This converts a few places using the global allocator for temporary memory to use a bump allocator instead. In practice, no wall time benefits show up on macro benchmarks, but we do see improved memory use (which is likely just a side effect of the OS committing less memory than we request)
This converts a few places using the global allocator for temporary memory to use a bump allocator instead. In practice, no wall time benefits show up on macro benchmarks, but we do see improved memory use (which is likely just a side effect of the OS committing less memory than we request)