Obviously running a x16 card at x16 is the best choice. Motherboards with multiple PCIe slots sometimes switch an x16 slot to run at x8 when there are multiple GPU's installed.

I am just wondering if there is any testing specific to GPUGrid (or other GPU crunching projects) regarding the impact of x16 vs. x8. Does crunching exploit the need for x16 (beyond x8) and if so, what is that average performance hit? Also, how does this change between different cards (GTX5xx vs. 4xx or 2xx series cards)?

I am running at x16, but just wondering for my own curiosity.

There is no difference between PCIE 2.0 x8 and x16 for most current cards. Not entirely sure about the new HD 7970 and GTX 680, but nothing before that really needs x16, except for the dual-GPU cards.

There is a performance impact of PCIe x16 vs x8, but it's dependent on many other factors:
The faster the GPU and the higher the CPU utilization of the running task, and the older the CPU is (has FSB), the higher the performance penalty will be of a GPU running at x8 instead of x16.
See the Credit per € / $ thread, especially messages 22056, 22058, 24141, and message 21451 in a different thread.

Excellent, thanks Zoltan. I thought I had read on this topic before, but didn't find those threads with the search for "x16" or "x8".