That is not a sneak performance peak. Its an article that states how fast it runs, then notates that its running un-optimized. It mentions that in this bench it is slightly slower than the sandy bridge, but it doesn't state anywhere whether or not the sandy is clocked the same as the ivy.
Don't expect any loss of performance, but don't expect any real gains either. Also, wait until real benches come out that show the true performance.
The benchmark they ran was only testing memory and cache performance, it tells you absolutely nothing about the computational performance of the chip, just that it might have slightly worse memory performance which really doesn't matter since there are very very few scenarios where your memory bandwidth even affects your performance today.