VECTOR
The VECTOR option controls whether the compiler enables the vector programming support and
automatically takes advantage of vector/SIMD instructions. The VECTOR option provides potential
performance improvements in the following cases:
- Binary floating-point data types: float and long double
- When the VECTOR option is specified with ARCH(12) and FLOAT(IEEE), the
long double
andfloat
data types can be processed in the vector registers. - Binary floating-point data type: double
- When the VECTOR option is specified with ARCH(11) or higher and FLOAT(IEEE), the
double
data types can be processed in the vector registers. - Built-in library functions
- When the VECTOR option is specified with ARCH(11) or higher, certain built-in library functions can take advantage of vector string instructions to accelerate the processing of strings of character data.
- Fixed-point decimal operations
- When the VECTOR option is specified with ARCH(12), the compiler can take advantage of vector decimal instructions that perform the operations in register operands, which can improve the performance of such applications.
- SIMD instructions
- When the AUTOSIMD suboption is in effect, the compiler generates code, when possible, using the SIMD instructions. SIMD instructions calculate several results at one time, which is faster than calculating each result sequentially.