The professional's tool

Mirada's workloads for in-car graphics and OpenGL ES 2.0 features emulate the functionality of future automotive graphics applications. The benchmark enables tomorrow’s 3D benchmarking today, before any actual OpenGL ES 2.0 in-car graphics application are even announced and available for automotives. Running these test contents on today’s development and prototype boards allows early hardware testing, evaluation and verification with realistic content and results eventually with better performing hardware for the end users.

Automotive Benchmark Application Tests

Dashboard test is a high-quality per-frame rendered instrumentation. Test will render a realistic real-time vehicle dashboard with an animated speedometer, tachometer, odometer and a fuel gauge. All instruments are rendered in 3d view. The displays are rendered to remind of traditional instruments. The needles will be animated and the display also contains other dynamic information system display elements. Dashboard test has two content levels, low and high. While the polygon load is identical for both tests, high end test uses far more fragment operations. Low load is reminiscent of single pass GL ES 1.x contents. It will contain a Blinn-Phong shading and possibly a texture lookup. The high-end test uses a photorealistic shader to produce a lively, high-quality display with realistic materials.

Navigation test is based on the common ‘navigator’ utility usage scenario. A positioning signal is fed from a GPS location system to a navigation application that will then present a dynamic map view based on the location information and will offer routing instructions for street navigation. The navigation test renders a street map view with 3D-buildings in part of the city blocks. A blip representing the automobile navigates through the streets through a predetermined route with directional instructions.

The user interface test renders a hypothetical user interface showing vehicle information system display and environmental controls with dynamic functionality and various 3D user interface models.

Automotive Benchmark Feature Tests:

• Polygon throughput: The test rates the polygon throughput of the OpenGL ES 2.0 implementation. The test renders 300,000 triangles per frame at maximum and tries to use as less as possible of the screen space with very low pixel weight to avoid fillrate load. Depth-buffer will be disabled.
• Fillrate: The test renders a scene with a high ratio of polygon interlap that tests the OpenGL ES 2.0 implementation’s fill-rate. The test renders triangles as less as possible to fill the screen with very high pixel weight. Depth-buffer will be disabled.