arXiv subjects

Luca Benini

Publications and source records attributed to Luca Benini.

At least 163 records · Page 9Linked to original sources

arXiv2025

A Reliable, Time-Predictable Heterogeneous SoC for AI-Enhanced Mixed-Criticality Edge Applications

Next-generation mixed-criticality Systems-on-chip (SoCs) for robotics, automotive, and space must execute mixed-criticality AI-enhanced sensor processing and control workloads, ensuring reliable and time-predictable execution of critical tasks sharing resources with non-critical tasks, while also fitting within a sub-2W power envelope. To tackle these multi-dimensional challenges, in this brief, we present a 16nm, reliable, time-predictable heterogeneous SoC with multiple programmable accelerators. Within a 1.2W power envelope, the SoC integrates software-configurable hardware IPs to ensure predictable access to shared resources, such as the on-chip interconnect and memory system, leading to tight upper bounds on execution times of critical applications. To accelerate mixed-precision mission-critical AI, the SoC integrates a reliable multi-core accelerator achieving 304.9 GOPS peak performance at 1.6 TOPS/W energy efficiency. Non-critical, compute-intensive, floating-point workloads are accelerated by a dual-core vector cluster, achieving 121.8 GFLOPS at 1.1 TFLOPS/W and 106.8 GFLOPS/mm2.

cs.AR↗

arXiv2025

A Multicast-Capable AXI Crossbar for Many-core Machine Learning Accelerators

To keep up with the growing computational requirements of machine learning workloads, many-core accelerators integrate an ever-increasing number of processing elements, putting the efficiency of memory and interconnect subsystems to the test. In this work, we present the design of a multicast-capable AXI crossbar, with the goal of enhancing data movement efficiency in massively parallel machine learning accelerators. We propose a lightweight, yet flexible, multicast implementation, with a modest area and timing overhead (12% and 6% respectively) even on the largest physically-implementable 16-to-16 AXI crossbar. To demonstrate the flexibility and end-to-end benefits of our design, we integrate our extension into an open-source 288-core accelerator. We report tangible performance improvements on a key computational kernel for machine learning workloads, matrix multiplication, measuring a 29% speedup on our reference system.

cs.AR↗

arXiv2025

Evaluating IOMMU-Based Shared Virtual Addressing for RISC-V Embedded Heterogeneous SoCs

Embedded heterogeneous systems-on-chip (SoCs) rely on domain-specific hardware accelerators to improve performance and energy efficiency. In particular, programmable multi-core accelerators feature a cluster of processing elements and tightly coupled scratchpad memories to balance performance, energy efficiency, and flexibility. In embedded systems running a general-purpose OS, accelerators access data via dedicated, physically addressed memory regions. This negatively impacts memory utilization and performance by requiring a copy from the virtual host address to the physical accelerator address space. Input-Output Memory Management Units (IOMMUs) overcome this limitation by allowing devices and hosts to use a shared virtual paged address space. However, resolving IO virtual addresses can be particularly costly on high-latency memory systems as it requires up to three sequential memory accesses on IOTLB miss. In this work, we present a quantitative evaluation of shared virtual addressing in RISC-V heterogeneous embedded systems. We integrate an IOMMU in an open-source heterogeneous RISC-V SoC consisting of a 64-bit host with a 32-bit accelerator cluster. We evaluated the system performance by emulating the design on FPGA and implementing compute kernels from the RajaPERF benchmark suite using heterogeneous OpenMP programming. We measure the transfers and computation time on the host and accelerators for systems with different DRAM access latencies. We first show that IO virtual address translation can account for 4.2% up to 17.6% of the accelerator's runtime for gemm (General Matrix Multiplication) at low and high memory bandwidth. Then, we show that in systems containing a last-level cache, this IO address translation cost falls to 0.4% and 0.7% under the same conditions, making shared virtual addressing and zero-copy offloading suitable for such RISC-V heterogeneous SoCs.

