diff options
Diffstat (limited to 'lib/Target/X86/README-SSE.txt')
| -rw-r--r-- | lib/Target/X86/README-SSE.txt | 20 |
1 files changed, 13 insertions, 7 deletions
diff --git a/lib/Target/X86/README-SSE.txt b/lib/Target/X86/README-SSE.txt index 7d901af..624e56f 100644 --- a/lib/Target/X86/README-SSE.txt +++ b/lib/Target/X86/README-SSE.txt @@ -923,15 +923,21 @@ The insertps's of $0 are pointless complex copies. //===---------------------------------------------------------------------===// -If SSE4.1 is available we should inline rounding functions instead of emitting -a libcall. +[UNSAFE FP] -floor: roundsd $0x01, %xmm, %xmm -ceil: roundsd $0x02, %xmm, %xmm +void foo(double, double, double); +void norm(double x, double y, double z) { + double scale = __builtin_sqrt(x*x + y*y + z*z); + foo(x/scale, y/scale, z/scale); +} -and likewise for the single precision versions. +We currently generate an sqrtsd and 3 divsd instructions. This is bad, fp div is +slow and not pipelined. In -ffast-math mode we could compute "1.0/scale" first +and emit 3 mulsd in place of the divs. This can be done as a target-independent +transform. -Currently, SelectionDAGBuilder doesn't turn calls to these functions into the -corresponding nodes and some targets (including X86) aren't ready for them. +If we're dealing with floats instead of doubles we could even replace the sqrtss +and inversion with an rsqrtss instruction, which computes 1/sqrt faster at the +cost of reduced accuracy. //===---------------------------------------------------------------------===// |
