如题,
我的两个复数的类型定义分别如下:
cmplxFloatImRe_t gRxChPhaseComp[12];
cmplx32ReIm_t *input;
而
typedef struct cmplxFloatImRe_t_
{
float imag; /*!< @brief imaginary part */
float real; /*!< @brief real part */
} cmplxFloatImRe_t;
typedef struct cmplx32ReIm_t_
{
int32_t real; /*!< @brief real part */
int32_t imag; /*!< @brief imaginary part */
} cmplx32ReIm_t;
/* Compensation */
for (antIndx = 0; antIndx < numAnt; antIndx++)
{
re = (float)input[antIndx].real*gRxChPhaseComp[antIndx].real – input[antIndx].imag*gRxChPhaseComp[antIndx].imag;
im = (float)input[antIndx].real*gRxChPhaseComp[antIndx].imag + input[antIndx].imag*gRxChPhaseComp[antIndx].real;
for (antIndx = 0; antIndx < numAnt; antIndx++)
{
re = (float)input[antIndx].real*gRxChPhaseComp[antIndx].real – input[antIndx].imag*gRxChPhaseComp[antIndx].imag;
im = (float)input[antIndx].real*gRxChPhaseComp[antIndx].imag + input[antIndx].imag*gRxChPhaseComp[antIndx].real;
input[antIndx].real = (int32_t)re;
input[antIndx].imag = (int32_t)im;
}
input[antIndx].imag = (int32_t)im;
}
上述代码怎么处理更加高效快捷的?目前的执行效率太慢了。。
Shine:
请尝试打开cache和使用-o优化选项。
user4162505:
回复 Shine:
你好,,这个具体怎么操作的? 有大概的使用指导例程啥的么? 谢谢!
Shine:
回复 user4162505:
请参考下面的文档。
5 Discussion on Cache Strategy
www.ti.com/…/swra552.pdf