cs.AR↗

arXiv2025

LocalViT: Analyzing Locality in Vision Transformers

The aim of this paper is to study the influence of locality mechanisms in vision transformers. Transformers originated from machine translation and are particularly good at modelling long-range dependencies within a long sequence. Although the global interaction between the token embeddings could be well modelled by the self-attention mechanism of transformers, what is lacking is a locality mechanism for information exchange within a local region. In this paper, locality mechanism is systematically investigated by carefully designed controlled experiments. We add locality to vision transformers into the feed-forward network. This seemingly simple solution is inspired by the comparison between feed-forward networks and inverted residual blocks. The importance of locality mechanisms is validated in two ways: 1) A wide range of design choices (activation function, layer placement, expansion ratio) are available for incorporating locality mechanisms and proper choices can lead to a performance gain over the baseline, and 2) The same locality mechanism is successfully applied to vision transformers with different architecture designs, which shows the generalization of the locality concept. For ImageNet2012 classification, the locality-enhanced transformers outperform the baselines Swin-T, DeiT-T, and PVT-T by 1.0%, 2.6% and 3.1% with a negligible increase in the number of parameters and computational effort. Code is available at https://github.com/ofsoundof/LocalViT.

cs.CV↗

arXiv2025

Croc: An End-to-End Open-Source Extensible RISC-V MCU Platform to Democratize Silicon

Ensuring a continuous and growing influx of skilled chip designers and a smooth path from education to innovation are key goals for several national and international "Chips Acts". Silicon democratization can greatly benefit from end-to-end (from silicon technology to software) free and open-source (OS) platforms. We present Croc, an extensible RISC-V microcontroller platform explicitly targeted at hands-on teaching and innovation. Croc features a streamlined OS synthesis and an end-to-end OS implementation flow, ensuring full, unconstrained access to the design, the design automation tools, and the implementation technology. Croc uses the industry-proven, open-source CVE2 core, implementing the RV32I(EMC) instruction set architecture (ISA), enabling students to define and implement their own ISA extensions. MLEM, a tapeout of Croc in IHP's open 130 nm node completed in eight weeks by a team of just two students, demonstrates the platform's viability for hands-on teaching in schools, universities, or even on a self-education path. In spring 2025, ETH Zurich will utilize Croc for its curricular VLSI class, involving up to 80 students, producing up to 40 OS application-specific integrated circuit layouts, and completing up to five student-led system-on-chip tapeouts. The lecture notes and exercises are already available under a Creative Commons license.

cs.AR↗

arXiv2025

BodySense: An Expandable and Wearable-Sized Wireless Evaluation Platform for Human Body Communication

Wearable, wirelessly connected sensors have become a common part of daily life and have the potential to play a pivotal role in shaping the future of personalized healthcare. A key challenge in this evolution is designing long-lasting and unobtrusive devices. These design requirements inherently demand smaller batteries, inevitably increasing the need for energy-sensitive wireless communication interfaces. Capacitive Human Body Communication (HBC) is a promising, power-efficient alternative to traditional RF-based communication, enabling point-to-multipoint data and energy exchange. However, as this concept relies on capacitive coupling to the surrounding area, it is naturally influenced by uncontrollable environmental factors, making testing with classical setups particularly challenging. This work presents a customizable, wearable-sized, wireless evaluation platform for capacitive HBC, designed to enable realistic evaluation of wearable-to-wearable applications. Comparative measurements of channel gains were conducted using classical grid-connected and wireless Data Acquisition (DAQ) across various transmission distances within the frequency range of 4 MHz to 64 MHz and revealed an average overestimation of 18.15 dB over all investigated distances in the classical setup.

eess.SP↗

arXiv2025

PlaceIT: Placement-based Inter-Chiplet Interconnect Topologies

2.5D integration technology is gaining traction as it copes with the exponentially growing design cost of modern integrated circuits. A crucial part of a 2.5D stacked chip is a low-latency and high-throughput inter-chiplet interconnect (ICI). Two major factors affecting the latency and throughput are the topology of links between chiplets and the chiplet placement. In this work, we present PlaceIT, a novel methodology to jointly optimize the ICI topology and the chiplet placement. While state-of-the-art methods optimize the chiplet placement for a predetermined ICI topology, or they select one topology out of a set of candidates, we generate a completely new topology for each placement. Our process of inferring placement-based ICI topologies connects chiplets that are in close proximity to each other, making it particularly attractive for chips with silicon bridges or passive silicon interposers with severely limited link lengths. We provide an open-source implementation of our method that optimizes the placement of homogeneously or heterogeneously shaped chiplets and the ICI topology connecting them for a user-defined mix of four different traffic types. We evaluate our methodology using synthetic traffic and traces, and we compare our results to a 2D mesh baseline. PlaceIT reduces the latency of synthetic L1-to-L2 and L2-to-memory traffic, the two most important types for cache coherency traffic, by up to 28% and 62%, respectively. It also achieve an average packet latency reduction of up to 18% on traffic traces. PlaceIT enables the construction of 2.5D stacked chips with low-latency ICIs.

