Increase min llvm version for assembly test

There seems to be a small difference in generated code.
In CI with llvm 16:

cmgt v0.16b, v1.16b, v0.16b

Locally with bundled llvm:

cmge v0.16b, v0.16b, #0

tests/assembly/simd-intrinsic-mask-reduce.rs

@@ -4,7 +4,7 @@
 //@ [x86] needs-llvm-components: x86
 //@ [aarch64] compile-flags: --target=aarch64-unknown-linux-gnu
 //@ [aarch64] needs-llvm-components: aarch64
-//@ [aarch64] min-llvm-version: 15.0
+//@ [aarch64] min-llvm-version: 18.0
 //@ assembly-output: emit-asm
 //@ compile-flags: --crate-type=lib -O
 
@@ -32,7 +32,7 @@ extern "rust-intrinsic" {
 pub unsafe fn mask_reduce_all(m: mask8x16) -> bool {
     // x86: movdqa
     // x86-NEXT: pmovmskb
-    // aarch64: cmge
+    // aarch64: cmgt
     // aarch64-NEXT: umaxv
     simd_reduce_all(m)
 }