x86 is, with all due respect, a horribly outdated architecture. It first arrived in 1978 and even if it has been upgraded twice to support longer words (first 32-bit and then 64-bit) there have been architectures that were superior, but didn’t manage to come establish themselves. Most of all because of the strong establishment of x86. Researchers at University of Texas have presented a processor they call TRIPS (Tera-op, Reliable, Intelligently adaptive Processing System). The foundation of this processor has been under development for seven years now by Professors Stephen Keckler, Doug Burger and Kathryn McKinley and belongs to a new kind of processor architectures called EDGE (Explicit Data Graph Execution).
The difference between this one and other conventional architectures is that this one can handle large blocks of data instead of a single instruction. Today’s retail processors have been upgraded with more processor cores to improve the parallel work capacity, but this has some pretty obvious limitations, but at the same time they remain compatible with current standards.
But instead they increase the amount of work for software developers, which have to rewrite their applications to make use of the new cores. EDGE is simply a possible solution when x86 can no longer be broadened with more cores. Each TRIPS chip contains two cores that can handle 16 operations per cycle or a whopping 1,024 instructions at the same time. Today’s processors can handle up to four operations per cycle. The researchers are hoping to increase the number of operations significantly soon.
Even if we, alas, doubt that TRIPS and EDGE will become a real alternative to consumers they may very well become a serious threat to processors like SUN’s Niagara and similar processors.