//===---------------------------------------------------------------------===//
-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);
+}
+
+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.
-and likewise for the single precision versions.
+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.
+
+//===---------------------------------------------------------------------===//
+
+This function should be matched to haddpd when the appropriate CPU is enabled:
+
+#include <x86intrin.h>
+double f (__m128d p) {
+ return p[0] + p[1];
+}
-Currently, SelectionDAGBuilder doesn't turn calls to these functions into the
-corresponding nodes and some targets (including X86) aren't ready for them.
+similarly, v[0]-v[1] should match to hsubpd, and {v[0]-v[1], w[0]-w[1]} should
+turn into hsubpd also.
//===---------------------------------------------------------------------===//