-qhot

In IBM® XL C/C++ for AIX® 16.1.0 or earlier releases, the -qhot option performs high-order loop analysis and transformations (HOT) during optimization.

If you specified -qhot when compiling your program with IBM XL C/C++ for AIX 16.1.0 or earlier releases, consider the following guidelines when migrating your program to IBM Open XL C/C++ for AIX 17.1.2:
  • If you used -qhot without any suboption in IBM XL C/C++ for AIX 16.1.0 or earlier releases, use -Ofast instead in IBM Open XL C/C++ for AIX 17.1.2.
  • If you used -qnohot, try using -Ofast in IBM Open XL C/C++ for AIX 17.1.2. However, if you encounter any floating-point precision issues, try using -O3 instead. If the issues can be resolved, you can go back to use -Ofast and achieve finer control over the floating-point optimizations with one or more of the following options:
    • -f[no-]honor-infinities
    • -f[no-]honor-nans
    • -f[no-]math-errno
    • -f[no-]finite-math-only
    • -f[no-]associative-math
    • -f[no-]reciprocal-math
    • -f[no-]signed-zeros
    • -f[no-]trapping-math
    • -ffp-contract
    • -f[no-]rounding-math
  • There is not an option that is functionally equivalent to -qhot=level=0|1|2. However, you can control the level of optimizations including some loop optimizations with the -O2, -O3, or -Ofast option.
  • The effect of -qhot=vector can be achieved by specifying -mllvm -vector-library=MASSV along with an optimization level that triggers loop vectorization.
  • The effect of -qhot=fastmath can be achieved by specifying -O3 -fapprox-func or -Ofast that invokes scalar MASS library and triggers loop vectorization.

Related information