The emergence of persistent shared memory in recent years has created new opportunities to rethink classic algorithmic problems in the presence of crash-restart failures, and also new challenges as algorithm designers must account for both failures and concurrency. Recent research has explored algorithm design techniques based on powerful base objects that provide special features to simplify recovery from failures as an alternative to the traditional approach of programming more directly using processor instructions. Notably, Friedman, Herlihy, Marathe, and Petrank introduced detectable objects, which allow an application to resolve the outcome of operations that may have been interrupted by a crash. We propose a transformation that replaces low-level memory operations in a conventional algorithm with operations on detectable base objects that leverage persistent memory to enable recovery from crash-restart failures. The transformation is almost universal, and can be used to obtain efficient solutions to complex and previously unsolved problems.

Persistent memory (PM) has emerged as a promising technology that enables data structure algorithms to preserve their consistent state after recovering from system failures. Detectable data structures have been proposed to resolve the response of the last operation of a crashed process. Designing detectable lock-based data structures is challenging due to the need to preserve the correctness properties of the underlying locks, such as mutual exclusion and deadlock-freedom, across failures. Therefore, lock-based detectable and persistent data structures are not as common as lock-free structures. In this work, we introduce DULL: a fast, scalable and Detectable Unrolled Lock-based Linked list. To the best of our knowledge, DULL is the fastest detectable lock-based linked list and the first detectable strictly-linearizable linked list. Experimental results indicate that DULL is several-fold faster than competitors in update workloads and equally fast in read-only scenarios.

The mutual exclusion (ME) problem has been of interest to the scientific community since it was first defined by Dijkstra. Various algorithms have been developed to solve the problem, like the MCS and CLH queue-based locks. The problem was generalized into the recoverable mutual exclusion (RME) problem by Golab and Ramaraju to accommodate the possibility of process crash failures. Since then, multiple RME algorithms have been presented in the literature that vary in design and performance. Furthermore, non-uniform memory access (NUMA) architecture has become mainstream in designing modern distributed systems, stimulating the development of NUMA-aware mutex locks. None of the existing NUMA-aware mutex locks are recoverable to the best of our knowledge. In addition, none of the transformation techniques in the literature, such as flat-combining and cohort-locking, is a black-box transformation. Precisely, each of the existing transformation techniques requires specific characteristics of, and possible modifications to, the underlying NUMA-oblivious lock. In this work, we propose the Recoverable Filter (RF) lock, a black-box transformation approach that exploits memory locality to transform a NUMA-oblivious recoverable mutex lock into a NUMA-aware one. Practical experiments are conducted using two existing RME algorithms, Golab and Hendler's (GH) and Jayanti, Jayanti, and Joshi's (JJJ). The two RME locks are transformed into NUMA-aware locks using the proposed RF and the existing cohort algorithms. Results show that, in multi-socket configurations, our transformation boosts the performance of the NUMA-oblivious RME locks by up to 45%. The RME locks transformed using the proposed RF lock are slower than their non-recoverable cohort variants by up to 9%. Outcomes demonstrate that the overhead of our algorithm is minimal when using a single socket. Moreover, a deeper empirical assessment shows that the gap in performance between GH and JJJ is due to the entry section of JJJ, not its exit section.

A multi-path integrate-and-fire circuit is proposed as a new method of providing prosthesis sensory feedback in an effort to improve the response to dynamic stimuli. The proposed integrate and fire (I&F) utilizes adjustable thresholds and offsets on each path to provide two degrees of freedom to correctly match afferent firing frequencies. The proposed I&F was designed in TSMC 65 nm and its functionality was confirmed through simulations and comparing the response to afferent firing patterns expected for three sets of excitations.

A noise-cancelling sturdy multi-stage noise-shaping (NC-SMASH) structure that cancels first stage quantization error is proposed. Without any modification to existing analog blocks in the traditional sturdy MASH (SMASH) structure, adding digital noise-cancelling filters (NCFs) will cancel the first stage quantization error. This quantization error cancellation, unlike the traditional MASH structure, is less sensitive to analog loop filter matching and digital filters. This is because the leakage of the first stage quantization error is shaped by the overall noise transfer function (NTF) of the proposed NC-SMASH structure. Therefore, this technique allows using low gain opamps without signal-to-noise ratio degradation and reducing the number of the first stage quantization levels. Detailed Matlab simulation results are provided to prove the efficiency of the proposed modulator.