diff options
| author | Yukihiro "Matz" Matsumoto <[email protected]> | 2021-08-20 16:57:33 +0900 |
|---|---|---|
| committer | Yukihiro "Matz" Matsumoto <[email protected]> | 2021-08-21 15:25:58 +0900 |
| commit | 36632f55f418bd711112b72d7fe1b6b1745058e5 (patch) | |
| tree | 30a9a0e83fced5759b5da8f2354b7ba02280e2b8 /include/mrbconf.h | |
| parent | 4ea80913c27264df0c7088367109c5a0678879f1 (diff) | |
| download | mruby-36632f55f418bd711112b72d7fe1b6b1745058e5.tar.gz mruby-36632f55f418bd711112b72d7fe1b6b1745058e5.zip | |
boxing_word.h: embed `mrb_float` in `mrb_value` if possible.
Embedding reduce memory consumption, sacrificing precision. It clips least
significant 2 bits from `mrb_float`, so if you need to keep float precision,
define `MRB_USE_FLOAT_FULL_PRECISION`.
`MRB_WORD_BOXING` and `MRB_INT64`:
`mrb_float` (`double`) is embedded in `mrb_value` clipped last 2 bits.
`MRB_WORD_BOXING` and `MRB_INT64` and `MRB_USE_FLOAT_FULL_PRECISION`:
`mrb_float` is allocated in the heaps wrapped by `struct RFloat`.
`MRB_WORD_BOXING` and `MRB_INT32` and `MRB_USE_FLOAT32`:
`mrb_float` (`float`) is embedded in `mrb_value` clipped last 2 bits.
In addition, to reserve bit space in the `mrb_value`, maximum inline
symbol length become 4 (instead of 5) in the configuration.
`MRB_WORD_BOXING` and `MRB_INT32`:
Assume `MRB_USE_FLOAT_FULL_PRECISION` and allocate Float values in heap.
Diffstat (limited to 'include/mrbconf.h')
| -rw-r--r-- | include/mrbconf.h | 3 |
1 files changed, 3 insertions, 0 deletions
diff --git a/include/mrbconf.h b/include/mrbconf.h index 5d54a1c22..7b29cb0ba 100644 --- a/include/mrbconf.h +++ b/include/mrbconf.h @@ -42,6 +42,9 @@ #error Cannot define MRB_USE_FLOAT32 and MRB_NO_FLOAT at the same time #endif +/* if defined mruby allocates Float objects in the heap to keep full precision if needed */ +//#define MRB_USE_FLOAT_FULL_PRECISION + /* add -DMRB_NO_METHOD_CACHE to disable method cache to save memory */ //#define MRB_NO_METHOD_CACHE /* size of the method cache (need to be the power of 2) */ |