Compute Solution for Tesla's Full Self-Driving Computer
The DOI matches its Crossref record; 4 of the 4 references it cites were found in a registry or an index.
Abstract
Tesla's full self-driving (FSD) computer is the world's first purpose-built computer for the highly demanding workloads of autonomous driving. It is based on a new System on a Chip (SoC) that integrates industry standard components such as CPUs, ISP, and GPU, together with our custom neural network accelerators. The FSD computer is capable of processing up to 2300 frames per second, a 21× improvement over Tesla's previous hardware and at a lower cost, and when fully utilized, enables a new level of safety and autonomy on the road.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“In the past few years, some companies and universities have presented application specific integrated circuits (ASICs) based on the complementary metal oxide semiconductor (CMOS) technology that are capable to compute and store information in the same unit.”
From Hardware implementation of memristor-based artificial neural networks (2024) · cites this paper as [21]
Citations
Cites
In-Datacenter Performance Analysis of a Tensor Processing Unit
Jouppi, N. P., Young, C., Patil, N., Patterson, D., Agrawal, G., Bajwa, R., Bates, S., Bhatia, S., Boden, N., Borchers, A., et al. (2017) · Proceedings of the 44th Annual International Symposium on Computer Architecture · doi:10.1145/3079856.3080246 · Cited by 4,537 · cited 1×
“Thus, each MAC unit computes the dotproduct locally with no data movement within the MAC array, unlike the systolic array computations implemented in many other processors.(3,4) This results in lower power and less cell area than systolic array implementations.”
Deep Convolutional Neural Networks for Image Classification: A Comprehensive Review
Rawat, W., & Wang, Z. (2017) · Neural Computation · doi:10.1162/neco_a_00990 · Cited by 3,503
Object Recognition with Gradient-Based Learning
LeCun, Y., Haffner, P., Bottou, L., & Bengio, Y. (1999) · Lecture Notes in Computer Science · doi:10.1007/3-540-46805-6_19 · Cited by 947
Floating point HOG implementation for real-time multiple object detection
Komorkiewicz, M., Kluczewski, M., & Gorgoń, M. (2012) · International Conference on Field Programmable Logic and Applications (FPL) · doi:10.1109/fpl.2012.6339159 · Cited by 52
Research works printed but not linked: 1
- horowitz, 1999, Computing's energy problem, IEEE Int Solid-State Circuits Conf Dig Tech Papers, 10
Cited by
Items 101–200 of 206
Falcão, G., & Ferreira, J. D. (2023) · Communications of the ACM · doi:10.1145/3589995 · Cited by 4
Czekalla, M., & Seibel, A. (2023) · Springer tracts in additive manufacturing · doi:10.1007/978-3-031-42983-5_24 · Cited by 4
Zhang, W., Zhao, F., & Liu, Z. (2025) · Sustainability · doi:10.3390/su17041753 · Cited by 7
Xu, Y., Chen, S., Yang, X., Xiang, Y., Yu, J., Ding, W., Wang, J., & Wang, Y. (2025) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2025.3548504 · Cited by 4
“Currentembeddedrobotsaremainlyequippedwithembedded GPUs [2], [31] or NPUs [17], [19] to enable efficient neural network inference at the edge.”
Hardware Specialization: From Cell to Heterogeneous Microprocessors Everywhere
Swaminathan, K., & Vega, A. J. (2021) · IEEE Micro · doi:10.1109/mm.2021.3114882 · Cited by 4
EN-T: Optimizing Tensor Computing Engines Performance via Encoder-Based Methodology
Wu, Q., Gui, Y., Zeng, Z., Wang, X., Liang, H., & Jin, X. (2024) · IEEE International Conference on Computer Design: VLSI in Computers and Processors (ICCD) · doi:10.1109/iccd63220.2024.00097 · Cited by 4
“The primary distinction among these architectures lies in the optimization and alteration of the data flow paths and interconnect topologies within the multiplier arrays, aiming to enhance data reuse for matrix multiplication within the Tensor Computing Units (TCUs).”