cs.AR↗

arXiv2025

A Multi-level Compiler Backend for Accelerated Micro-kernels Targeting RISC-V ISA Extensions

High-performance micro-kernels must fully exploit today's diverse and specialized hardware to deliver peak performance to DNNs. While higher-level optimizations for DNNs are offered by numerous compilers (e.g., MLIR, TVM, OpenXLA), performance-critical micro-kernels are left to specialized code generators or handwritten assembly. Even though widely-adopted compilers (e.g., LLVM, GCC) offer tuned backends, their CPU-focused input abstraction, unstructured IR, and general-purpose best-effort design inhibit tailored code generation for innovative hardware. We think it is time to widen the classical hourglass backend and embrace progressive lowering across a diverse set of structured abstractions to bring domain-specific code generation to compiler backends. We demonstrate this concept by implementing a custom backend for a RISC-V-based accelerator with hardware loops and streaming registers, leveraging knowledge about the hardware at levels of abstraction that match its custom ISA. We use incremental register allocation over structured IRs, while dropping classical spilling heuristics, and show up to 90% FPU utilization across key DNN kernels. By breaking the backend hourglass model, we reopen the path from domain-specific abstractions to specialized hardware.

cs.PL↗

arXiv2025

ArtistIC: An Open-Source Toolchain for Top-Metal IC Art and Ultra-High-Fidelity GDSII Renders

Open-source projects require outreach material to grow their community, secure funds, and strengthen their influence. Numbers, specifications, and facts alone are intangible to uninvolved people; using a clear brand and appealing visual material is thus ample to reach a broad audience. This is especially true for application-specific integrated circuits (ASICs) during the early stages of the development cycle without running prototype systems. This work presents ArtistIC, an open-source framework to brand ASICs with top-metal art and to render GDSII layouts with ultra-high fidelity reaching render densities below 25 nm/px and gigapixels-scale resolutions.

cs.OH↗

arXiv2025

Towards Reliable Systems: A Scalable Approach to AXI4 Transaction Monitoring

In safety-critical SoC applications such as automotive and aerospace, reliable transaction monitoring is crucial for maintaining system integrity. This paper introduces a drop-in Transaction Monitoring Unit (TMU) for AXI4 subordinate endpoints that detects transaction failures including protocol violations or timeouts and triggers recovery by resetting the affected subordinates. Two TMU variants address different constraints: a Tiny-Counter solution for tightly area-constrained systems and a Full-Counter solution for critical subordinates in mixed-criticality SoCs. The Tiny-Counter employs a single counter per outstanding transaction, while the Full-Counter uses multiple counters to track distinct transaction stages, offering finer-grained monitoring and reducing detection latencies by up to hundreds of cycles at roughly 2.5x the area cost. The Full-Counter also provides detailed error logs for performance and bottleneck analysis. Evaluations at both IP and system levels confirm the TMU's effectiveness and low overhead. In GF12 technology, monitoring 16-32 outstanding transactions occupies 1330-2616 um2 for the Tiny-Counter and 3452-6787 um2 for the Full-Counter; moderate prescaler steps reduce these figures by 18-39% and 19-32%, respectively, with no loss of functionality. Results from a full-system integration demonstrate the TMU's robust and precise monitoring capabilities in safety-critical SoC environments.

cs.AR↗

arXiv2025

TCDM Burst Access: Breaking the Bandwidth Barrier in Shared-L1 RVV Clusters Beyond 1000 FPUs

As computing demand and memory footprint of deep learning applications accelerate, clusters of cores sharing local (L1) multi-banked memory are widely used as key building blocks in large-scale architectures. When the cluster's core count increases, a flat all-to-all interconnect between cores and L1 memory banks becomes a physical implementation bottleneck, and hierarchical network topologies are required. However, hierarchical, multi-level intra-cluster networks are subject to internal contention which may lead to significant performance degradation, especially for SIMD or vector cores, as their memory access is bursty. We present the TCDM Burst Access architecture, a software-transparent burst transaction support to improve bandwidth utilization in clusters with many vector cores tightly coupled to a multi-banked L1 data memory. In our solution, a Burst Manager dispatches burst requests to L1 memory banks, multiple 32b words from burst responses are retired in parallel on channels with parametric data-width. We validate our design on a RISC-V Vector (RVV) many-core cluster, evaluating the benefits on different core counts. With minimal logic area overhead (less than 8%), we improve the bandwidth of a 16-, a 256-, and a 1024--Floating Point Unit (FPU) baseline clusters, without Tightly Coupled Data Memory (TCDM) Burst Access, by 118%, 226%, and 77% respectively. Reaching up to 80% of the cores-memory peak bandwidth, our design demonstrates ultra-high bandwidth utilization and enables efficient performance scaling. Implemented in 12-nm FinFET technology node, compared to the serialized access baseline, our solution achieves up to 1.9x energy efficiency and 2.76x performance in real-world kernel benchmarkings.

