Table 1.
Comparison between area, memory, and the technology node used in a few neuromorphic chips.
| Architecture | mm 2 | Memory (Mb) | Technology | Neurons(Physical) |
|---|---|---|---|---|
| /Core | /Core | /core | ||
| ODIN (Frenkel et al., 2018) | 0.086 | 0.28 | STM 28 nm | 256(1) |
| TrueNorth (Akopyan et al., 2015) | 0.10 | 0.10 | Samsung 28 nm | 256(1) |
| NeuronFlow (Moreira et al., 2020) | 0.1 | 0.12 | TSMC 28 nm | 1024(1) |
| Loihi2 (Davis, 2021) | 0.21 | 1.5 | Intel 4 nm | Flex(1) |
| Loihi (Davies et al., 2018) | 0.41 | 2.0 | Intel 14 nm | 1024(1) |
| ReckOn (Frenkel and Indiveri, 2022) | 0.45 | 1.1 | FDSOI 28 nm | 256(16) |
| μBrain (Stuijt et al., 2021) | 1.42 | 0.15 | TSMC 40 nm | 336(336) |
| SpiNNaker (Furber et al., 2014) | 5.6 | 0.12 | 130 nm | Flex(1) |
| SpiNNaker2 (Höppner et al., 2021) | 1.09 | 1.0 | FDX 22 nm | Flex (64) |
| Tianjic (Deng et al., 2020) | 0.092 | 0.17 | UMC 28 nm | 256(16) |
| SENECA | 0.47 | 2.3 | FDX 22 nm | Flex(8) |
The amount of memory can be used as an indication of the number of neurons and synapses per core.