Use a relaxed load before XCHG to not lock cache line on contention (AMD spinlock recommendation per <https://probablydance.com/2019/12/30/measuring-mutexes-spinlocks-and-how-bad-the-linux-scheduler-really-is/>)

include/boost/smart_ptr/detail/spinlock_w32.hpp

@@ -66,6 +66,11 @@ public:
 
     bool try_lock()
     {
+        if( *const_cast< long const volatile* >( &v_ ) != 0 )
+        {
+            return false;
+        }
+
         long r = BOOST_SP_INTERLOCKED_EXCHANGE( &v_, 1 );
 
         BOOST_COMPILER_FENCE