Everyone gets excited when they hear about new software, but it takes a powerful video card to see what’s being touted on the display. Hence the rollout of Nvidia’s two newest Quadro video cards at SIGGRAPH 2015: the double-slot Quadro M5000 and its slightly less powerful sibling, the single-slot Quadro M4000. Both cards are based on Nvidia’s Maxwell 2 architecture as found in the flagship Quadro M6000 card, and while both can light up four 4K displays thanks to having four DisplayPort 1.2 connectors each, the M5000 has an added DVI-I connector. They both possess 8 Gigabytes of GDDR5 GPU memory with the M4000’s GPU memory’s speed being 192 GB/sec to the faster M5000’s 211 GB/sec.
Another differentiating factor between the two cards comes from the number of processing cores, with the 5000’s 2,408 Cuda cores nearly double that of the 4000’s 1,664.
But it is in the use of Nvidia’s DesignWorks, a suite of tools and technologies designed especially for professional graphics and advanced rendering applications, that a new set of libraries, algorithms and tools will enable users working with the video cards to work more effectively and efficiently with physically based rendering and physically based materials. DesignWorks includes Nvidia’s Iray SDK (software development kit) for simulating the physical behavior of light and materials, the MDL SDK (Material Definition Language) for materials and lights that can be shared across renderers, the VMaterials Library consisting of hundreds of ready to use real world materials, the OptiX framework for programmable ray tracing and the DesignWorks VR tool set actualized for application developers for enabling a virtual reality workflow.
The Quadro M5000 has a boost clock speed of around 1,050MHz, with FP32 performance rated at 4.3 TFLOPs. The M4000’s boost clock speed is about 780MHz, with an FP32 performance of 2.6 TFLOPs. Prices for the two video cards have yet to be announced.