1.0.23.16: more generic assembly op optimizations on x86 and x86-64
commit2cd35e262d4be338e419114137ebf75e36e950f9
authorNikodemus Siivola <nikodemus@random-state.net>
Wed, 3 Dec 2008 16:31:10 +0000 (3 16:31 +0000)
committerNikodemus Siivola <nikodemus@random-state.net>
Wed, 3 Dec 2008 16:31:10 +0000 (3 16:31 +0000)
tree150cd740b8d689d2bad86c181b27445ed2e1d148
parent695734b2aff0e6b7ee7ea6f0424d3c0b46d088ca
1.0.23.16: more generic assembly op optimizations on x86 and x86-64

 * Cleanup: replace bunch of magic numbers with fixnum-tag-mask and
   n-fixnum-tag-bits.

 * More cases of "one test to check both argument types against fixnum".

 * Use CMOV where appropriate. (Thanks to Vitaly Mayatskikh)
src/assembly/x86-64/arith.lisp
src/assembly/x86/arith.lisp
src/runtime/alloc.c
version.lisp-expr