common-lisp.net / projects/cmucl/cmucl.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 441a76a) | patch
Floating-point micro-optimizations
authorRaymond Toy <toy.raymond@gmail.com>
Sun, 19 Aug 2012 14:58:06 +0000 (07:58 -0700)
committerRaymond Toy <toy.raymond@gmail.com>
Sun, 19 Aug 2012 14:58:06 +0000 (07:58 -0700)
commit5293d44e876c4d08a1a1f19b037a8afd54961eb8
tree04ddd5a14dcef262100ecd04ae208d4ef59ed3d2tree
parent441a76a6b156e8eedc5ef9bb72c5772d89b657bccommit | diff
Floating-point micro-optimizations

o Convert x/n to x*(1/n) when n is a power of two since 1/n has an
  exact representation.
o Convert 2*x to x+x.
src/compiler/float-tran.lisp diff | blob | blame | history
CMU Common Lisp