Soft Error Masking Method Based on Approximate Logic Circuit
An approximate logic, soft error technology, applied in the field of soft error shielding, can solve problems such as logic errors, precision tolerance errors, timing errors, etc., to achieve the effect of improving scalability, ensuring scalability, and cost-effectiveness
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0035] Embodiment one: see figure 1 As shown, a soft error masking method based on an approximate logic circuit includes the following steps:
[0036] 1) Determine the space where the original circuit needs logic error shielding and the approximate one-way space of the original circuit, such as figure 2 Shown, where ∑ err It is the area that needs to be shielded in the original circuit. The logic function of the 1-approximation circuit reaches 81% of the original circuit, and the number of logic gates consumed is 44.4% of the original circuit. The function of the 0-approximation circuit is 81% of the original circuit. The quantity is 27.9% of the original circuit, it can be seen that the approximate circuit has a higher cost performance;
[0037] 2) Logically synthesize the experimental circuit through the open source abc tool of the Berkeley Research Center, and use the obtained circuit netlist file in BLIF format as the original circuit form;
[0038] 3) Carry out the lo...
Embodiment 2
[0050] Embodiment two: see Figures 5 to 7 Shown is the reference circuit cmb in the LGSynth93 library, which has 16-bit input and 4-bit output. The original circuit contains 51 logic gates, while the 1-approximation circuit contains 28 logic gates. The 1-approximation circuit saves 59.7% of logic gates and 46.4% of power consumption while achieving 95.8 logic functions. Figure 7 In , the simulation time is set to 1 us, and the input combinations are 0, 26, 65535, 9876, .... The symbol "_org" represents the waveform of the original circuit, "_appro" represents the waveform of the approximate circuit, "_faulttolerance" represents the output waveform of the fault-tolerant system combined with the original circuit and the 1-approximate circuit, "q", "r", "s" and " t” indicates a 4-bit output terminal. Depend on Figure 7 It can be seen that there are two logic errors from 1 to 0. When the 1->0 transition occurs at the output terminal of s, the shielding system composed of the...
Embodiment 3
[0051] Embodiment three: see Figures 8 to 10 Shown is the reference circuit x2 in the LGSynth93 library, which has 10-bit input and 7-bit output, and its original circuit complexity is slightly higher than cmb. The original circuit contains 85 logic gates, while the 1-approximation circuit contains 33 logic gates. The 1-approximation circuit saves 61.7% of the number of logic gates and 23.2% of power consumption while achieving 94.3% of logic functions. Figure 10 Among them, the input combination is a set of random values, where the symbol "_org" represents the waveform of the original circuit, "_appro" represents the waveform of the approximate circuit, "_faulttolerance" represents the output waveform of the fault-tolerant system combined with the original circuit and the 1-approximate circuit, " l", "k", "m", "n", "o", "p" and "q" represent 7-bit output terminals. The glitches of 1->0 jumps are observed in all the output terminals of the original circuit of x2. At the k a...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


