From 57003a6768146c20befe5fc24d8a9b9bcb7ca0fd Mon Sep 17 00:00:00 2001 From: Benjamin Kramer Date: Mon, 19 Mar 2012 00:43:34 +0000 Subject: [PATCH] Add a note for -ffast-math optimization of vector norm. llvm-svn: 153031 --- llvm/lib/Target/X86/README-SSE.txt | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) diff --git a/llvm/lib/Target/X86/README-SSE.txt b/llvm/lib/Target/X86/README-SSE.txt index a581993c3c61..624e56fa0f64 100644 --- a/llvm/lib/Target/X86/README-SSE.txt +++ b/llvm/lib/Target/X86/README-SSE.txt @@ -922,3 +922,22 @@ _test2: ## @test2 The insertps's of $0 are pointless complex copies. //===---------------------------------------------------------------------===// + +[UNSAFE FP] + +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); +} + +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. + +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. + +//===---------------------------------------------------------------------===//