Related papers: tuGEMM: Area-Power-Efficient Temporal Unary GEMM A…
General matrix multiplication (GEMM) is the computational backbone of modern AI workloads, and its efficiency is critically dependent on effective tiling strategies. Conventional approaches employ symmetric tile buffering, where the…
The advent of 1-bit large language models (LLMs), led by BitNet b1.58, has spurred interest in ternary LLMs. Despite this, research and practical applications focusing on efficient edge inference for ternary LLMs remain scarce. To bridge…
General matrix-matrix multiplications with double-precision real and complex entries (DGEMM and ZGEMM) in vendor-supplied BLAS libraries are best optimized for square matrices but often show bad performance for tall & skinny matrices, which…
Geometric graph neural networks (GNNs) excel at capturing molecular geometry, yet their locality-biased message passing hampers the modeling of long-range interactions. Current solutions have fundamental limitations: extending cutoff radii…
Mixed-precision quantization improves the budget--accuracy trade-off for large language models (LLMs) by allocating more bits to sensitive modules. However, automating this allocation at LLM scale faces a unique combination of constraints:…
In most domains of network analysis researchers consider networks that arise in nature with weighted edges. Such networks are routinely dichotomized in the interest of using available methods for statistical inference with networks. The…
Sparse general matrix-matrix multiplication (SpGEMM) is a critical operation in many applications. Current multithreaded implementations are based on Gustavson's algorithm and often perform poorly on large matrices due to limited cache…
In this work, we propose a ferroelectric FET(FeFET) time-domain compute-in-memory (TD-CiM) array as a homogeneous processing fabric for binary multiplication-accumulation (MAC) and content addressable memory (CAM). We demonstrate that: i)…
The recently proposed "generalized min-max" (GMM) kernel can be efficiently linearized, with direct applications in large-scale statistical learning and fast near neighbor search. The linearized GMM kernel was extensively compared in with…
A Time Projection Chamber (TPC) is an ideal device for the detection of charged particle tracks in a large volume covering a solid angle of almost $4\pi$. The high density of hits on a given particle track facilitates the task of pattern…
Generalized Sparse Matrix-Matrix Multiplication (SpGEMM) is a ubiquitous task in various engineering and scientific applications. However, inner product based SpGENN introduces redundant input fetches for mismatched nonzero operands, while…
Ternary weight quantization (e.g., BitNet b1.58) offers a promising path to mitigate the memory bandwidth bottleneck in Large Language Model (LLM) inference. However, conventional compute platforms lack native support for ternary-weight…
To address the increasing computational demands of artificial intelligence (AI) and big data, compute-in-memory (CIM) integrates memory and processing units into the same physical location, reducing the time and energy overhead of the…
Mobile edge generation (MEG) is an emerging technology that allows the network to meet the challenging traffic load expectations posed by the rise of generative artificial intelligence~(GAI). A novel MEG model is proposed for deploying GAI…
In recent years Gas Electron Multipliers have proven to be reliable amplification stages at high beam rates, and can be used also in Time Projection Chambers. Our group developed a 1 dm$^3$ active volume double-GEM TPC, with spatial…
The application of on-device language models (ODLMs) on resource-constrained edge devices is a multi-dimensional problem that strikes a fine balance between computational effectiveness, memory, power usage, and linguistic capacity across…
Medical image segmentation is essential for clinical diagnosis and treatment planning. Although transformer-based methods have achieved remarkable results, their high computational cost hinders clinical deployment. To address this issue, we…
The sizes of GPU applications are rapidly growing. They are exhausting the compute and memory resources of a single GPU, and are demanding the move to multiple GPUs. However, the performance of these applications scales sub-linearly with…
Unified multimodal models (UMMs) achieve strong performance in both understanding and generation by learning a shared latent space, yet they often exhibit functional inconsistency between these two capabilities. We observe that this issue…
Tensor processing units (TPUs), specialized hardware accelerators for machine learning tasks, have shown significant performance improvements when executing convolutional layers in convolutional neural networks (CNNs). However, they…