[llvm] [GlobalISel] Add G_ABS computeKnownBits, add ComputeKnownBitsCache assertion to computeNumSignBits (PR #154413)

[Pragyansh Chaturvedi via llvm-commits]

================
@@ -249,10 +249,20 @@ define <2 x i64> @smull_zext_v2i32_v2i64(ptr %A, ptr %B) nounwind {
 ; CHECK-GI-NEXT:    movi d0, #0x00ffff0000ffff
 ; CHECK-GI-NEXT:    mov v1.s[1], v2.s[0]
 ; CHECK-GI-NEXT:    and v0.8b, v1.8b, v0.8b
-; CHECK-GI-NEXT:    mov v1.s[0], v0.s[0]
-; CHECK-GI-NEXT:    mov v1.s[1], v0.s[1]
+; CHECK-GI-NEXT:    mov w8, v0.s[0]
----------------
r41k0u wrote:

This was a goof, I was increasing the depth while calling computeKnownBitsImpl for computeNumSignBits, which was causing this. Fixing that and reupdating the tests removed this.

https://github.com/llvm/llvm-project/pull/154413