diff --git a/examples/sub_threads.lua b/examples/sub_threads.lua
index cfdaf64..c18f29c 100644
--- a/examples/sub_threads.lua
+++ b/examples/sub_threads.lua
@@ -18,36 +18,48 @@
 -- OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
 -- THE SOFTWARE.
 
+-- Sub-thread processing example in Lua using llthreads - 1,000 quick sub-thread execution
+
+-- luajit sub_threads.lua
+
 local llthreads = require"llthreads"
+
+local num_threads = tonumber(arg[1] or 1000)
  
---- level 0 string literal enclosure [[  ]]
+-- level 0 string literal enclosure [[  ]] of child execution code
 local thread_code = [[  
-    print("CHILD: received params:", ...)    
-    local llthreads = require"llthreads" -- need to re-declare under this scope
+    local num_threads = ...
+    print("CHILD: received from ROOT params:", ...)    
+    local llthreads = require"llthreads" -- need to re-declare this under this scope
     local t = {} -- thread storage table
-    -- create a new child thread - with level 1 literal string [=[ ]=] enclosures, level 2 would be [==[ ]==]
+
+    -- create a new child sub-thread execution code - it requires level 1 literal string [=[ ]=] enclosures, level 2 would be [==[ ]==]
     local executed_child_code = [=[ 
-        return 'Hello from child sub-thread, new input params:', ... 
+        return "Hello from child sub-thread, new input params:", ... 
     ]=]
-    -- first child sub-thread
-    local thread1 = llthreads.new(executed_child_code , "number:", 1001, "nil:", nil, "bool:", true) 
-    assert(thread1:start()) -- start new child sub-thread
-    table.insert(t, 1, thread1) -- insert thread into thread table
-    -- second child sub-thread
-    local thread2 = llthreads.new(executed_child_code , "number:", 1002, "nil:", nil, "bool:", true) 
-    assert(thread2:start()) -- start new child sub-thread
-    table.insert(t, 1, thread2) -- insert thread into thread table
-    -- wait for all child sub-threads to complete before returning to root
-    while true do        
-        print("PARENT: child returned: ", t[1]:join()) 
-        table.remove(t,1)
-        if (#t <= 0) then break end
+
+    -- create 1000 sub-threads - which creates an incremental 30% / 20% utilization spike on the two AMD cpu cores
+    print("CHILD: Create sub threads:", num_threads)
+    for i=1,num_threads do
+        -- create child sub-thread with code to execute and the input parmeters
+        local thread = llthreads.new(executed_child_code , "number:", 1000 + i, "nil:", nil, "bool:", true) 
+        assert(thread:start()) -- start new child sub-thread
+        table.insert(t, thread) -- append the thread at the end of the thread table
+    end
+
+    -- wait (block) for all child sub-threads to complete before returning to ROOT
+    while true do
+        -- always wait on the first element, since order is not important        
+        print("CHILD: sub-thread returned: ", t[1]:join()) 
+        table.remove(t,1) -- always remove the first element
+        if (#t == 0) then break end
     end
     return ...  -- return the parents' input params back to the root
 ]]
 
 -- create child thread.
-local thread = llthreads.new(thread_code, "number:", 1234, "nil:", nil, "bool:", true)
+local thread = llthreads.new(thread_code, num_threads, "number:", 1000, "nil:", nil, "bool:", true)
 -- start joinable child thread.
 assert(thread:start())
-print("Root: child returned: ", thread:join())
+-- wait for all child and child sub-threads to finish
+print("ROOT: child returned: ", thread:join())