perf(msgpack): add SIMD integer optimizations

- Add SIMD-accelerated byte swap functions for u32/u64
- Implement memory alignment optimizations for better performance
- Add batch conversion functions for integer arrays
- Replace std.mem.writeInt with optimized fast paths
- Add comprehensive alignment and batch conversion tests

Author: jinzhongjia