Re: Benchmarking A DB-Intensive Script

[idle]

Hi, you can use Time::HiRes, like this:

use Time::HiRes qw(gettimeofday);
my $start=gettimeofday;
next part of code...
my $next=gettimeofday;
printf "time spended  %.4f" ,($next-$start);
next part of code...
my $next_1=gettimeofday;
printf "time spended  %.4f" ,($next_1-$next);

to see where exactly it is slow.

[bernanke01]

Hi, I was thinking of something like this using Benchmark as well, but then how do I add the values across many (say 10k) iterations of the loop? I didn't see a way to do that with either Benchmark or Time::HiRes -- did I miss it?

[BrowserUk]

Take a look at Benchmark::Timer. You can wrap a start/stop pairing around a piece of code and it accumulates the iterations and average the times taken for you. You can also run several different and overlapping pairs concurrently, which makes it easy to start with a course granularity and then home in onto those parts where you need finer granularity.

It allows you to get finer granularity than one of the per subroutine profilers without resorting to the time expensive per line profilers.