Tried parallelizing the code, but the compiler can already do that automatically

.clang-format

@@ -1,7 +1,10 @@
-BasedOnStyle: GNU
+BasedOnStyle: LLVM
 IndentWidth: 4
 TabWidth: 4
 UseTab: Always
 BreakBeforeBraces: Attach
 ColumnLimit: 100
-AlignReturnType: Always
+PenaltyReturnTypeOnItsOwnLine: 1000000
+AlwaysBreakAfterDefinitionReturnType: None
+AlwaysBreakAfterDefinitionReturnType: None
+SeparateDefinitionBlocks: Always

Makefile

@@ -6,7 +6,7 @@ ts := $(shell /usr/bin/date "+%d-%m__%H_%M_%S")
 .DEFAULT_GOAL = MD
 
 MD: $(SRC)/MD.cpp
-	$(CC) $(CFLAGS) $(SRC)MD.cpp -lm -march=native -O2 -ftree-vectorize -funroll-loops -pg -g -o ./out/MD
+	$(CC) $(CFLAGS) $(SRC)MD.cpp -lm -march=native -mtune=native -mavx -O2 -ftree-vectorize -funroll-loops -pg -g -o ./out/MD
 
 MDorig: $(SRC)/MD-original.cpp
 	$(CC) $(CFLAGS) $(SRC)MD-original.cpp -lm -O2 -pg -o ./out/MD-original