We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:


References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Emerging Technologies

Title: Proposal of Analog In-Memory Computing with Magnified Tunnel Magnetoresistance Ratio and Universal STT-MRAM Cell

Abstract: In-memory computing (IMC) is an effectual solution for energy-efficient artificial intelligence applications. Analog IMC amortizes the power consumption of multiple sensing amplifiers with analog-to-digital converter (ADC), and simultaneously completes the calculation of multi-line data with high parallelism degree. Based on a universal one-transistor one-magnetic tunnel junction (MTJ) spin transfer torque magnetic RAM (STT-MRAM) cell, this paper demonstrates a novel tunneling magnetoresistance (TMR) ratio magnifying method to realize analog IMC. Previous concerns include low TMR ratio and analog calculation nonlinearity are addressed using device-circuit interaction. Peripheral circuits are minimally modified to enable in-memory matrix-vector multiplication. A current mirror with feedback structure is implemented to enhance analog computing linearity and calculation accuracy. The proposed design maximumly supports 1024 2-bit input and 1-bit weight multiply-and-accumulate (MAC) computations simultaneously. The 2-bit input is represented by the width of the input (IN) pulses, while the 1-bit weight is stored in STT-MRAM and the x7500 magnified TMR (m-TMR) ratio is obtained by latching. The proposal is simulated using 28-nm CMOS process and MTJ compact model. The integral nonlinearity is reduced by 57.6% compared with the conventional structure. 9.47-25.4 TOPS/W is realized with 2-bit input, 1-bit weight and 4-bit output convolution neural network (CNN).
Comments: 2021 submitted to IEEE Transactions on Circuits and Systems-I, regular papers
Subjects: Emerging Technologies (cs.ET)
Cite as: arXiv:2110.03937 [cs.ET]
  (or arXiv:2110.03937v1 [cs.ET] for this version)

Submission history

From: Hao Cai [view email]
[v1] Fri, 8 Oct 2021 07:13:30 GMT (3934kb)

Link back to: arXiv, form interface, contact.