Just did a non-PGO regular `-d:danger` run. Times went up 1.9x but memory usage patterns were the same with `gc:arc` using much more RSS than `gc:boehm` or `gc:markAndSweep`. It's a pretty tiny program.
- NvP: s.add('x') 100M times HashBackupJim
- Re: NvP: s.add('x') 100M times Araq
- Re: NvP: s.add('x') 100M times HashBackupJim
- Re: NvP: s.add('x') 100M times cblake
- Re: NvP: s.add('x') 100M times Araq
- Re: NvP: s.add('x') 100M times cblake
- Re: NvP: s.add('x') 100M times cblake
- Re: NvP: s.add('x') 100M times cumulonimbus
- Re: NvP: s.add('x') 100M times cblake
- Re: NvP: s.add('x') 100M times b3liever
- Re: NvP: s.add('x') 100M times cblake
- Re: NvP: s.add('x') 100M times b3liever
- Re: NvP: s.add('x') 100M times cblake
- Re: NvP: s.add('x') 100M times sschwarzer
- Re: NvP: s.add('x') 100M times sschwarzer
- Re: NvP: s.add('x') 100M times HashBackupJim
- Re: NvP: s.add('x') 100M times cblake