Originally Posted by Strat79
For AMD, the Shader's, or "Stream Processors", are the main FPS pushers so to speak. Their clock speed is tied to the Core clock speed, so whatever the core speed is, the shader speed is. That is why you don't see it on AMD cards. Cuda Cores are the same thing as Shaders, Cuda cores are the FPS pushers for Nvidia cards. They are architecturally different, that is the reason there is a big difference in the number of Shaders for each company. They work differently and cannot be compared by clock speeds or numbers between the two companies. On Nvidia cards(for the most part anyway), the shader speed is not tied to the core clock, so you can change them independently.
Basically they both have the same stuff in them, just named differently for marketing reasons. Both have shaders as the main powerhouse. AMD calls theirs "Stream Processors" and Nvidia calls theirs "Cuda Cores".
Based on that, your card selection should be based primarily on benchmarks (Synthetic/real world) between your preferred cards and not just based on what certain details are because both companies use very different architectures.
Nvidia stresses CUDA cores because they(i.e. shaders) are used in encoding but quite a few and do accelerate the process a lot. They do that to get the attention of people who seek the graphics card for that purpose.
AMD at times focuses on "Stream Processors" when listing but again, they are both shaders. (I don't know how present day video encoding/transcoding performance fares with amd gpus so I'll make no comment on that).
Apps (example, flash/MPC) which allow hardware acceleration work excellently with both manufacturers anyway.