Benchmark Results: Autodesk 3ds Max 2011, Cinebench, VUE, And MatchMover
This is the 3ds Max space scene (oddly, included in the SPECapc test for 3ds Max 9) that we include as a component of our processor reviews. While we don’t have another single-socket workstations to compare it to, there are multiple other reference points if you look through older CPU stories.
Maxon Cinebench r11.5
Cinebench is a two-part test that tests rendering and OpenGL performance as a standalone application, based on Maxon’s Cinema4D 3D animation software.
The OpenGL performance in CineBench comes in at 75% higher than the Z400--a respectable increase, but not the 100% or higher increase most of the other tests exposed. This is likely due to the smaller difference in graphics subsystem performance versus the more significant CPU disparity.
Cinebench rendering is a different story. It shows a 172% performance increase between the two systems. Once again, rendering takes full advantage of the additional processor cores (and threads via Hyper-Threading), while the OpenGL test is limited by the graphics card.
e-on VUE 8 PLE
This landscape render test throws a high-polygon landscape with procedurally-generated trees at the system. If the trees were not procedurally generated, the scene would likely overwhelm the system’s memory subsystem.
The Vue render shows a 132% improvement, consistent with the other render test results.
Autodesk MatchMover 2011
This test takes footage of a moving camera in 2D (a 720p clip of a brief walk on the Walk of Fame in Hollywood) and tracks the movement in 3D space. Even though the tests were done with the same software version, they were performed on different operating systems.
Since this test is also I/O heavy, it does not show the improvement the purely CPU-based tests (like rendering) do, and receives no additional benefit from the higher-performance graphics card. The 46% improvement is fairly close to the 56% difference in processor speed between the two systems, possibly indicating that the test is not well-optimized for threading, either.
