Compressed sensing recovery using computational memory

Gallo, Manuel Le; Sebastian, Abu; Cherubini, Giovanni; Giefers, Heiner; Eleftheriou, Evangelos

doi:10.1109/iedm.2017.8268469

Cited by 20 publications

(22 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The effect of device imperfections and failures on the final reconstruction NMSE is discussed in [5]. We found that the AMP recovery can tolerate conductance variations due to programming errors (up to 20%), and up to 20% stuck-SET and stuck-RESET device failures.…”

Section: Discussionmentioning

confidence: 98%

“…The PCM chip is interfaced to a hardware platform comprising two FPGA boards and an analog frontend (AFE) board. The layout, picture, and specifications of the experimental PCM chip with integrated read/write circuitry can be found in [5].…”

Section: B Physical Implementation On Prototype Pcm Chipmentioning

confidence: 99%

“…In this paper, we propose an implementation of a CS recovery algorithm, namely Approximate Message Passing (AMP), based on memristive crossbar arrays, of which we presented a preliminary version in [5]. We experimentally investigate the impact of this memristive implementation on the performance of AMP, in particular on the reconstruction accuracy.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Compressed Sensing With Approximate Message Passing Using In-Memory Computing

Gallo

Sebastian

Cherubini

et al. 2018

IEEE Trans. Electron Devices

View full text Add to dashboard Cite

In-memory computing is a promising non-von Neumann approach where certain computational tasks are performed within resistive memory units by exploiting their physical attributes. In this paper, we propose a new method for fast and robust compressed sensing of sparse signals with approximate message passing recovery using in-memory computing. The measurement matrix for compressed sensing is encoded in the conductance states of resistive memory devices organized in a crossbar array. This way, the matrix-vector multiplications associated with both the compression and recovery tasks can be performed by the same crossbar array without intermediate data movements at potential O(1) time complexity. For a signal of size N, the proposed method achieves a potential O(N)fold recovery complexity reduction compared with a standard software approach. We show the array-level robustness of the scheme through large-scale experimental demonstrations using more than 256k phase-change memory devices.

show abstract

Section: Discussionmentioning

confidence: 98%

Section: B Physical Implementation On Prototype Pcm Chipmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Compressed Sensing With Approximate Message Passing Using In-Memory Computing

Gallo

Sebastian

Cherubini

et al. 2018

IEEE Trans. Electron Devices

View full text Add to dashboard Cite

show abstract

“…If G p or G n > G x , and |G p − G n | < 0.25 G x , where G x is a threshold, both the devices are RESET and the conductance difference is programmed to the device which had the higher conductance. For the hardware implementation, the RESET pulse shape could be determined from the PCM programming curve 2,9,20 . Here, the stochastic RESET behavior is simulated using an abrupt conductance transition to a distribution of mean 1 µS and standard deviation 0.5 µS.…”

Section: The Overall Model Description and Validationmentioning

confidence: 99%

A phase-change memory model for neuromorphic computing

Nandakumar

Gallo

Boybat

et al. 2018

Journal of Applied Physics

120

View full text Add to dashboard Cite

Phase-change memory (PCM) is an emerging non-volatile memory technology that is based on the reversible and rapid phase transition between the amorphous and crystalline phases of certain phase-change materials. The ability to alter the conductance levels in a controllable way makes PCM devices particularly well-suited for synaptic realizations in neuromorphic computing. A key attribute that enables this application is the progressive crystallization of the phase-change material and subsequent increase in device conductance by the successive application of appropriate electrical pulses. There is significant inter and intra-device randomness associated with this cumulative conductance evolution and it is essential to develop a statistical model to capture this. PCM also exhibits a temporal evolution of the conductance values (drift) which could also influence applications in neuromorphic computing. In this paper, we have developed a statistical model that describes both the cumulative conductance evolution and conductance drift. This model is based on extensive characterization work on 10,000 memory devices. Finally, the model is used to simulate supervised training of both spiking and non-spiking artificial neuronal networks.

show abstract

“…However, breaking the dichotomy of the processor and memory as separate units would vastly transform the computing landscape by allowing processing directly on the memory elements-so-called "memcomputing" or in-memory computing. Electronic implementations of such systems, able to carry out complex tasks such as scalar multiplication, bulk-bitwise operations, correlation detection, and compressed sensing recovery, are now emerging [18][19][20][21][22] . It is also suggested that such computational memory machines could solve certain nondeterministic polynomial (NP) problems in polynomial (P) time by exploiting attributes such as inherent parallelism, functional polymorphism and information overhead 23,24 .…”

Section: Introductionmentioning

confidence: 99%