Added more content to docs 2

docs/HTML/DataFrame.html

@@ -1425,7 +1425,7 @@ <H2><font color="blue">API Reference with code samples</font></H2>
 <BR><HR COLOR="Gray" SIZE="5">
 
   <H2><font color="blue">Multithreading</font></H2>
-    In general, multithreading could be very unintuitive. Often you think by using multithreading you enhance the performance of your program. But in fact, you are hindering it. To do effective multithreading, you must do two things repeatedly; measure and adjust. It is recommended to start with a single-threaded version and when that is <I>working correctly</I>, establish a baseline, take measurements, and implement a multithreading solution.<BR>
+    In general, multithreading could be very unintuitive. Often you think by using multithreading you enhance the performance of your program. But in fact, you are hindering it. To do effective multithreading, you must do two things repeatedly; measure and adjust. In general (rule of thumb), you should use multithreading in two contradictory situations. First, when you have intensive CPU-bound operations like mathematical equations that can independently utilize different cores. Second, when you have multiple I/O-bound operations that can go on independently while they wait for each other. They key word here is <I>independently</I>. You must also realize that multithreading has an inherent overhead that not only affects your process but also other processes running on the same node. It is recommended to start with a single-threaded version and when that is <I>working correctly</I>, establish a baseline, take measurements, and implement a multithreading solution.<BR>
 DataFrame uses multithreading extensively and provides granular tools to adjust your environment. Let’s divide the multithreading subject in DataFrame into two categories:<BR>
 
   <H4>1. User Multithreading</H4>