FPGAs target edge AI computing designs
Efinix Trion Titanium FPGAs are fabricated on a 16-nm process node and incorporate the company’s Quantum fabric for compute acceleration, machine learning, and deep learning. Combined with Efinix RISC-V SoCs, Titanium FPGAs form the compute core and adaptive hardware acceleration for complete embedded system-in-package (SiP) designs.
Leveraging the Quantum fabric’s enhanced exchangeable logic and routing (XLR) cell and 2X efficiency improvement, along with highly configurable embedded memory blocks and dedicated high-speed DSP blocks, Titanium FPGAs pack plenty of processing power into a die size that is just a quarter of the area of previous-generation Trion devices. The low power consumption of the 16-nm node enables Titanium devices to consume a third of the power of Trion FPGAs and overcome all of the thermal issues associated with highly-integrated applications.
The Titanium family comprises FPGAs with 25,000 to 500,000 logic elements housed in BGA packages.
>> This article was originally published on our sister site, EDN.