1gh v1.2 - windows 8 64 bit. sapphire 7970 ghz edition 3gb, 1040 core 1500 ram clock speed. ~1120 CPM
Just for fun:
cudapts (linux) v2014-01-12 GTX690, stock clock settings, 1780 c/m.
Do you mind to try 1gh miner on this one? That would give a general idea how much OpenCL is inferior to CUDA.
Unfortunately, that machine also has one of my coin wallets on it, and I don't run untrusted binaries on it. 
(Mind you, I'm not saying there's anything wrong with the 1gh miner - but there's been a lot of malware floating around.)
But - probably the best case thus far is the EC2 results showing 480 c/m vs about 780. One might expect the same per-core results to extend to the 690, so maybe it would get about 1150? (total SWAG)
I totally understand your concern about running unknown binaries, yet, maybe there is a way to arrange it securely without much hassle? For example, in VM with mapped PCI device if you are already using such setup for development?
480cpm must be something else, 1gh miner yields 687cpm on g2.2xlarge and I believe there is no more room for optimization (but it is possible that different Nvidia OpenCL compiler versions may produce different results).
