30 fps does double the cycles available to render a single frame. If the engine is simply brute force rendering everything again for each frame, this basically "doubles the GPU performance".
However there are many techniques available that reuse old data and amortize the rendering cost to several frames. The smaller variation between frames, the smaller the cost to render a single frame. In a 60 fps game, the camera and animation moves at half the speed (difference from a frame to another). For example CSM scrolling (
https://d3cw3dd2w32x2b.cloudfront.net/wp-content/uploads/2012/08/CSM-Scrolling.pdf) provides a pretty much constant performance hit for sunlight shadow map rending, independent of the frame rate. In a 60 fps game, the camera moves at half the speed (per frame), so you have to update roughly half the shadow map data per frame compared to a 30 fps game. A similar efficiency gain for complex material combining (including decal rendering) can be achieved with virtual texturing. The visible scene rendering samples a single texel from the virtual texture cache (this is very fast). All the complex material blending is rendered to the virtual texture cache, and reused for as many frames as the surface is visible. In a 60 fps game, each generated surface pixel (in virtual texture cache) is visible for twice as many frames compared to a 30 fps game. Again the cost is independent of the frame rate.
Compute shaders allow us to run more complex algorithms on the GPU. As soon as current generation consoles are no longer the main development platform, we will start seeing some clever rendering pipelines that do not have to use that much brute force to generate a similar end result. 60 fps will eventually be a better choice. The performance impact will be minimized, while the positive impact on game play quality will remain.
Nobody targets 30fps. 30fps is a side effect of poor optimization.
This is not true. Many console games target 30 fps from the beginning of the project. The target frame rate is decided at early stage of a console game project (locked 30 fps, locked 60 fps, or variable). The selected frame rate affects many technical decisions, project schedule and tasks (it takes time to create new effects, higher frame rate allows less special effects). The frame rate is monitored throughout the project. No matter what frame rate was chosen, a huge deal of programmer time is spent in optimizing the bottlenecks. If you are running at 60 fps, you have to micro-optimize every single thing you are running on the GPU. Every GPU instruction counts. If you are running at 30 fps, you have more effects to optimize, more draw calls, more geometry. Optimization is as important as it is for the 60 fps game, but the focus is more on the fluctuating parts of the rendering process (for example object and shadow map rendering), as the relative cost of the constant cost passes (for example post effects) is smaller.
