Programs/Benchmarks in OpenCL and CUDA

I am looking for the same program or benchmark expressed in both CUDA and OpenCL. My goal is to take these example programs an try to characterize the performance differences.

Obviously, the NVIDIA SDKs offer quite a few examples that work for me, but I am looking for more. Any ideas?

We have been playing around benchmarking with this OpenCL program.

I also would love to see a CUDA port, just to compare the performance difference.

There is also this benchmark, but we are waiting on a revision.

Sandra Light is out, but Nvidia has issues with that app too.…20&start=20

We need more benchmarks that weren’t made on ATI first, and were better optimized for Nvidia from the beginning.

I wish Nvidia would put one out…