Hashtable

Frank Ch. Eigler fche@redhat.com
Mon Jul 10 15:20:00 GMT 2017


arkady.miasnikov wrote:

> [...]
> From the tests it appears that associative maps contribute significant
> part of the overhead. I was thinking about replacing the maps with a
> simple hashtable like this one (this is not production code)
> https://gist.github.com/larytet/c306d470f7b032c796efad15dcd609a9

The systemtap maps are already hash tables.  The complication here is
that you're using them in a highly contended way, so locking effects
dominate algorithmic ones.  PR10796 is probably more relevant to your
case.

- FChE



More information about the Systemtap mailing list