Re: [Libevent-users] Strange timeout scheduling in 1.4

[Dongsheng Song <dongsheng.song@xxxxxxxxx>]

On Mon, Aug 8, 2011 at 06:15, Leonid Evdokimov <leon@xxxxxxxxxxxx> wrote:

2.6.24 result is really surprising me.

Maybe you forgot the CPU not the same:

$ dmesg | grep GHz
[ÂÂÂ 0.056379] CPU0: Intel(R) Xeon(TM) CPU 3.20GHz stepping 04
[ÂÂÂ 0.144076] CPU1: Intel(R) Xeon(TM) CPU 3.20GHz stepping 04
[ÂÂÂ 0.240057] CPU2: Intel(R) Xeon(TM) CPU 3.20GHz stepping 04
[ÂÂÂ 0.336144] CPU3: Intel(R) Xeon(TM) CPU 3.20GHz stepping 04
[ÂÂÂ 0.432148] CPU4: Intel(R) Xeon(TM) CPU 3.20GHz stepping 04
[ÂÂÂ 0.528083] CPU5: Intel(R) Xeon(TM) CPU 3.20GHz stepping 04
[ÂÂÂ 0.620103] CPU6: Intel(R) Xeon(TM) CPU 3.20GHz stepping 04
[ÂÂÂ 0.712116] CPU7: Intel(R) Xeon(TM) CPU 3.20GHz stepping 04

$ cat /proc/version
Linux version 2.6.32-5-amd64 (Debian 2.6.32-35) (dannf@xxxxxxxxxx) (gcc version 4.3.5 (Debian 4.3.5-4) ) #1 SMP Mon Jun 13 05:49:32 UTC 2011

$ ./gettimeofday-cost
(CLOCK_REALTIME) gtod: 345.124579
(CLOCK_REALTIME) clock(mono): 475.158095
(CLOCK_REALTIME) memcpy: 4.811645
(CLOCK_MONOTONIC) gtod: 344.205039
(CLOCK_MONOTONIC) clock(mono): 475.315154
(CLOCK_MONOTONIC) memcpy: 5.014204
(CLOCK_PROCESS_CPUTIME_ID) gtod: 346.424803
(CLOCK_PROCESS_CPUTIME_ID) clock(mono): 474.858415
(CLOCK_PROCESS_CPUTIME_ID) memcpy: 5.041936