Some attempts at code optimization

1. Loop body

The difference in the two-segment loop body of the following program is only the use of local variables to reduce the one-time addition operation

Func Main () {buf: = make ([]byte, 100000000) var offset intvar sum int//do not use local variables, multiple addition operations T3: = time. Now () offset = 0sum = 0for I: = 0; I < Len (BUF); i++ {sum + = Int (buf[offset+i]) sum + = Int (Buf[offset+i])}t4: = time. Now (). Sub (T3) fmt. PRINTLN (T4)//Use local variable/*t1: = time. Now () offset = 0sum = 0for I: = 0; I < Len (BUF); i++ {Temp: = offset + isum + = Int (buf[temp]) sum + = Int (buf[temp])}t2: = time. Now (). Sub (t1) fmt. Println (T2) */return}

Results of multiple runs: the first paragraph (unused local variables) is always 7% or so slower than the second paragraph. The number of cycles in the experiment must be large enough (10^8 or more), otherwise the result is uncertain

Conclusion: In the case of a large number of cycles, local variables can reduce the running time.

2. The effect of type conversion on run time

The only difference between the two test programs is whether the type conversion is done

Result: The running time of the two is basically consistent

Conclusion: type conversion has little effect on program operation.

　　

　　

Some attempts at code optimization