UK memory technology company sureCore has developed in-memory compute designs for edge AI chips.
The company develops low power SRAM technology that can be embedded into system on chip designs. It is extending this by adding embedding arithmetic blocks deep within the memory array in a way that is compatible with its existing silicon-proven, low-power memory design.
This ‘CompuRAM’ mean data can be processed within the memory block rather than having to move it around a chip, saving energy on data transfers. With billions of operations for machine learning, this can make a significant difference to the overall power consumption of a chip at the edge of the IoT network.
“Our knowledge of memory technology means that we have been able to create a solution for the next technology demand of integrating arithmetic operations within the memory,” said Tony Stansfield, CTO of sureCore. “Cutting power consumption is what we do as we have proven with our existing technologies, such as our EverOn and PowerMiser SRAM families that enable near threshold operation and 50 percent dynamic power cuts respectively. Our solutions are all designed to making it possible to create products for the next generations of ultra-low power applications that could not exist without their power reducing techniques.”
- Samsung uses MRAM for AI in-memory computing
- NeuroBlade raises $83m for compute in memory chip
- ISSCC 2020 sees 3nm process, compute-in-memory
- Processing-in-memory advance uses neural networks
- AI satellite IoT module tackles chip shortage
Other articles on eeNews Europe
- EU takes aim at standards
- US looks to quantum-ready workers
- Ultra-low-power supercomputing chip targets AI, HPC and metaverse
- Framework raises $18m for sustainable laptop