Use Of Memory Bandwidth In GPU

Solution
Only an engineer can.
Such specs are used for marketing more than anything else.
It is fruitless to look at specs to make graphics card decisions.
The engineers select different component and specs to achieve a desired price/performance product solution.
Such specs are different between different gpu architectures and vendors; you can not validly compare them.

The best you can do is use a generalized hierarchy chart like tom's:
http://www.tomshardware.com/reviews/gaming-graphics-card-review,3107-7.html

You will find that you get comparable performance from any modern graphics cards at comparable price points.
Only an engineer can.
Such specs are used for marketing more than anything else.
It is fruitless to look at specs to make graphics card decisions.
The engineers select different component and specs to achieve a desired price/performance product solution.
Such specs are different between different gpu architectures and vendors; you can not validly compare them.

The best you can do is use a generalized hierarchy chart like tom's:
http://www.tomshardware.com/reviews/gaming-graphics-card-review,3107-7.html

You will find that you get comparable performance from any modern graphics cards at comparable price points.
 
Solution
GPUs need memory bandwidth to pull texture, geometry and other data from GPU memory into the GPU so it can render pixels and store the results back into memory. The more bandwidth the GPU has, the quicker those read and write operations can be completed and the more pixels the GPU can render per second, assuming the GPU has sufficient processing power to benefit from additional bandwidth and that all the necessary data fits in the GPU's memory.