The rumor mill already indicated that the upcoming Haswell processors from Intel would not offer a massive gain in performance based on the architecture when you look at it clock-for-clock opposed to the current generation Ivy bridge processors. A Russian website called OClab however was able to test an engineering sample of quad-core Intel Haswell processor clocked at 2.80GHz. The chip was tested with SuperPi 1M&32M, PiFast, wPrime 32M&1024M at at unknown platform whilst running Microsoft Windows 7 64-bit operating system. The obtained benchmark results could be very dodgy of course.
The Haswell architecture seems actually to be slightly slower than Ivy Bridge on the same clock-speed in single-threaded SuperPi 1M. Interesting is that it is faster (11 minutes 27.505 seconds) compared to its Ivi bridge (11 minutes 49.094 seconds) in the long run SuperPi 32M.
In the threaded PiFast benchmark Intel Haswell (24.01 seconds) reportedly again managed to leave Ivy Bridge (25.5 seconds) behind while operating at the same clock-speed of 2.80GHz. In wPrime 32M the Haswell chip showed better (13.86 seconds) results than the down-clocked Ivy Bridge chip (13.97 seconds); the results in longer wPrime 1024M benchmark were predictable: Haswell chip managed to finish the job in 431.171 seconds, less than a second faster than the Core i7-3770K at 2.8GHz.
Not exactly massive differences alright. Obviously we do not know the final clock frequencies and turbo modes 100% just yet, to the final performance numbers will look completely different from clock-for-clock ones.
Via Xbitlabs via oclab.ru:
Haswell versus Ivy Bridge clock for clock performance
