands:

Compiler Command
armclang -armpl=parallel
gcc -larmpl_lp64_mp

8-byte integer variables

To use 8-byte integer variables in C or Fortran, compile and link your code with the following commands:

Compiler Command
armclang -armpl=ilp64(Compile and link)
gcc -DINTEGER64 (Compile)
-larmpl_ilp64 (Link)

OpenMP and 64-bit integers

To use both OpenMP and 64-bit integers, use the following link flags:

Compiler Command
armclang -armpl=ilp64,parallel(Compile and link)
gcc -DINTEGER64(Compile)
-larmpl_ilp64_mp(Link)

Selecting target architecture

When using Arm Compiler For Linux, you can use the -mcpu flag to select which library to link with. There are four options: 

  • -mcpu=native allows the compiler to infer from the host system which libraries to use.
  • -mcpu=thunderx2t99 selects libraries tuned for Cavium ThunderX2 cores.
  • -mcpu=generic selects libraries that work on any AArch64 core.

Linking against static libraries

Use -static to link against the static rather than shared libraries.


Previous Next