We propose 2D Piezoelectric FET (PeFET) based compute-enabled non-volatile memory for ternary deep neural networks (DNNs). PeFETs hinge on ferroelectricity for bit storage and piezoelectricity for bit sensing, exhibiting inherently amenable features for computation-in-memory of dot products of weights and inputs in the signed ternary regime. PeFETs consist of a material with ferroelectric and piezoelectric properties coupled with Transition Metal Dichalcogenide channel. We utilize (a) ferroelectricity to store binary bits (0/1) in the form of polarization (-P/+P) and (b) polarization dependent piezoelectricity to read the stored state by means of strain-induced bandgap change in Transition Metal Dichalcogenide channel. The unique read mechanism of PeFETs enables us to expand the traditional association of +P (-P) with low (high) resistance states to their dual high (low) resistance depending on read voltage. Specifically, we demonstrate that +P (-P) stored in PeFETs can be dynamically configured in (a) a low (high) resistance state for positive read voltages and (b) their dual high (low) resistance states for negative read voltages, without afflicting a read disturb. Such a feature, which we name as Polarization Preserved Piezoelectric Effect Reversal with Dual Voltage Polarity (PiER), is unique to PeFETs and has not been shown in hitherto explored memories. We leverage PiER to propose a Strain-enabled Ternary Precision Computation-in-Memory (STeP-CiM) cell with capabilities of computing the scalar product of the stored weight and input, both of which are represented with signed ternary precision. Further, using multi word-line assertion of STeP-CiM cells, we achieve massively parallel computation of dot products of signed ternary inputs and weights. Our array level analysis shows 91% lower delay and improvements of 15% and 91% in energy for in-memory multiply-and-accumulate operations compared to near-memory design approaches based on 2D FET based SRAM and PeFET respectively. We also analyze the system-level implications of STeP-CiM by deploying it in a ternary DNN accelerator. STeP-CiM exhibits 6.11× - 8.91× average improvement in performance and 3.2× average improvement in energy over SRAM based near-memory design. We also compare STeP-CiM to near-memory design based on PeFETs showing 5.67× - 6.13× average performance improvement and 6.07× average energy savings.

Gravitational lensing studies of the Bullet Cluster suggested convincingly in favour of the existence of dark matter. However, it was performed without the knowledge of the original orientation of each galaxy before gravitational lensing. A potential improvement to this issue lies in the measurement of the original orientation from the polarization direction of radio waves emitted from each galaxy. In this context, Francfort et al. derived a formula that can utilize the information about the original orientation of each galaxy to obtain what is called shear. However, we demonstrate that shear in their formula should be replaced by reduced shear when the change in sizes of images of galaxies is taken into account. As the previous gravitational lensing analysis of the Bullet Cluster used reduced shear, we suggest applying our improved formula directly for the reanalysis once we obtain the polarization direction of radio waves. In particular, we show that our new formula can yield a more accurate analysis than the previous one, if the polarization direction can be measured more precisely than 10◦.