% IMPORTANT: The following is UTF-8 encoded.  This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.

@ARTICLE{Freye:912510,
      author       = {Freye, Florian and Lou, Jie and Bengel, Christopher and
                      Menzel, Stephan and Wiefels, Stefan and Gemmeke, Tobias},
      title        = {{M}emristive {D}evices for {T}ime {D}omain
                      {C}ompute-in-{M}emory},
      journal      = {IEEE journal on exploratory solid-state computational
                      devices and circuits},
      volume       = {8},
      number       = {2},
      issn         = {2329-9231},
      address      = {New York, NY},
      publisher    = {IEEE},
      reportid     = {FZJ-2022-05683},
      pages        = {119 - 127},
      year         = {2022},
      abstract     = {Analog compute schemes and compute-in-memory (CIM) have
                      emerged in an effort to reduce the increasing power hunger
                      of convolutional neural networks (CNNs), which exceeds the
                      constraints of edge devices. Memristive device types are a
                      relatively new offering with interesting opportunities for
                      unexplored circuit concepts. In this work, the use of
                      memristive devices in cascaded time-domain CIM (TDCIM) is
                      introduced with the primary goal of reducing the size of
                      fully unrolled architectures. The different effects
                      influencing the determinism in memristive devices are
                      outlined together with reliability concerns. Architectures
                      for binary as well as multibit multiply and accumulate (MAC)
                      cells are presented and evaluated. As more involved circuits
                      offer more accurate compute result, a tradeoff between
                      design effort and accuracy comes into the picture. To
                      further evaluate this tradeoff, the impact of variations on
                      overall compute accuracy is discussed. The presented cells
                      reach an energy/OP of 0.23 fJ at a size of 1.2 μm2 for
                      binary and 6.04 fJ at 3.2 μm2 for 4×4 bit MAC
                      operations.},
      cin          = {PGI-7 / JARA-FIT},
      ddc          = {530},
      cid          = {I:(DE-Juel1)PGI-7-20110106 / $I:(DE-82)080009_20140620$},
      pnm          = {5233 - Memristive Materials and Devices (POF4-523) /
                      BMBF-16ME0399 - Verbundprojekt: Neuro-inspirierte
                      Technologien der künstlichen Intelligenz für die
                      Elektronik der Zukunft - NEUROTEC II - (BMBF-16ME0399) /
                      BMBF-16ME0398K - Verbundprojekt: Neuro-inspirierte
                      Technologien der künstlichen Intelligenz für die
                      Elektronik der Zukunft - NEUROTEC II - (BMBF-16ME0398K)},
      pid          = {G:(DE-HGF)POF4-5233 / G:(DE-82)BMBF-16ME0399 /
                      G:(DE-82)BMBF-16ME0398K},
      typ          = {PUB:(DE-HGF)16},
      UT           = {WOS:000915312400008},
      doi          = {10.1109/JXCDC.2022.3217098},
      url          = {https://juser.fz-juelich.de/record/912510},
}