cs.AR↗

arXiv2025

AraXL: A Physically Scalable, Ultra-Wide RISC-V Vector Processor Design for Fast and Efficient Computation on Long Vectors

The ever-growing scale of data parallelism in today's HPC and ML applications presents a big challenge for computing architectures' energy efficiency and performance. Vector processors address the scale-up challenge by decoupling Vector Register File (VRF) and datapath widths, allowing the VRF to host long vectors and increase register-stored data reuse while reducing the relative cost of instruction fetch and decode. However, even the largest vector processor designs today struggle to scale to more than 8 vector lanes with double-precision Floating Point Units (FPUs) and 256 64-bit elements per vector register. This limitation is induced by difficulties in the physical implementation, which becomes wire-dominated and inefficient. In this work, we present AraXL, a modular and scalable 64-bit RISC-V V vector architecture targeting long-vector applications for HPC and ML. AraXL addresses the physical scalability challenges of state-of-the-art vector processors with a distributed and hierarchical interconnect, supporting up to 64 parallel vector lanes and reaching the maximum Vector Register File size of 64 Kibit/vreg permitted by the RISC-V V 1.0 ISA specification. Implemented in a 22-nm technology node, our 64-lane AraXL achieves a performance peak of 146 GFLOPs on computation-intensive HPC/ML kernels (>99% FPU utilization) and energy efficiency of 40.1 GFLOPs/W (1.15 GHz, TT, 0.8V), with only 3.8x the area of a 16-lane instance.

cs.AR↗

arXiv2025

Occamy: A 432-Core Dual-Chiplet Dual-HBM2E 768-DP-GFLOP/s RISC-V System for 8-to-64-bit Dense and Sparse Computing in 12nm FinFET

ML and HPC applications increasingly combine dense and sparse memory access computations to maximize storage efficiency. However, existing CPUs and GPUs struggle to flexibly handle these heterogeneous workloads with consistently high compute efficiency. We present Occamy, a 432-Core, 768-DP-GFLOP/s, dual-HBM2E, dual-chiplet RISC-V system with a latency-tolerant hierarchical interconnect and in-core streaming units (SUs) designed to accelerate dense and sparse FP8-to-FP64 ML and HPC workloads. We implement Occamy's compute chiplets in 12 nm FinFET, and its passive interposer, Hedwig, in a 65 nm node. On dense linear algebra (LA), Occamy achieves a competitive FPU utilization of 89%. On stencil codes, Occamy reaches an FPU utilization of 83% and a technology-node-normalized compute density of 11.1 DP-GFLOP/s/mm2,leading state-of-the-art (SoA) processors by 1.7x and 1.2x, respectively. On sparse-dense linear algebra (LA), it achieves 42% FPU utilization and a normalized compute density of 5.95 DP-GFLOP/s/mm2, surpassing the SoA by 5.2x and 11x, respectively. On, sparse-sparse LA, Occamy reaches a throughput of up to 187 GCOMP/s at 17.4 GCOMP/s/W and a compute density of 3.63 GCOMP/s/mm2. Finally, we reach up to 75% and 54% FPU utilization on and dense (LLM) and graph-sparse (GCN) ML inference workloads. Occamy's RTL is freely available under a permissive open-source license.

cs.AR↗

arXiv2025

Spatz: Clustering Compact RISC-V-Based Vector Units to Maximize Computing Efficiency

