Enable the static (compile-time) check on the expressions, if it's disabled run-time check will be performed.
#define EXMAT_STATIC_CHECK 0
Enable SIMD code generation
#define EXMAT_ENABLE_SIMD 1
Enable Intel SSE. For GCC , add the compiler switch "-msse"
#define EXMAT_ENABLE_SSE 1
Enable Intel SSE2. For GCC , add the compiler switch "-msse2"
#define EXMAT_ENABLE_SSE2 1
Enable memory alignment check for every SIMD container construction
#define ENABLE_ALIGNMENT_CHECK 0
The memory alignment required by the SIMD operations (in bytes)
#define SIMD_ALIGN_BYTE 16
Enable run-time dispatching of approximated functions
#define EXMAT_ENABLE_AM_DISPATCH 1
Enable aliasing analysis system. Disable it for faster compilation
#define ENABLE_ALIAS_ANALYSIS 1
Enable temporary analysis system. Disable it for faster compilation
#define ENABLE_TMP_ANALYSIS 1
The default analysis that will preformed during assignment, there are 3 options you can choose from:
1) ASSIGN_ANALYSIS::NONE
2) ASSIGN_ANALYSIS::ALIAS
3) ASSIGN_ANALYSIS::TEMPORARY
#define EXMAT_DEFAULT_ANALYSIS ASSIGN_ANALYSIS::ALIAS|ASSIGN_ANALYSIS::TEMPORARY
Enable compile-time loop unroller
#define ENABLE_UNROLLING 1
No. of floating point operations to be allowed when unrolling a row
enum { MAX_UROLLROW_NOP = 15 };
No. of floating point operations to be allowed when unrolling the whole matrix
enum { MAX_FULLUROLL_NOP = 200 };
No. of floating point operations to be allowed when unrolling vector dot operation during matrix mutliplication
enum { MAX_MULDOTUROLL_ENOP = 12 };