ARM and GLOBALFOUNDRIES have inked a multi-year deal to jointly develop improved system-on-chip solutions to design ARM processors for GLOBALFOUNDRIES’ 20-nm and FinFET process technologies.
As per the deal, the long-standing partnership will now also incorporate graphics processors, which are turning out to be a key element in mobile devices. ARM will deliver a comprehensive platform of ARM Artisan Physical IP, which includes POP IP solutions, memory compilers and standard cell libraries. The resulting solutions will provide a new level of power-efficiency and system performance for tablets, smartphones and ultrathin notebooks.
Through their long-standing partnership, ARM and GLOBALFOUNDRIES have jointly optimized ARM Cortex-A series processors, which include numerous demonstrations of power efficiency and performance advantages on 28 nm and a 20 nm test-chip deployment that is presently running through GLOBALFOUNDRIES fab situated in Malta, New York. This deal widens the earlier efforts by accelerating production IP platforms, which will allow for customer designs on 20 nm and support quick shifting to three-dimensional FinFET transistor technology. This joint development will quickly deliver system-on-chip solutions for clients utilizing advanced ARM CPUs and GPUs in portable devices.
GLOBALFOUNDRIES intends to develop improved benchmark analysis and implementations for advanced, energy-efficient ARM Mali graphic processor and ARM Cortex processor technologies, speeding up customers’ own system-on-chip designs utilizing the respective technologies. The full platform of ARM Artisan Physical IP on GLOBALFOUNDRIES’ 20 nm-LPM and FinFET processes as well as POP IP products offer basic building blocks for system-on-chip designers. This platform is based on the current Artisan physical IP platforms for several GLOBALFOUNDRIES’ process technologies, which include 28, 55, and 65 nm, and the Cortex-A9 POP technology for 28 nm SLP, now offered for licensing from ARM. The GLOBALFOUNDRIES 20 nm-LPM technology is a complete, cost-efficient platform, providing up to 40% improvement in performance and double the gate density of 28 nm.