diff --git a/Makefile b/Makefile
index 4d3a205..e61d401 100644
--- a/Makefile
+++ b/Makefile
@@ -1,3 +1,4 @@
+CONFIG_NATIVE := true
 CONFIG_CXX_ALLOCATOR := true
 CONFIG_UBSAN := false
 CONFIG_SEAL_METADATA := false
@@ -8,6 +9,11 @@ endef
 
 CPPFLAGS := -D_GNU_SOURCE
 SHARED_FLAGS := -O2 -flto -fPIC -fvisibility=hidden -fno-plt -pipe -Wall -Wextra $(call safe_flag,-Wcast-align=strict) -Wcast-qual -Wwrite-strings
+
+ifeq ($(CONFIG_NATIVE),true)
+    SHARED_FLAGS += -march=native
+endif
+
 CFLAGS := -std=c11 $(SHARED_FLAGS) -Wmissing-prototypes
 CXXFLAGS := -std=c++14 $(SHARED_FLAGS)
 LDFLAGS := -Wl,-z,defs,-z,relro,-z,now,-z,nodlopen,-z,text
diff --git a/README.md b/README.md
index b76e66c..38d3de2 100644
--- a/README.md
+++ b/README.md
@@ -60,6 +60,10 @@ make command as follows:
 
 The available configuration options are the following:
 
+* `CONFIG_NATIVE`: `true` (default) or `false` to control whether the code is
+  optimized for the detected CPU on the host. If this is disabled, setting up a
+  custom `-march` higher than the baseline architecture is highly recommended
+  due to substantial performance benefits for this code.
 * `CONFIG_CXX_ALLOCATOR`: `true` (default) or `false` to control whether the
   C++ allocator is replaced for slightly improved performance and detection of
   mismatched sizes for sized deallocation (often type confusion bugs). This