Displaying 2 results from an estimated 2 matches for "e7f4b1c8".
2012 Apr 16
0
[LLVMdev] Representing -ffast-math at the IR level
...specify specific
fpmath metadata when creating an operation.
Ciao, Duncan.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: fastm.diff
Type: text/x-patch
Size: 18788 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20120416/e7f4b1c8/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: fastm-clang.diff
Type: text/x-patch
Size: 1497 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20120416/e7f4b1c8/attachment-0001.bin>
-------------- next...
2012 Apr 14
9
[LLVMdev] Representing -ffast-math at the IR level
The attached patch is a first attempt at representing "-ffast-math" at the IR
level, in fact on individual floating point instructions (fadd, fsub etc). It
is done using metadata. We already have a "fpmath" metadata type which can be
used to signal that reduced precision is OK for a floating point operation, eg
%z = fmul float %x, %y, !fpmath !0
...
!0 = metadata