Have we tried increasing the block size limit on a test coin yet?  We should make sure the functions operate as expected under high and bursts of saturation workloads.