Does overclocking the GPU help with making the rendering process faster, make the rasterizer drawing frames faster or it just works for GPU computing for example make the OpenGL and DirectX11 compute shaders work faster and CUDA and OpenCL parallel computing platforms?

There is no real technical difference between gfx ‘shaders’ and GPGPU 'compute', both use the same cores and will equally benefit from higher frequency.
Dedicated gfx HW is mainly the non programmable stuff like ROPs or TMUs, and now RT, and those units usually also depend on the same core clock afaik.
Probably the more interesting related question is ‘What gives more benefit - increasing core clock or memory clock?’, which should depend on the application.
But even more important: How much does overclocking decrease lifetime? We know mining GPUs die rather quickly because 24/7 operation, so i guess overclocking has a similar effect.
Considering HW becomes more and more expensive, i won't do any overclocking anymore.