During compilation we seem to be picking up float.h from Intel LLVM include dir as opposed to clang we use for compiling. This results in some undefined macros for MAX/MIX size types.
[ 63%] Building CXX object samples/0_MatrixMultiply/CMakeFiles/MatrixMultiply.dir/MatrixMultiply.cpp.o
/home/pvelesko/space/chipStar/fix-icpx-include/samples/0_MatrixMultiply/MatrixMultiply.cpp:317:15: error: use of undeclared identifier 'FLT_EPSILON'
317 | float eps = FLT_EPSILON * 4;
During compilation we seem to be picking up
float.h
from Intel LLVM include dir as opposed toclang
we use for compiling. This results in some undefined macros for MAX/MIX size types.Workaround: