Futuremark’s 3DMark 11 Review
Posted by Robert Dow on December 8th 2010 | Discuss
Categories:
Software Review
Tags:
benchmark
futuremark
First look at new benchmark
DirectX 11 has been with us since Windows 7 was introduced in late October 2009, almost 14 months ago. And it’s not as if no one didn’t know it was coming, or no major OEM or game developer didn’t know what the specifications were, and even have access to some of the tools. So it’s hard to understand why so few games are available that support DirectX 11 other than they are cheap console ports and were never designed for DirectX 11.
OK, that explains the lazy PC hating game developers, but not a company whose life blood depends on being state of the art. So Futuremark was way late to the DX11 party, but as Jon always says when he’s late, “I’m worth waiting for.”
Futuremark’s 3DMark 11 is definitely worth waiting for. The span of tests, the art work, and the test data are impressive.
And the stress the games put on today’s AIBs is also impressive. It’s doubtful any of the GPU and AIB suppliers are going to sing the praises of 3DMark 11 since the FPS scores are in the teens. We’re assuming Futuremark expects the GPUs to grow into the test and maybe get better scores with their next generation parts.
3DMark 11 tests for volumetric lighting, tessellation, and physics, either combined, or individually. Already we’ve heard from one GPU supplier who criticizes the way 3DMark 11 tests for tessellation – the mud-slinging has started.
3DMark 11 tests
Graphics Test 1 has lots of light sources with many casting shadows in the scene. Volumetric lighting illuminates the cloudy water. Post processing is used for various camera lens effects. There is no tessellation in this test.
Graphics Test 2 has a few lights which cast shadows, volumetric lighting through cloudy water and tessellation on the structures, corals, rock and sea bed. Post processing adds depth of field and other camera lens effects.
Graphics Test 3 features tessellation on the pillars, statues and some of the vegetation. Volumetric lighting from the sun casts shadows in the scene. Post processing adds depth of field and other lens effects.
Graphics Test 4 also uses tessellation on the pillars, statues and some of the vegetation. There is volumetric lighting and multiple light sources with several casting shadows. Post processing adds various camera lens effects.
Physics Test focuses on CPU performance by simulating rigid body physics with a large number of objects. This test runs at a fixed screen resolution for all presets. There is no post processing, volumetric lighting or tessellation.
Combined Test. This test combines CPU and GPU workloads. The CPU handles rigid-body physics while the GPU is tasked with volumetric lighting, tessellation, post processing as well as simulating soft-body physics using DirectCompute.
As the test scores indicate this new test is stressful on the current batch of AIBs.
There are three versions of the benchmark, the basic edition which is free and will give you a basic score, the Advanced Edition which sells for a mere $19.95 and gives you everything but demo looping and image quality tools, and the crème de la crème Professional Edition for $995 which gives you everything including priority support.
The new benchmark has three preset tests, and two custom ones
Entry preset (E) is designed for benchmarking with a low level of load on the graphics card. The benchmark runs in 1024 x 600 resolution making it suitable for most entry level DirectX 11 capable systems including notebooks and netbooks.
Performance preset (P) is for benchmarking with a moderate load on the AIB. The benchmark runs in 1280 x 720 (720p) making it suitable for most DirectX 11 capable gaming PCs, though at launch high end hardware may be required to achieve a fluid frame rate.
Extreme preset (X) This preset is designed for benchmarking with a very heavy load on the graphics card. The benchmark runs in 1920 x 1080 (1080p). The Extreme preset extends the lifetime of the benchmark by representing the likely loads used by high end games in years to come. At launch it offers enthusiasts a suitable benchmark load for competing at the extreme end of system performance.
Custom settings can be run with customized settings offering a fine degree of control over resolution, anti-aliasing, tessellation, shadow quality, texture quality and many other graphical settings. When using a resolution with an aspect ratio other than 16:9, the benchmark runs in a 16:9 letterbox.
Individual test scores
The Advanced and Professional Edition users can see the individual scores for the graphics, physics and combined tests on 3dmark.com. These scores can be used to compare a single component such as the GPU or CPU when using the online services on 3dmark.com.
To arrive at the graphics test score each of the Graphics Tests produces a result in frames per second. The Graphics Score is calculated by multiplying the harmonic mean of four test results with a scaling constant: It is the scaling constant for the graphics score and the fps results for Graphics Tests 1-4. Futuremark says the scaling brings the score in line with traditional 3DMark score levels, but from a FPS basis it doesn’t, the scores are way low.
What do we think?
Futuremark has competition now. Unigine, SiSoft Sandra, and others will challenge 3DMark 11 for mind share and significance. Futuremark has the brand and is established at most OEMs. However, the new scoring is always like a new pair of shoes, the user has to get broken in to them. So there’s going to a lot whopping and hollowing till the GPU and AIB suppliers get comfortable with the new benchmark.
It would sure be a lot easier on the industry and the users if that could be avoided. We released this issue of TechWatch the same day the 3DMark 11 and the new Nvidia GTX 570 were released and by the time you’ve read this the forums and fan boyz will have lit up the internet with their opinions and remedies.
