Related papers: DiP: A Scalable, Energy-Efficient Systolic Array f…

Deep Neural Networks (DNNs) and Large Language Models (LLMs) have revolutionized artificial intelligence, yet their deployment faces significant memory and computational challenges, especially in resource-constrained environments.…

In the big data era of observational oceanography, passive acoustics datasets are becoming too high volume to be processed on local computers due to their processor and memory limitations. As a result there is a current need for our…

We present FastBoost, a parameter-efficient neural architecture that achieves state-of-the-art performance on CIFAR benchmarks through a novel Dynamically Scaled Progressive Attention (DSPA) mechanism. Our design establishes new efficiency…

We propose DFModel, a modeling framework for mapping dataflow computation graphs onto large-scale systems. Mapping a workload to a system requires optimizing dataflow mappings at various levels, including the inter-chip (between chips)…

Transformer architectures, particularly Diffusion Transformers (DiTs), have become widely used in diffusion and flow-matching models due to their strong performance compared to convolutional UNets. However, the isotropic design of DiTs…

Diffusion-based generation is increasingly powering production content pipelines; however, deploying these models at scale remains a significant challenge. Model weights frequently exceed the memory capacity of commodity GPUs, while the…

Most of the existing work on FPGA acceleration of Convolutional Neural Network (CNN) focus on employing a single strategy (algorithm, dataflow, etc.) across all the layers. Such an approach does not achieve optimal latency on complex and…

Benefiting from the self-attention mechanism, Transformer models have attained impressive contextual comprehension capabilities for lengthy texts. The requirements of high-throughput inference arise as the large language models (LLMs)…

The research interest in specialized hardware accelerators for deep neural networks (DNN) spikes recently owing to their superior performance and efficiency. However, today's DNN accelerators primarily focus on accelerating specific…

In-memory computing hardware accelerators allow more than 10x improvements in peak efficiency and performance for matrix-vector multiplications (MVM) compared to conventional digital designs. For this, they have gained great interest for…

Processing-in-Memory (PIM) enhances memory with computational capabilities, potentially solving energy and latency issues associated with data transfer between memory and processors. However, managing concurrent computation and data flow…

Region proposal is critical for object detection while it usually poses a bottleneck in improving the computation efficiency on traditional control-flow architectures. We have observed region proposal tasks are potentially suitable for…

Deep neural networks (DNNs) have been widely applied in our society, yet reducing power consumption due to large-scale matrix computations remains a critical challenge. MADDNESS is a known approach to improving energy efficiency by…

Training large models with distributed data parallelism (DDP) requires frequent communication of gradients across workers, which can saturate bandwidth. Infrequent communication strategies (e.g., Local SGD) reduce this overhead but, when…

Scientific workflows bridge scientific challenges with computational resources. While dispel4py, a stream-based workflow system, offers mappings to parallel enactment engines like MPI or Multiprocessing, its optimization primarily focuses…

It is a challenging task to train large DNN models on sophisticated GPU platforms with diversified interconnect capabilities. Recently, pipelined training has been proposed as an effective approach for improving device utilization. However,…

The state-of-the-art accelerators for Convolutional Neural Networks (CNNs) typically focus on accelerating only the convolutional layers, but do not prioritize the fully-connected layers much. Hence, they lack a synergistic optimization of…

In recent years, Dynamic Sparse Training (DST) has emerged as an alternative to post-training pruning for generating efficient models. In principle, DST allows for a more memory efficient training process, as it maintains sparsity…

Current inference systems for Mixture-of-Experts (MoE) models primarily employ static parallelization strategies. However, these static approaches cannot consistently achieve optimal performance across different inference scenarios, as they…

Diffusion Transformer (DiT), an emerging diffusion model for image generation, has demonstrated superior performance but suffers from substantial computational costs. Our investigations reveal that these costs stem from the static inference…