Repository navigation
v0.5.2
MPL-v0.5.2 Release Notes (Jan 12, 2025)
This release consists of two updates:
- More robust support for Arm-based Macs. #214 fixed a weak memory bug in the scheduler. MPL now successfully runs all of the
parallel-ml-benchwithout error (tested on a MacBook Air M2, 1-8 cores). - Major performance fix: promotion stacks (#211). Benchmarks with excessively deep call-stacks now perform as expected; for example, the
trees/test-unbalancedexample (which walks a highly unbalanced tall tree) now is orders of magnitude faster. Previously, deep call-stacks incurred a non-constant performance penalty at each PCall promotion. This fix extends the PCall calling convention with an auxiliary deque of promotable frames, ensuring that every promotion is constant-time.