-qsimd

In IBM® XL C/C++ for AIX® 16.1.0 or earlier releases, the -qsimd=noauto option disabled the conversion of loop array operations into vector instructions.

In IBM Open XL C/C++ for AIX 17.1.1, use -fno-vectorize and -fno-slp-vectorize options to achieve the same effect to disable auto vectorization features. For an LTO build, you need to add -Wl,-bplugin_opt:-vectorize-loops=false -Wl,-bplugin_opt:-vectorize-slp=false on the link step.

Related information

• #pragma nosimd
• The "Clang command line argument reference" section in the Clang documentation