Performance

Welcome to the OpenSolaris performance community!

This community is intended to serve as a place for people working on general
performance issues in OpenSolaris to ask questions, raise issues and
otherwise talk about OpenSolaris performance.
Our goal is to be the fastest performing system in the world without
compromises in other areas. This is the place for you to voice your ideas,
concerns and suggestions. We hope that the discussion within this community
will be helpful for both users and developers of OpenSolaris.

OpenSolaris Performance Principles

We feel that it is important to state major principles that guide the development
and discussion. For the OpenSolaris performance work we suggest the following
set of guiding principles:

• • If another major system is faster than OpenSolaris, it is a bug.
  • Performance is a goal, correctness is a constraint.
  • You can not improve performance if you can not measure it.

OpenSolaris Performance Community Goals

• • Improve the performance and efficiency of OpenSolaris.
  • Develop and improve measurement tools and methodology to facilitate the
          above.
  • Foster awareness of the above throughout the OpenSolaris community.

The Performance Wiki

• Solaris Performance Wiki RSS Feed

Discussion Topics

The primary areas of interest are:

• • Performance measurements
  • Performance improvements
  • Tools

Some initial areas to look at are

• • Microbenchmarks performance
  • Scheduler performance and CMT (multi-{theaded,cored} chips) support
  • NUMA support (see overview for more info)
  • DTrace based tools to better enable performance observability

We are looking forward to lively performance discussions at this forum! 

Links

• DTrace OpenSolaris Community
• Observability Community
• Core-aware Version of psrinfo(1M) command
• Enhance Solaris for Intel Project performance work