Very interesting post...

Andreas Strangmüller wrote:

Conditions: 
Hardware: Dual AMD Opteron 6376, 32 x 2.3 GHz (Turbo Core off) 
OS: Windows 7 Pro 64-Bit 
GUI: no 
Settings: all engines default settings 
Large Tables: no 
Position: starting position 
Time: 20 seconds 

UCI commands:
setoption name threads value 1 (to 32) 
go movetime 20000 

The tests were run in console mode. 

Here the values from 1 to 32 threads, starting position, with 20 seconds of computing time. 
nps = nodes per second 

Threads factor: Komodo, Houdini, Stockfish and Zappa Table-factor-1-32-threads 

Komodo, Houdini and Zappa are almost equal up to 16 threads (factor 11.94 - 11.37 - 11.34). 
Stockfish DD and also the latest Stockfish version lies somewhat behind (factor 8.01 - 9.79). 

Komodo scales still excellent beyond 16 threads. Also Zappa shows a very good SMP implementation. 
Beyond 16 threads Houdini and Stockfish DD benefit much less than the other tested engines. 

Increase from 16 to 32 threads: 

Komodo TCECr (11,94 - 20,60 = 73%
Zappa Mexico II (11,37 - 16,46 = 45%
Stockfish 140513 ( 9,79 - 14,21 = 45%
Stockfish DD ( 8,01 - 10,18 = 27%
Houdini 4 Pro (11,34 - 13,48 = 19%

Threads factor: Komodo, Houdini, Stockfish and Zappa Graphic-factor-1-32-threads