r*y
2 楼
not necessarily
The old rule is do everything in multiplication rather than division
An American intern once used this "rule" to change my code, then
she was crying loud...
【在 J*****n 的大作中提到】
: double 类型 x,y
: (y/x)/x 是不是比 y/(x*x) 更加好一些?
The old rule is do everything in multiplication rather than division
An American intern once used this "rule" to change my code, then
she was crying loud...
【在 J*****n 的大作中提到】
: double 类型 x,y
: (y/x)/x 是不是比 y/(x*x) 更加好一些?
C*n
3 楼
不好说
两个的结果可能不一致
division
【在 r****y 的大作中提到】
: not necessarily
: The old rule is do everything in multiplication rather than division
: An American intern once used this "rule" to change my code, then
: she was crying loud...
两个的结果可能不一致
division
【在 r****y 的大作中提到】
: not necessarily
: The old rule is do everything in multiplication rather than division
: An American intern once used this "rule" to change my code, then
: she was crying loud...
相关阅读
Haar Wavelet的thresholding到底是怎么算的啊[转载] Is the "bench" in Matlab reliableRe: Is there any documentation available for Intel Fortran 90 under li请教：求最值问题any body debugged MPI programms?大家能不能seriously讨论一下这个topicHi, guys. Re: molecular simulation怎么检测小于1毫秒的时间段？有人在 Windows 2K + cygwin 下配置 GSL 成功吗？Implementation of Genetic Algorithm急问，网络图论问题！！！请教一个Gaussian quadrature的问题let me ask a question again关于VF/VC中cannot save file...in use..的问题Speed comparison of various number crunching packages[转载] any procedures to compute and show the permutations?A question about IFA300The Future of Scientific ComputingHermite interpolationMONTE CARLO METHOD QUESTION