C ++ executes the memory memcpy Efficiency Test

 

During the memcpy operation, although it is a memory operation, it still consumes a little CPU. Today we tested the efficiency of executing memcpy in a single thread. The result is for configuring the work thread in TCP epoll.

The quantity has guiding significance. The following example shows how to execute memcpy quickly based on 8 K memory. One thread can copy 500 m in about 1 second. If the server bandwidth or network card reach the upper limit of 1 GB, the network I/O work thread can be opened with two threads. Considering the message parsing loss, the three threads are sufficient to resist the maximum Hardware load.

The test result is as follows:

Intel (r) Xeon (r) CPU e5405 @ 2.00 GHz

Do memcpy speed: 12.27 MS/MB
Each threads can do memcpy 667.645 MB

 

 1 # Include <iostream>
 2 # Include <sys/time. h>
 3 # Include < String . H>
 4  
5   Using   Namespace STD;
 6  
 7   Int Main ( Int Argc, Char * Argv [])
 8 {
 9           Long Len = 8192 ;
 10           Int Loop = 200 ;
 11           Char * P = New   Char [Len];
 12           Char * Q = P;
 13           Struct Timeval start, end;
 14 Gettimeofday (& START, null );
 15           For (Int I = 0 ; I <loop; ++ I)
 16 {
 17                   Char * P = New   Char [Len];
 18 * P = Char (I );
 19 Memcpy (p, q, Len );
 20 Delete [] P;
21 }
 22 Gettimeofday (& End, null );
 23 Cout < "  Do memcpy speed:  " <(End. TV _sec-start. TV _sec )* 1000 + Double (End. TV _usec-start. TV _usec)/(LEN * loop/ 1000 / 1000 )/Loop < "  MS/MB \ n " ;
 24 Cout < "  Each thread can do memcpy  " < Double (LEN) * loop/ 1000 / 1000 /(End. TV _sec-start. TV _sec) + Double (End. TV _usec-start. TV _usec )/ 1000 / 1000 ) < " MB \ n  " ;
 25  
 26 }