The ever-increasing computational and storage requirements of modern applications and the slowdown of technology scaling pose major challenges to designing and implementing efficient computer architectures. To mitigate the bottlenecks of typical processor-based architectures on both the instruction and data sides of the memory, we present Spatz, a compact 64-bit floating-point-capable vector processor based on RISC-V's Vector Extension Zve64d. Using Spatz as the main Processing Element (PE), we design an open-source dual-core vector processor architecture based on a modular and scalable cluster sharing a Scratchpad Memory (SCM). Unlike typical vector processors, whose Vector Register Files (VRFs) are hundreds of KiB large, we prove that Spatz can achieve peak energy efficiency with a latch-based VRF of only 2 KiB. An implementation of the Spatz-based cluster in GlobalFoundries' 12LPP process with eight double-precision Floating Point Units (FPUs) achieves an FPU utilization just 3.4% lower than the ideal upper bound on a double-precision, floating-point matrix multiplication. The cluster reaches 7.7 FMA/cycle, corresponding to 15.7 DP-GFLOPS and 95.7 DP-GFLOPS/W at 1 GHz and nominal operating conditions (TT, 0.80V, 25C), with more than 55% of the power spent on the FPUs. Furthermore, the optimally-balanced Spatz-based cluster reaches a 95.0% FPU utilization (7.6 FMA/cycle), 15.2 DP-GFLOPS, and 99.3 DP-GFLOPS/W (61% of the power spent in the FPU) on a 2D workload with a 7x7 kernel, resulting in an outstanding area/energy efficiency of 171 DP-GFLOPS/W/mm2. At equi-area, the computing cluster built upon compact vector processors reaches a 30% higher energy efficiency than a cluster with the same FPU count built upon scalar cores specialized for stream-based floating-point computation.

cs.AR↗

arXiv2025

A "New Ara" for Vector Computing: An Open Source Highly Efficient RISC-V V 1.0 Vector Processor Design

Vector architectures are gaining traction for highly efficient processing of data-parallel workloads, driven by all major ISAs (RISC-V, Arm, Intel), and boosted by landmark chips, like the Arm SVE-based Fujitsu A64FX, powering the TOP500 leader Fugaku. The RISC-V V extension has recently reached 1.0-Frozen status. Here, we present its first open-source implementation, discuss the new specification's impact on the micro-architecture of a lane-based design, and provide insights on performance-oriented design of coupled scalar-vector processors. Our system achieves comparable/better PPA than state-of-the-art vector engines that implement older RVV versions: 15% better area, 6% improved throughput, and FPU utilization >98.5% on crucial kernels.

cs.AR↗

arXiv2025

Toward Attention-based TinyML: A Heterogeneous Accelerated Architecture and Automated Deployment Flow

One of the challenges for Tiny Machine Learning (tinyML) is keeping up with the evolution of Machine Learning models from Convolutional Neural Networks to Transformers. We address this by leveraging a heterogeneous architectural template coupling RISC-V processors with hardwired accelerators supported by an automated deployment flow. We demonstrate Attention-based models in a tinyML power envelope with an octa-core cluster coupled with an accelerator for quantized Attention. Our deployment flow enables end-to-end 8-bit Transformer inference, achieving leading-edge energy efficiency and throughput of 2960 GOp/J and 154 GOp/s (0.65 V, 22 nm FD-SOI technology).

cs.AR↗

arXiv2024

Open-Source Heterogeneous SoCs for AI: The PULP Platform Experience

Since 2013, the PULP (Parallel Ultra-Low Power) Platform project has been one of the most active and successful initiatives in designing research IPs and releasing them as open-source. Its portfolio now ranges from processor cores to network-on-chips, peripherals, SoC templates, and full hardware accelerators. In this article, we focus on the PULP experience designing heterogeneous AI acceleration SoCs - an endeavour encompassing SoC architecture definition; development, verification, and integration of acceleration IPs; front- and back-end VLSI design; testing; development of AI deployment software.

cs.AR↗

arXiv2024

EnhancePPG: Improving PPG-based Heart Rate Estimation with Self-Supervision and Augmentation

Heart rate (HR) estimation from photoplethysmography (PPG) signals is a key feature of modern wearable devices for health and wellness monitoring. While deep learning models show promise, their performance relies on the availability of large datasets. We present EnhancePPG, a method that enhances state-of-the-art models by integrating self-supervised learning with data augmentation (DA). Our approach combines self-supervised pre-training with DA, allowing the model to learn more generalizable features, without needing more labelled data. Inspired by a U-Net-like autoencoder architecture, we utilize unsupervised PPG signal reconstruction, taking advantage of large amounts of unlabeled data during the pre-training phase combined with data augmentation, to improve state-of-the-art models' performance. Thanks to our approach and minimal modification to the state-of-the-art model, we improve the best HR estimation by 12.2%, lowering from 4.03 Beats-Per-Minute (BPM) to 3.54 BPM the error on PPG-DaLiA. Importantly, our EnhancePPG approach focuses exclusively on the training of the selected deep learning model, without significantly increasing its inference latency

eess.SP↗