Add LTO in configure, add more benchmarks

bench/include/BH/Bench.h

@@ -1,6 +1,7 @@
 #ifndef BH_BENCH_H
 #define BH_BENCH_H
 
+
 typedef struct BH_Bench BH_Bench;
 typedef void (*BH_BenchCallback)(BH_Bench *);
 
@@ -22,4 +23,19 @@ int BH_BenchIter(BH_Bench *state);
 
 int BH_BenchRun(void);
 
+
+#if defined(__GNUC__) || defined(__clang__)
+#define BH_BenchDoNotOptimize(arg) \
+    do { \
+        __asm__ __volatile__("" : : "r"(*((char volatile*) &arg)) : "memory"); \
+    } while (0)
+#else
+    #define BH_BenchDoNotOptimize(arg) \
+        do { \
+            volatile void* _ptr = (void*)&(arg); \
+            (void)_ptr; \
+        } while (0)
+#endif
+
+
 #endif /* BH_BENCH_H */