[llvm] [DAGCombiner] Enable constant folding of (bitcast int_c0) -> fp_c0 (PR #82289)

[Simon Pilgrim via llvm-commits]

================
@@ -320,7 +320,8 @@ define <3 x double> @extvselectsetcc_crash(<2 x double> %x) {
 ; X64-LABEL: extvselectsetcc_crash:
 ; X64:       # %bb.0:
 ; X64-NEXT:    vcmpeqpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm1
-; X64-NEXT:    vmovsd {{.*#+}} xmm2 = [1.0E+0,0.0E+0]
+; X64-NEXT:    movabsq $4607182418800017408, %rax # imm = 0x3FF0000000000000
----------------
RKSimon wrote:

vector loads vs rematerialization vs scalar constants has never been ideal on x86

https://github.com/llvm/llvm-project/pull/82289