Filtered by: Simulation × Clear all

DGNA: Dissecting GPU NUMA Architecture through Microbenchmarking and Data Analysis

Changxi Liu, Yun Chen, Trevor E. Carlson 2026-07-25
Cache Coherence Memory Hierarchy Numa Microarchitecture Simulation × HPC

The problem is that modern GPU memory architectures, particularly NUMA mechanisms within L2 and DRAM, remain a black-box, hindering optimization and simulation. DGNA introduces a methodology using microbenchmarking and Gaussian mixture models to measure L2 and DRAM latency without relying on intrinsic instructions. Applied to NVIDIA's A100 and H100 GPUs, DGNA reveals NUMA node architecture, SM-NUMA relationships, and NUMA-aware memory allocation strategies for cache coherence. This matters because it is the first work to detail GPU memory subsystem NUMA architecture, enabling better application optimization and accurate simulator design.

PDF

A Flexible Sparsity-Aware FPGA Accelerator with Column-Wise Compression for Efficient CNN Inference

Amirhossein Zarei, Shervin Vakili 2026-07-25
Microarchitecture Simulation × Sparsity Accelerator Inference

Problem: Efficient CNN inference on resource-constrained FPGAs is challenged by the irregularity of sparsity patterns and hardware overhead. Method: SparHiXcel-v2 introduces a column-wise kernel compression technique and a hardware-algorithm co-design framework with ordering optimization and multi-phase structured pruning. Finding: On a cost-effective AMD Kintex UltraScale+ FPGA, SparHiXcel-v2 achieves over 2.5 TOPS and 210 GOP/s/W for VGG16, and over 1.1 TOPS and 72 GOP/s/W for ResNet18 with modest accuracy degradation. Why it matters: This work provides a flexible and efficient FPGA accelerator that balances sparsity flexibility and hardware efficiency for resource-constrained CNN inference.

PDF

Qoreo: Choreographic Programming for Quantum Distributed Systems

Jennifer Paykin, Steven Baldasty, Joseph P. Near, Christian Skalka 2026-07-25
Quant Ph Distributed Communication Network Simulation × Hardware

Qoreo introduces a choreographic programming language for quantum distributed systems, where an entire protocol is expressed as a single global program rather than separate actor processes. The method combines a local quantum language with linear types to enforce the no-cloning principle, a choreographic language for inter-actor communication, and a process language for network nodes. The authors prove type safety and define endpoint projection (EPP), which automatically derives a deadlock-free network of independent processes from any choreography, with the metatheory fully mechanized in Rocq. This matters because it eliminates tedious and error-prone manual coordination, ensuring correct quantum operations and deadlock-free execution for distributed quantum systems.

PDF