W1zzard over at techPowerUp! reckons that AMD may have sent out bad HD 4830 reference review samples to members of the press. The issue was discovered when shader counting support for the new HD 4830 was added to GPU-Z and partner cards showed a count of 640 stream processors (or shaders) and the reference sample from AMD which was sent to techPowerUp! and another site only had 560 shaders.
Of course, this could have really screwed up the HD 4830 launch for AMD since the missing shaders account for about a 10% drop in performance according to the testing performed.
As you can see, the FPS changes linearly with the number of shaders in the GPU. Please note that this is data from some real empiric testing. All those ten tests were performed with a real running HD 4850 (running at HD 4830 clocks for better comparability), without reboot in between, shader count was changed on the fly.
If you look back up at the first graph, you can easily see that the AMD 4830 result matches the result I got with 560 shaders and the PowerColor 4830 matches 640 shaders.
This test confirms that 3DMark06 Perlin is a valid test to determine the relative shader performance of a RV770 card. It also confirms that the register contents reflect the active number of shaders and that AMD's sample has the shading power of a HD 4850 with 560 shaders and that the PowerColor card is equally as fast as a HD 4850 with 640 shaders.
You can read the rest of the story over here.