C2: vectorization fails on simple ByteBuffer loop

This showed up with some panama benchmarks.
The following code snippet (extracted from a ByteBuffer benchmark) is not vectorized:

    public static void testByteLong4(byte[] dest, long[] src, int start, int stop) {
        for (int i = start; i < stop; i++) {
            UNSAFE.putLongUnaligned(dest, 8 * i + baseOffset, src[i]);
        }
    }