For your needs, choose the P3000.
Aside from driver differences, there are some underlying differences in the GPUs. If you just play AAA 3D games on them, there is no difference, but if you do simulations like CAD, complex splatting and n-body, even a far cheaper P2000 GPU can beat a GTX 1080 Ti.
Why? Its because the single precision part of the GPU cores are functional in both GPUs but only the double precision scheduler/handler is functional in the Quadro workstation GPUs, allowing much faster DP calculations, and the GTX series suffer from extremely inefficient driver software based DP calculations than hardware based.
GTX 1070-
FP32 Single Precision performance:
7500 GFLOPS
FP64 Double Precision performance:
~234 GFLOPS (1/32 of FP32)
"Who cares?"
Games mainly use single precision as what matters is high framerate (note overclocking and instability), not extreme precision and reliability.
Quadro P3000
FP32 Single Precision performance:
7500 GFLOPS
FP64 Double Precision performance:
3750 GFLOPS (1/2 of FP32)
Hardware implementation of DP boosts DP calculation rates (Double Precision Floating Point Units DP FPUs)
Double precision calculations are used in simulations as any miscalculations can cost the user plenty.
SP vs DP, look at "Mandelbrot FP32 vs FP64"
http://www.ozone3d.net/public/jegx/201009/geexlab_fractal_julia_fp64_glsl.jpg
Note that since most of THW rate GPUs by gaming performance, GTX 1070 performs equal to Quadro P3000. But in the realm of high precision graphics, Quadro P3000 is miles ahead.
And P3000 has the licenses you need, and is tested more rigorously for reliability.