SearcharxivSearch

arXiv subjects

William P. Risk

Publications and source records attributed to William P. Risk.

4 recordsLinked to original sources

A Scalable NorthPole System with End-to-End Vertical Integration for Low-Latency and Energy-Efficient LLM Inference

A vertically integrated, end-to-end, research prototype system combines 288 NorthPole neural inference accelerator cards, offline training algorithms, a high-performance runtime stack, and a containerized inference pipeline to deliver a scalable and efficient cloud inference service. The system delivers 115 peta-ops at 4-bit integer precision and 3.7 PB/s of memory bandwidth across 18 2U servers, while consuming only 30 kW of power and weighing 730 kg in a 0.67 m^2 42U rack footprint. The system can run 3 simultaneous instances of the 8-billion-parameter open-source IBM Granite-3.3-8b-instruct model at 2,048 context length with 28 simultaneous users and a per-user inter-token latency of 2.8 ms. The system is scalable, modular, and reconfigurable, supporting various model sizes and context lengths, and is ideal for deploying agentic workflows for enterprise AI applications in existing data center (cloud, on-prem) environments. For example, the system can support 18 instances of a 3-billion-parameter model or a single instance of a 70-billion-parameter model.

cs.DC

A high-performance integrated single-photon detector for telecom wavelengths

We have integrated a commercial avalanche photodiode (APD) and the circuitry needed to operate it as a single-photon detector (SPD) onto a single PC-board. At temperatures accessible with Peltier coolers (~200-240K), the PCB-SPD achieves high detection efficiency (DE) at 1308 and 1545 nm with low dark count probability (e.g. ~10-6/bias pulse at DE=20%, 220 K), making it useful for quantum key distribution (QKD). The board generates fast bias pulses, cancels noise transients, amplifies the signals, and sends them to an on-board discriminator. A digital blanking circuit suppresses afterpulsing.

quant-ph

Autocompensating Quantum Cryptography

Quantum cryptographic key distribution (QKD) uses extremely faint light pulses to carry quantum information between two parties (Alice and Bob), allowing them to generate a shared, secret cryptographic key. Autocompensating QKD systems automatically and passively compensate for uncontrolled time dependent variations of the optical fiber properties by coding the information as a differential phase between orthogonally-polarized components of a light pulse sent on a round trip through the fiber, reflected at mid-course using a Faraday mirror. We have built a prototype system based on standard telecom technology that achieves a privacy-amplified bit generation rate of ~1000 bits/s over a 10-km optical fiber link. Quantum cryptography is an example of an application that, by using quantum states of individual particles to represent information, accomplishes a practical task that is impossible using classical means.

quant-ph

Enhanced Autocompensating Quantum Cryptography System

We have improved the hardware and software of our autocompensating system for quantum key distribution by replacing bulk optical components at the end stations with fiber-optic equivalents and implementing software that synchronizes end-station activities, communicates basis choices, corrects errors and performs privacy amplification over a local area network. The all fiber-optic arrangement provides stable, efficient and high-contrast routing of the photons. The low bit error rate leads to high error correction efficiency and minimizes data sacrifice during privacy amplification. Characterization measurements made on a number of commercial avalanche photodiodes are presented that highlight the need for improved devices tailored specifically for quantum information applications. A scheme for frequency shifting the photons returning from Alice's station to allow them to be distinguished from backscattered noise photons is also described. OCIS codes: 030.5260, 060.0060, 060.2360, 230.2240, 270.5570.

quant-ph