Use own implementation of pow_int_ instead of llvm

Remove LLVM intrinsics code related to `Num.powInt`.

compiler/builtins/bitcode/src/lib.rs

@@ -10,3 +10,25 @@
 pub fn i64_to_f64_(num: i64) -> f64 {
     num as f64
 }
+
+#[no_mangle]
+pub fn pow_int_(mut base: i64, mut exp: i64) -> i64 {
+    let mut acc = 1;
+
+    while exp > 1 {
+        if (exp & 1) == 1 {
+            acc *= base;
+        }
+        exp /= 2;
+        base *= base;
+    }
+
+    // Deal with the final bit of the exponent separately, since
+    // squaring the base afterwards is not necessary and may cause a
+    // needless overflow.
+    if exp == 1 {
+        acc *= base;
+    }
+
+    acc
+}