Although the current in 32 nanometer manufactured Sandy Bridge platform is still the undisputed performance crown in the notebook market continues to begrudge Intel does not break and, after: a new generation of quad-core processors for desktops and notebooks, the competition will continue to keep distance.
After Sandy Bridge in the last year, a new architecture has been presented referred to in Intel-speak as tock is Ivy Bridge subsequent shrink to the next smaller manufacturing process dar because unusually large modifications especially in the graphics part of the manufacturer speaks of a Tick .
The second part is devoted to comprehensive benchmark tests the processor side, while the graphics performance of the HD 4000 GPU in a separate report have been investigated. About the later following dual-core models, we can unfortunately not reveal anything, but they are as soon as possible to drive through our test course.
The core of Ivy Bridge is fundamentally based on its predecessor, which includes familiar features such as hyperthreading and Turbo Boost 2.0. With regard to the precise question on how we want our Sandy Bridge review refer to dwell on this point to the specific features of the new generation in detail, show up at these various minor tweaks that lead to a slight improvement in performance per MHz at a rate of about 5 percent.
Thus, MOV instructions, which so far the contents of a register has been copied in a different, much more elegant now solved: With the introduced with Sandy Bridge Physical Register File is possible in conjunction with other changes, just a reference to the Register to deposit desired goal - to offload the execution units.
There were other alterations with respect to the divider unit, and AVX Shift-/Rotate-Operationen and SSE instructions, which should be accelerated by 6 additional Split-/Load-Register. Also, Intel has revised the prefetch: the allocation of the cache is dynamically adjusted and now preferred bandwidth-heavy applications, which specifically effective in exploiting multithreading scarce capacity of the fast cache.