PAALM: Power Density Aware Approximate Logarithmic Multiplier Design
PAALM: Power Density Aware Approximate Logarithmic Multiplier Design
复制标题
DOI：
10.1145/3566097.3567884
复制
发表时间：
2023-01
期刊：
影响因子：
--
通讯作者：
Shuyuan Yu;S. Tan
中科院分区：
文献类型：
--
作者：
Shuyuan Yu;S. Tan
近似的硬件设计可以显着降低功耗或能耗。然而，最近的一项研究表明，由于功率密度的增加，近似设计可能会导致不必要的高温和相关的可靠性问题。在这项工作中，我们试图通过首次提出一种新颖的功率密度感知近似对数乘法器（称为 PAALM）设计来缓解这一重要问题。由于其严格的数学基础，新的乘法器设计基于近似对数乘法器 (ALM) 框架。这个想法是基于等效数学公式重新设计现有 ALM 设计的高计算开关活动，以便可以在不损失精度的情况下降低功率密度，同时以一些面积开销为代价。我们的结果表明，与定点乘法器基线相比，所提出的 PAALM 设计可以在 8/16 位精度下分别提高 11.5%/5.7% 的功率密度和 31.6%/70.8% 的面积。并且还实现了极低的误差偏差：8/16 位精度分别为 -0.17/0.08。除此之外，我们进一步在卷积神经网络（CNN）中实现 P AALM 设计，并在 CIFAR10 数据集上进行测试。结果表明，通过误差补偿，PAALM 可以达到与定点乘法器基线相同的推理精度。我们还在离散余弦变换 (DCT) 应用中评估 PAALM。结果表明，通过误差补偿，PAALM 与 ALM 设计相比图像质量平均提高 8.6dB。
Approximate hardware designs can lead to significant power or energy reduction. However, a recent study showed that approximated designs might lead to unwanted higher temperature and related reliability issues due to the increased power density. In this work, we try to mitigate this important problem by proposing a novel power density aware approximate logarithmic multi-plier (called PAALM) design for the first time. The new multiplier design is based on the approximate logarithmic multiplier (ALM) framework due to its rigorous mathematics based foundation. The idea is to re-design the high computing switch activities of existing ALM designs based on equivalent mathematical formula so that the power density can be reduced at no accuracy loss while at costs of some area overheads. Our results show that the proposed PAALM design can improve 11.5%/5.7% of power density and 31.6%/70.8% of area with 8/16-bit precision when compared with the fixed-point multiplier baseline, respectively. And also achieves extremely low error bias: -0.17/0.08 for 8/16-bit precision, respectively. On top of this, we further implement the P AALM design in a Convolutional Neural Network (CNN) and test it on CIFAR10 dataset. The results show that with error compensation, PAALM can achieve the same inference accuracy as the fixed-point multiplier baseline. We also evaluate the PAALM in a discrete cosine transformation (DCT) application. The results show that with error compensation, PAALM can improve the image quality of 8.6dB in average when compared to the ALM design.