Update notes

Author: abhinav-nath

README.md

@@ -22,6 +22,7 @@ Simple and easy to understand code examples for most of the Concurrent APIs prov
   - [How many subtasks/threads to break the original task into?](./notes/performance-in-multithreading.md#n--)
     - [Hyper-threading](./notes/performance-in-multithreading.md#hyper-threading)
   - [Inherent cost of Parallelization and Aggregation](./notes/performance-in-multithreading.md#inherent-cost-of-parallelization-and-aggregation)
+  - [Can we always break any task into multiple subtasks?](./notes/performance-in-multithreading.md#can-we-always-break-any-task-into-multiple-subtasks)
 - [Throughput](./notes/performance-in-multithreading.md#throughput)
 
 #### Thread synchronization and locks

notes/performance-in-multithreading.md

@@ -70,4 +70,11 @@ We cannot run all the threads 100% in parallel, but we can get close to that!
 
 ### Inherent cost of Parallelization and Aggregation
 
-![Inherent cost of Parallelization and Aggregation](../images/latency_5.png)
+![Inherent cost of Parallelization and Aggregation](../images/latency_5.png)
+
+
+### Can we always break any task into multiple subtasks?
+
+The answer is NO!
+
+- Small and trivial tasks are not worth breaking and running in parallel.