[llvm] [memprof] Speed up caller-callee pair extraction (PR #116184)

[Kazu Hirata via llvm-commits]

================
@@ -1303,6 +1303,10 @@ Error IndexedMemProfReader::deserializeV3(const unsigned char *Start,
   FrameBase = Ptr;
   CallStackBase = Start + CallStackPayloadOffset;
 
+  // Compute the number of elements in the radix tree array.
+  RadixTreeSize = (RecordPayloadOffset - CallStackPayloadOffset) /
----------------
kazutakahirata wrote:

The overestimation like that (that is, reality N elements, estimate N + 1 elements) is OK.  The only use of the estimate is to reserve enough bits in a BitVector introduced in this patch:

```
BitVector Worklist(RadixTreeSize);
```

I don't use the estimate to data structures allocated by other people.

https://github.com/llvm/llvm-project/pull/116184