H

high-performance computing

XStore is a high-performance key-value store for time-series

Conduit Systems is an open-source system architecture platform for designing, simulating, and orchestrating data movement across chiplet-based computing systems. It focuses on memory-to-compute pathways, enabling modular hardware design, simulation-driven validation, and runtime-aware scheduling. Built under AGPL-3.0+, it serves as the foundation for next-generation AI and high-performance computing architectures. https://roxanneardary.com/conduitsystems/

CircuitPath is an open-source framework for advanced chiplet packaging and system integration, enabling modular AI and HPC architectures through standardized chiplet models, interposer layouts, and simulation-driven design. Built under AGPL-3.0+, it provides a fully extensible ecosystem for modeling bandwidth, thermal behavior, and data movement across multi-die systems, empowering contributors to shape the future of open chiplet-based computing. https://roxanneardary.com/circuitpath/

NeuroML-based neural simulator

An application that requires significant computation to be resilient is an essential feature, and dealing with faults is critical to running an application in some high-performance computing environments. The Dependability Library for Iterative Applications (DeLIA) proposed was designed to be applications that use high-performance techniques and have synchronization moments.

The NAStJA framework provides an easy way to enable massively parallel simulations for a wide range of multi-physics applications based on stencil algorithms.

Python API and CLI for remote submission and monitoring of Python applications on HPC.

A collection of high-performance, composable .NET primitives and data structures. This repository provides reusable building blocks for advanced threading, queueing, and priority management scenarios, designed for reliability and efficiency.

API documentation can be found at https://counterpointcollective.gitlab.io/prestoprimitives/.

Archive for all my labs at the "Principles of Performance Oriented Coding" course at Babes-Bolyai University.

SuperConga: Superconductivity simulation framework on GPU.

Standard Template Adaptive Parallel Library: https://parasollab.web.illinois.edu/research/stapl/

Service Discovery and Load Balancer implementation in Rust. Features include service registration, health checks, and load balancing with Prometheus metrics integration. Designed for high-performance microservices architecture.

Github repository of VerHem project source tree. A tensor-valued (up to 18 components) solving, post-processing and visualization parallel suite for 3D GL equations and complex geometries with adaptive mesh.

An application that requires significant computation to be resilient is an essential feature, and dealing with faults is critical to running an application in some high-performance computing environments. The Dependability Library for Iterative Applications (DeLIA) proposed was designed to be applications that use high-performance techniques and have synchronization moments.

A NAStJA Application