Unfortunately @RCAFBrat you appear to be confusing GPU PCIe bandwidth with GPU vRAM bandwidth.
The PCIe bandwidth is how fast data can be fed from the CPU/RAM to the GPU via the PCIe connection lanes.
The 256/512bit finalbarrage is asking about is how fast the GPU (CUDA for example) cores can communicate with the GPU vRAM.
There are two main factors in what affects a GPU’s memory bandwidth, the first is its clockspeed (MHz). The higher the clockspeed the more memory transfers a GPU can do per second.
The ‘bit’ width eg 256bit is how wide the data transfer between GPU cores and the vRAM is, or how much data can be sent in a single transfer.
Having a higher # of ‘bits’ in the bandwidth isn’t necessarily useful if your GPU doesn’t have the capability to process that data. That’s why greater core counts go hand in hand with faster, and wider memory connections, to keep the GPU cores fed.
Judging by the usage stats I see in GPU-Z when rendering; Example GPU: 100% Memory: 63% I would say the memory is fast enough to keep the GPU working as best it can.