arXiv subjects

Moein Khazraee

Publications and source records attributed to Moein Khazraee.

3 recordsLinked to original sources

arXiv2025

Platform Architecture for Tight Coupling of High-Performance Computing with Quantum Processors

We propose an architecture, called NVQLink, for connecting high-performance computing (HPC) resources to the control system of a quantum processing unit (QPU) to accelerate workloads necessary to the operation of the QPU. We aim to support every physical modality of QPU and every type of QPU system controller (QSC). The HPC resource is optimized for real-time (latency-bounded) processing on tasks with latency tolerances of tens of microseconds. The network connecting the HPC and QSC is implemented on commercially available Ethernet and can be adopted relatively easily by QPU and QSC builders, and we report a round-trip latency measurement of 3.96 microseconds (max) with prospects of further optimization. We describe an extension to the CUDA-Q programming model and runtime architecture to support real-time callbacks and data marshaling between the HPC and QSC. By doing so, NVQLink extends heterogeneous, kernel-based programming to the QSC, allowing the programmer to address CPU, GPU, and FPGA subsystems in the QSC, all in the same C++ program, avoiding the use of a performance-limiting HTTP interface. We provide a pattern for QSC builders to integrate with this architecture by making use of multi-level intermediate representation dialects and progressive lowering to encapsulate QSC code.

quant-ph↗

arXiv2023

Rosebud: Making FPGA-Accelerated Middlebox Development More Pleasant

We introduce an approach to designing FPGA-accelerated middleboxes that simplifies development, debugging, and performance tuning by decoupling the tasks of hardware-accelerator implementation and software-application programming. Rosebud is a framework that links hardware accelerators to a high-performance packet processing pipeline through a standardized hardware/software interface. This separation of concerns allows hardware developers to focus on optimizing custom accelerators while freeing software programmers to reuse, configure, and debug accelerators in a fashion akin to software libraries. We show the benefits of the Rosebud framework by building a firewall based on a large blacklist and porting the Pigasus IDS pattern-matching accelerator in less than a month. Our experiments demonstrate that Rosebud delivers high performance, serving ~200 Gbps of traffic while adding only 0.7-7 microseconds of latency.

cs.AR↗

arXiv2022

TopoOpt: Co-optimizing Network Topology and Parallelization Strategy for Distributed Training Jobs

We propose TopoOpt, a novel direct-connect fabric for deep neural network (DNN) training workloads. TopoOpt co-optimizes the distributed training process across three dimensions: computation, communication, and network topology. We demonstrate the mutability of AllReduce traffic, and leverage this property to construct efficient network topologies for DNN training jobs. TopoOpt then uses an alternating optimization technique and a group theory-inspired algorithm called TotientPerms to find the best network topology and routing plan, together with a parallelization strategy. We build a fully functional 12-node direct-connect prototype with remote direct memory access (RDMA) forwarding at 100 Gbps. Large-scale simulations on real distributed training models show that, compared to similar-cost Fat-Tree interconnects, TopoOpt reduces DNN training time by up to 3.4x.

cs.NI↗