模拟退火算法在低功耗BIST中的应用

Simulated annealing algorithm applied in low power BIST scheme

提出了应用模拟退火算法在一定长度的测试矢量集中寻找有效测试矢量的近似最优分组 ,在尽量减少面积开销的同时减少有效测试矢量的个数 ,并且通过置入种子的方法使LFSR产生近似最优分组的矢量 ,因此在保障故障覆盖率的前提下达到了降低测试功耗的目的 .实验表明 ,采用此方法可降低测试功耗 70 %以上 ,而故障覆盖率维持不变 .此外 ,由于减少了测试矢量 ,测试时间也大为缩短 ,在实时系统中 。

An approach to approximately optimal group test vectors in a certain length of test patterns is proposed to decrease the number of test vectors based on simulated annealing algorithm. By the scheme of reseeding, this approach makes linear feedback shift register (LFSR) generate optimized groups of vectors, so as to reduce the power consumption without any loss of fault coverage. The experiment result shows that more than 70% power consumption can be reduced while keeping the fault coverage invariable. In addition, the test time is greatly shortened with decreased number of test vectors, which is important in real time device.