Reference counting garbage collection

[Marcin 'Qrczak' Kowalczyk]

Wed, 22 Aug 2001 21:25:11 -0400, Tim Peters <tim.one at home.com> pisze:

> + Simple M&S is slower given Python's extremely high dynamic memory
> throughput (*everything* in Python is heap-allocated, from execution
> frames to ints,

Would it work for Python to not heap-allocate ints but tag them in
the lowest bit, assuming pointers are even?

I know that 101 ints are allocated statically, and that it would slow
down generic object manipulation, so maybe savings would not be that
great. But it should be easy to test: just modify various macros.

This should seriously happen (if at all) after int/long unification,
because it takes away one bit of precision of ints. After proper
unification the user would not see the difference.

It works for OCaml.

-- 
 __("<  Marcin Kowalczyk * qrczak at knm.org.pl http://qrczak.ids.net.pl/
 \__/
  ^^                      SYGNATURA ZASTĘPCZA
QRCZAK