Semiconductor & Computer Engineering
Template:mpu ETANN (Electronically Trainable Analog Neural Network) was the first commercial neural processor, introduced by Intel in 1992. Implemented on a 1.0 µm process, this chip incorporated 64 analog neurons and 10,240 analog synapses.
Overview
The ETANN was originally announced at the 1989 International Joint Conference on Neural Networks (IJCNN). The chip was implemented using an analog nonvolatile floating gate technology on Intel's CHMOS-III 1µm nonvolatile memory technology. The chip integrates a total of 64 analog neurons and 1024 analog nonvolatile synapses. The network calculated the dot product between the 64x64 nonvolatile EEPROM analog synaptic weight array and a 64-element analog input vector. The chip was reported the calculations to exceed 1.3 billion interconnections per second.
Die
- CHMOS-III 1µm nonvolatile memory technology
- 93.15 mm² die size
- 8.1 mm x 11.5 mm
Floor plan:
| core voltage | 5 V (50 dV, 500 cV, 5,000 mV) + |
| designer | Intel + |
| die area | 93.15 mm² (0.144 in², 0.932 cm², 93,150,000 µm²) + |
| die length | 8.1 mm (0.81 cm, 0.319 in, 8,100 µm) + |
| die width | 11.5 mm (1.15 cm, 0.453 in, 11,500 µm) + |
| first announced | 1989 + |
| first launched | 1989 + |
| full page name | intel/etann + |
| instance of | microprocessor + |
| ldate | 1989 + |
| manufacturer | Intel + |
| market segment | Artificial Intelligence + |
| model number | 80170NX + |
| name | ETANN + |
| part number | 80170NX + |
| power dissipation | 2 W (2,000 mW, 0.00268 hp, 0.002 kW) + |
| power dissipation (average) | 1.5 W (1,500 mW, 0.00201 hp, 0.0015 kW) + |
| process | 1,000 nm (1 μm, 0.001 mm) + |