This is a meta issue to track performance improvements to be made in $q for 1.3.0.

The main problems to be solved within the scope of this issue:

• Reduction of deferred instance memory to reduce GC pressure. One quick win is to move instance methods to prototype instead of created inside constructor.
• Some learnings should be taken from bluebird, which has a very fast promise implementation.

Related issues:

• $q is probably due for a rewrite #5223 (suggests rewriting $q, mentions memory pressure)
• Decouple $q in AngularJS 2.0 (and/or switch to bluebird) #6697 (focused on decoupling promise lib in Angular)

There is a PR from @caitp which updates A+ compliance, adds ES6-like constructor API, and begins to experiment with porting bluebird to $q (but with many tests failing): #8067. I've suggested simplifying that PR to not include perf improvements.

Benchmarks:

• JSPerf comparing $q, Native Promise, and Bluebird Promise
• JSPerf comparing many promise libraries, not including $q