Well let me explain. I'm a 3D artist and I have been working on movies, comercials etc.
The new mac pro uses 2x W9000 FirePro. One W9000 have 4 teraflop on single precision and 1 teraflop on double precision. Flop is like horse power on a car. Max power consuption on one W9000 is 275 watt. Quadro K6000 has 5.2 terraflop on single precision and 1.7 terraflop on double precision and max power is 225 watt. Not only the Quadro have more power but also more power per watt so the K6000 is a very power efficient card GPU. Quadro K6000 has 12 GB of vram instead of 6 GB like W9000.
Now if you look to benchmarks you'll notice that every card that has GK110 cip, outperforms every single AMD GPU out there on geometry/tesselation. Just look at unigine results. Unigine is a great benchmark to measure geometric\tessletation potential or power. Comparing W9000 and K6000 in poly pulling power or geometry/tesselation power is like comparing HD 7970 and GTX 780 or GTX Titan. Now if we talk about 3D CAD or DCC software and viewports we well see similar result like unigine. Quadro dominate most of the Specviewperf 11 tests in every review, exept AMD youtube channels obvious
). Also 90 % of the GPU raytrace engine use CUDA and not OpenCL. The only OpenCL raytrace engine out there that is used on mass post production is vray and AMD has bad support for this app. HD 7970\7950 has poor result compare to nvidia GPUs in vray.
So K6000 has more power in flops, more power per watt, less power consumption, double ram capacity, more poly/geometric/tesselation power and hell... a lot better support from ISVs.
PS. If a GPU has low power consumption automaticaly run cooler and quieter.
Hope I did answer your question. You you need more info just ask.
