Table 1 Targeted applications based on power and data distribution requirement
From: Leveraging stochasticity in memristive synapses for efficient and reliable neuromorphic systems
READ voltage (V) | READ Power (μW) | Avg. Power (μW) | Avg. Energy (pJ) | Data ‘0’ (% of noise) | Data ‘1’ (% of noise) | Targeted Application | Comment | |
|---|---|---|---|---|---|---|---|---|
3 kΩ | 30 kΩ | |||||||
0.55 | 20.71 | 20.19 | 20.45 | 20.45 | ~100% | ~93.48% | Biased coin flip | Better data distribution with bias and high power |
0.6 | 21.63 | 19.93 | 20.78 | 20.78 | ~100% | ~48.68% | Biased coin flip | Better data distribution with bias and high power |
0.65 | 23.74 | 18.40 | 21.07 | 21.07 | ~0.12% | ~0.52% | Reliable SNN | Almost perfect READ operation |
0.7 | 27.52 | 13.88 | 20.7 | 20.7 | ~0.02% | ~0.52% | Reliable SNN | Almost perfect READ operation |
0.75 | 31.58 | 10.14 | 20.86 | 20.86 | ~100% | ~18% | Biased coin flip | Poor data distribution with bias and low power |
0.8 | 8.71 | 4.83 | 6.77 | 6.77 | ~100% | ~68.46% | Biased coin flip | Poor data distribution with bias and low power |
0.85 | 2.54 | 2.18 | 2.36 | 2.36 | ~100% | ~98.22% | Biased coin flip | Poor data distribution with bias and low power |
