projects / mit-scheme.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 51456d8) | patch
Add fma, fused-multiply/add.
authorTaylor R Campbell <campbell@mumble.net>
Tue, 13 Aug 2019 23:25:14 +0000 (23:25 +0000)
committerTaylor R Campbell <campbell@mumble.net>
Wed, 14 Aug 2019 01:01:51 +0000 (01:01 +0000)
commit0a95c33d1160a6a284cec310ecf00ae04a29137c
tree35cd557c65d44d91bc4b5d474e29f5fafbb29808tree | snapshot
parent51456d87d356559d3ed7f6e2395a7ae6de159e42commit | diff
Add fma, fused-multiply/add.

Not yet open-coded anywhere.  Will be a huge pain on x86.  No aarch64
flonum open-coding at all yet.

(Maybe flo:fast-fma? should return false if it's not open-coded...)
doc/ref-manual/numbers.texi diff | blob | history
src/microcode/configure.ac diff | blob | history
src/microcode/flonum.c diff | blob | history
src/microcode/fma.c [new file with mode: 0644] blob
src/microcode/fma.h [new file with mode: 0644] blob
src/microcode/makegen/files-core.scm diff | blob | history
src/relnotes/fma [new file with mode: 0644] blob
src/runtime/primitive-arithmetic.scm diff | blob | history
src/runtime/runtime.pkg diff | blob | history
tests/runtime/test-arith.scm diff | blob | history
MIT/GNU Scheme, Pucked