Im discovering the limits of outdoor longer range architectural ue4 has. however my issues appears to be gpu based. ram and cpu are only around 30 - 40% but the frame rate drops to about 5 fps when the camera turns to the long view. i have x2 980gtx (though i have read that ue4 will only be using a single card). everything has been converted to static mesh from bsp and unfortunatly i have been forced to use all realtime lighting (1st project and ****ed up something somewhere that crashes editor towards the end of the light compile usually about 5 or 6 hours into it).
My question is:
Will i get more power in ue4 from investing in a quadro card over a titanx card and if so how much more power? links to performance stats etc… would be great as well.
Or will a Tesla card serve my requirements better?
AFAIK, Teslas are not for rendering image, but rather for GPGPU computing.
Quadro is not a good idea for game products, we have a K6000 and it has worse performance than 780Ti when running UE4.
As for Titan X, I’ve seen benchmarks showing 980Ti outperforming it in some cases, others were very close to almost no difference between the two (3-5%). Considering the price I would go for 980Ti, but don’t expect big boost in performance over 980 in either case.
I would rather look into how you can optimize the scene better (number and range of dynamic lights, objects casting shadows, LOD, culling, check for shader complexity, transparency overlays, etc.). If you press Control + Shift + , you should see a window with detailed breakdown of what’s costing you how much rendering time.
my sincere thanks. your comment has let me understand and learn more and lead me to what is looking like a workable solution.