Edit Values
Loihi 2
General Info
Designer	Intel
Manufacturer	Intel
Market	Artificial Intelligence
Introduction	September 30, 2021 (announced) September 30, 2021 (launched)
Shop	Amazon
Neuromorphic Specs
Neurons	1,000,000
Synapses	120,000,000
Microarchitecture
Process	Intel 4
Transistors	2,300,000,000
Technology	CMOS
Die	31 mm²
Multiprocessing
Max SMP	16,384-Way (Multiprocessor)
Vcore	0.50 V-1.25 V
Succession
Loihi

Loihi 2 (pronounced low-ee-hee 2) is the successor to Loihi, a second-generation neuromorphic research test chip designed by Intel Labs that was introduced in late 2021. The chip uses an asynchronous spiking neural network (SNN) to implement adaptive self-modifying event-driven fine-grained parallel computations used to implement learning and inference with high efficiency. The chip is a 128-neuromorphic cores many-core IC fabricated on Intel 4 process and features a unique programmable microcode learning engine for on-chip SNN training.

Overview

Introduced in September 2021, Loihi 2 is the successor to Loihi, a neuromorphic research test chip. Like its predecessor, Loihi, Loihi 2 consists of an asynchronous spiking neural network (SNN) meaning instead of manipulating signals, the chip sends spikes along activate synapses. Connections are asynchronous and highly timed-based. Neuromorphic cores containing many neurons are interlinked and receive spikes from elsewhere in the network. When received spikes accumulate for a certain period of time and reach a set threshold, the core will fire off its own spikes to its connected neurons. Preceding spikes reinforce each other and the neuron connections while spikes that follow will inhibit the connection, declining the connectivity until all activities are halted.

Loihi 2 is fabricated on pre-production Intel 4 process and has a total of 1,000,000 artificial neurons and 120 million synapses. In addition to the 128 neuromorphic cores, there are 6 managing Lakemont cores.

Key changes from Loihi

• Intel 14 nm -> Intel 4 process
• 1.10x transistors
• 0.9x synapse (120,000,000, down from 130,000,000)
• 7.7x neurons (1,000,000, up from 130,000)
• 2x microprocessor count (6, up from 3)
• 3D mesh scaling (from 2D)
• 32-bit integer payload (from binary)
• Rich programmable neuron cores
• 3-factor learning
• Faster circuits
• new interfaces
  • gigabit ethernet

Die

• Intel 4 process
• 2,300,000,000 transistors
  • 128 neuromorphic cores + 6 x86 cores
• 31 mm² die size
  • Core area 0.21 mm²

Bibliography

• Some information was obtained directly from Intel

See also

• Loihi
• TrueNorth