rp2: Enable optimisations (comp goto, map cache, fast attr).

Computed goto costs 1800 bytes for 5-10% performance.

Map caching and attr fast path costs 130 bytes for up to 30%.

Net effect of those three optimisations:
bm_chaos.py         +16.059% (+/-0.09%)
bm_fannkuch.py      +11.145% (+/-0.01%)
bm_fft.py           +14.604% (+/-0.01%)
bm_float.py         +26.849% (+/-0.08%)
bm_hexiom.py        +34.039% (+/-0.03%)
bm_nqueens.py       +18.333% (+/-0.06%)
bm_pidigits.py       +4.472% (+/-0.03%)
misc_aes.py         +28.765% (+/-0.09%)
misc_mandel.py      +27.116% (+/-0.05%)
misc_pystone.py     +40.299% (+/-0.20%)
misc_raytrace.py    +22.812% (+/-0.07%)

Also enable other EXTRA-level optimisations (module const, return_if_expr,
triple_tuple_assign, factorial, mpz bitwise).

Signed-off-by: Jim Mussared <[email protected]>

Author: jimmo

ports/rp2/mpconfigport.h

@@ -55,16 +55,11 @@
 #define MICROPY_EMIT_INLINE_THUMB_FLOAT         (0)
 #define MICROPY_EMIT_INLINE_THUMB_ARMV7M        (0)
 
+// Optimisations
+#define MICROPY_OPT_COMPUTED_GOTO               (1)
+
 // Features currently overriden for rp2, planned to be brought in line with
 // other ports
-#define MICROPY_COMP_MODULE_CONST               (0)
-#define MICROPY_COMP_RETURN_IF_EXPR             (0)
-#define MICROPY_COMP_TRIPLE_TUPLE_ASSIGN        (0)
-#define MICROPY_OPT_COMPUTED_GOTO               (0)
-#define MICROPY_OPT_LOAD_ATTR_FAST_PATH         (0)
-#define MICROPY_OPT_MAP_LOOKUP_CACHE            (0)
-#define MICROPY_OPT_MATH_FACTORIAL              (0)
-#define MICROPY_OPT_MPZ_BITWISE                 (0)
 #define MICROPY_PY_BUILTINS_EXECFILE            (0)
 #define MICROPY_PY_BUILTINS_NOTIMPLEMENTED      (0)
 #define MICROPY_REPL_EMACS_KEYS                 (0)