Harpertown: New 45nm Xeons from Intel
Posted on: 09/17/2007 05:00 AM

3ds Max 9 64-bit
Autodesk∆ 3ds Max∆ 9 software builds on our promise to deliver state-of-the-art tools for creative and media professionals. Developed as a total animation package with a deep, productive feature set designed to accelerate workflow, 3ds Max is the leader in 3D animation for game development, design visualization, visual effects, and education.
3ds Max is a widely used, scalable application that is most frequently run on high-end, multi-processor workstations and servers like the ones we're looking at today. In the past I have used a variety of test files for benchmarking. This time around, I used more of a "real world" test scene created by one of our forum and IRC regulars (opus13).



The new 3ds Max workload scales pretty well with cores and clock speed. The Harpertown system's eight cores almost double the performance over Woodcrest's four cores, with the slower (clock speed) Clovertowns falling right in between.

Power Usage

I touched on the whole power usage thing in my last article, but I'll be expanding on it quite a bit in this one. First up, I will present the actual wattage used over a predefined test period, as logged by the Exotech power meter. Remember, only the machine itself is plugged in to the meter.

The "test period" here, and in all subsequent graphs, is a period that starts five seconds before the machine leaves the idle state and starts the workload. The period ends five seconds after the slowest machine in the test returns to idle after completing the workload.



As you can see, the power usage graphs favor the faster machine. But, that doesn't give us the overall image of the "performance-per-watt" that everyone seems to be so interested in lately. So...

"Performance-per-Watt"

In this graph, I break down the power usage into Joules or, Watt-seconds. "Total Joules" represents the amount of energy used by the machine over the entire test period (as previously defined above). The "Workload Joules" represent the amount of power required to complete the actual benchmark, from the second the machine leaves the idle state until the second it returns to the idle state after completing the test.



"Workload Joules" are probably the best representation of true "performance-per-watt" of the given test systems. Faster machines return to idle quicker, thus using less power in the long run. But, depending on the end user's actual usage model, getting things done faster will probably help to save more power in the end.


Printed from 2CPU.com (http://www.2cpu.com/contentteller.php?ct=articles&action=pages&page=harpertown_new_45nm_xeons_from_intel,5.html)