The relentless growth in computational demands—from scientific simulations, AI training/inference, and large-scale data analytics—places immense pressure on high-performance computing (HPC) systems, and as architectures evolve toward exascale and beyond, the growing complexity of software stacks, heterogeneous hardware, and diverse workloads makes performance measurement, modeling, and efficient library implementations more critical than ever. This track centers on the three foundational pillars that bridge hardware capabilities and application performance: performance measurement methodologies for deep system-level insight, performance modeling techniques to guide architectural design and optimization, and high-performance libraries—including operator libraries, AI frameworks, and compute kernels—that convert peak theoretical performance into sustained practical throughput. We invite cutting-edge contributions spanning the full spectrum from low-level micro-benchmarking and profiling to high-level prediction, auto-tuning, and domain-specific optimizations, and we welcome researchers and practitioners from academia and industry to share their latest advances in performance engineering for parallel and distributed computing systems.
We are pleased to announce the important dates for PDCAT 2026.
Please mark your calendars!
All deadlines are based on Anywhere on Earth (AoE), at midnight on the specified date.