Ectothermic animals in areas characterised by seasonal changes are susceptible to extreme fluctuations in temperature. To survive through varied temperatures, ectotherms have developed unique strategies. This study focuses on synaptic transmission function at cold temperatures, as it is a vital component of ectothermic animals' survival. For determining how synaptic transmission is influenced by an acute change in temperature (20°C to 10°C within a minute) and chronic cold (10°C), the crayfish (Procambarus clarkii) neuromuscular junction (NMJ) was used as a model. To simulate chronic cold conditions, crayfish were acclimated to 15°C for 1 week and then to 10°C for 1 week. They were then used to examine the synaptic properties associated with the low output nerve terminals on the opener muscle in the walking legs and high output innervation on the abdominal deep extensor muscle. The excitatory postsynaptic potentials (EPSPs) of the opener NMJs increased in amplitude with acute warming (20°C) after being acclimated to cold; however, the deep extensor muscles showed varied changes in EPSP amplitude. Synaptic transmission at both NMJs was enhanced with exposure to the modulators serotonin or octopamine. The membrane resistance of the muscles decreased 33% and the resting membrane potential hyperpolarised upon warm exposure. Analysis of haemolymph indicated that octopamine increases during cold exposure. These results suggest bioamine modulation as a possible mechanism for ensuring that synaptic transmission remains functional at low temperatures.

The emergence of cloud-computing has raised important privacy questions about the data that users share with remote servers. While data in transit is protected using standard techniques like Transport Layer Security (TLS), most cloud providers have unrestricted plaintext access to user data at the endpoint. Fully Homomorphic Encryption (FHE) offers one solution to this problem by allowing for arbitrarily complex computations on encrypted data without ever needing to decrypt it. Unfortunately, all known implementations of FHE require the addition of noise during encryption which grows during computation. As a result, sustaining deep computations requires a periodic noise reduction step known as bootstrapping. The cost of the bootstrapping operation is one of the primary barriers to the wide-spread adoption of FHE.In this paper, we present an in-depth architectural analysis of the bootstrapping step in FHE. First, we observe that secure implementations of bootstrapping exhibit a low arithmetic intensity (< 1 Op/byte), require large caches (> 100MB) and as such, are heavily bound by the main memory bandwidth. Consequently, we demonstrate that existing workloads observe marginal performance gains from the design of bespoke highthroughput arithmetic units tailored to FHE. Secondly, we propose several cache-friendly algorithmic optimizations that improve the throughput in FHE bootstrapping by enabling up to 3.2× higher arithmetic intensity and 4.6× lower memory bandwidth. Our optimizations apply to a wide range of structurally similar computations such as private evaluation and training of machine learning models. Finally, we incorporate these optimizations into an architectural tool which, given a cache size, memory subsystem, the number of functional units and a desired security level, selects optimal cryptosystem parameters to maximize the bootstrapping throughput.Our optimized bootstrapping implementation represents a best-case scenario for compute acceleration of FHE. We show that despite these optimizations, bootstrapping (as well as other applications with similar computational structure) continue to remain bottlenecked by main memory bandwidth. We thus conclude that secure FHE implementations need to look beyond accelerated compute for further performance improvements and to that end, we propose new research directions to address the underlying memory bottleneck. In summary, our answer to the titular question is: yes, but only after addressing the memory bottleneck!

No abstract