Update Optimize performance.md

Author: codehariom

Optimize performance.md

@@ -1,5 +1,5 @@
 
-***Cache data objects***
+## Cache data objects***
 
  **Define a cached function:**
 
@@ -30,7 +30,7 @@ clears values from all cached functions, either in-memory or on-disk.
 
 These snippets demonstrate how Streamlit allows you to cache expensive computations or data retrievals to improve performance.
 
-***Cache global resources***
+## Cache global resources
 
 E.g. TensorFlow session, database connection, etc.
 
@@ -53,4 +53,26 @@ Create and return a non-data object
     foo.clear()
 #Clear all global resources from cache
 
-    st.cache_resource.clear()
+    st.cache_resource.clear()
+
+ ## Deprecated caching
+
+ These Streamlit snippets showcase caching with the `@st.cache` decorator:
+
+**Define a cached function:**
+
+    `@st.cache
+decorator marks the function `foo()` for caching.
+
+     Inside `foo()
+perform some expensive computation and return data.
+
+ **Execute the cached function:**
+
+    `d1 = foo(ref1)` executes `foo()` with the argument `ref1` and caches the result in `d1`.
+
+    `d2 = foo(ref1)` retrieves the cached result for the same argument, so `foo()` is not executed again, and `d1 == d2`.
+
+    `d3 = foo(ref2)` calls `foo()` with a different argument `ref2`, so the function executes again.
+
+These snippets demonstrate how Streamlit allows you to cache function results to avoid redundant computation and improve performance.