Nvidia to compete with Xeon and Opteron processors

In the enterprise market Intel and AMD are the only two big players (VIA’s presence is debatable), with Intel wielding a significant advantage in most cases. Now Nvidia wants a piece of the action but it plans to drop the complexity and inefficiencies of the traditional x86 method deployed by AMD and Intel.

Instead Nvidia will create high performance CPUs from ARM architecture cores where the x86 processing doesn’t matter so much. Simply put, Nvidia wants to use its own ARM based processors in Tesla based supercomputers.

Project “Boulder” as it is named is an ultra-high performance system-on-chip with 8-16+ “fat” ARM-compatible cores as well as high-bandwidth interconnects and I/O. Project Boulder is due in 2014 and is aimed at serial processing tasks while Nvidia’s GPUs are in charge of the parallel processing tasks.

Project Boulder will sit alongside Project “Denver” which will combine Maxwell GPUs with Denver based ARM CPUs to create the Echelon system chip. Since both Denver and Boulder will arrive in 2014 we are assuming that this is when Nvidia is planning an all-out-assault on the processing market with its ARM-based innovations.

Source