Lee, S., Kim, B., Choi, Y., & Lee, H. (2025) · ACM Transactions on Architecture and Code Optimization · doi:10.1145/3711821 · Cited by 6
“In particular, systolic array-based architectures have gained popularity in commercial NPUs [4, 30, 35, 40, 48, 49, 52] since the introduction of Google’s Tensor Processing Unit [40].”
InfluentialXie, T., Qi, Y., Wen, J., Wan, Z., Dong, Y., Wang, Z., Cai, S., Liang, Y., Jia, T., Wang, Y., et al. (2026) · International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS) · doi:10.1145/3779212.3790147 · Cited by 3
“For example, at the circuit level , redundancy-based approaches, e.g., dual modular redundancy (DMR) [37–39], are widely adopted in autonomous vehicles for error detection but require multiple copies of computations to be performed.”
Enabling Plug-and-Play in Cyber-Physical Systems Using MPSoC-FPGAs
Passaretti, D., Steiger, M., & Pionteck, T. (2023) · IEEE Access · doi:10.1109/access.2023.3325742 · Cited by 3
“For instance, Tesla and AMD-Xilinx have introduced a Full Self-Driving (FSD) hardware architecture for Advanced Driver-Assistance Systems (ADAS) [25], [26].”
BackgroundTopology-Aware Virtualization over Inter-Core Connected Neural Processing Units
Feng, D., Feng, E., Du, D., Xu, P., Xia, Y., Chen, H., & Zhao, R. J. (2025) · Proceedings - International Symposium on Computer Architecture/Proceedings · doi:10.1145/3695053.3731002 · Cited by 6
An Architecture for Human Action Recognition in Smart Cities Video Surveillance Systems
Llaurado-Fons, J. M., Martínez, A. C., Pujol, F. A., & Mora, H. (2021) · Springer proceedings in complexity · doi:10.1007/978-3-030-62066-0_5 · Cited by 3
Quality of Fault Injection Strategies on Hardware Accelerator
Guinebert, I., Barrilado, A., Delmas, K., Galtié, F., & Pagetti, C. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-14835-4_15 · Cited by 3
Surazhevsky, I. A., Minnekhanov, A. A., & Demin, V. A. (2021) · Nanobiotechnology Reports · doi:10.1134/s2635167621020154 · Cited by 3
Minimizing Age of Processed Information Over Unreliable Wireless Network Channels
Meesena, W., Nikunram, C., Turner, S. J., & Supittayapornpong, S. (2024) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2024.3520913 · Cited by 2
“A notable example of a system that employs a set of dedicated processors is Tesla’s full self-driving system [5].”
Unified Memory Protection with Multi-granular MAC and Integrity Tree for Heterogeneous Processors
Lee, S., Na, S., Choi, J., Pyo, J., & Huh, J. (2025) · Proceedings - International Symposium on Computer Architecture/Proceedings · doi:10.1145/3695053.3731066 · Cited by 3
Cloud Is Closer Than It Appears: Revisiting the Tradeoffs of Distributed Real-Time Inference
Sharma, P., Qiu, H., & Srivastava, M. B. (2025) · International Conference on Computer Communications and Networks (ICCCN) · doi:10.1109/icccn65249.2025.11133852 · Cited by 3
“Systems such as Waymo [4] and Tesla’s FSD software [5] are prime examples of this design philosophy, which favors tightly bounded and deterministic compute pipelines on onboard GPUs.”
Early Reliability Estimation in Hardware Accelerators using Improved Colored Petri Nets
Castillo, E. C. V., da Silva, F. A., Condia, J. E. R., Guerrero-Balaguera, J.-D., & Glaß, M. (2025) · IEEE International Test Conference (ITC) · doi:10.1109/itc58126.2025.00074 · Cited by 2
“…of the three architectural variants for Dot-Product Units (DPUs), which are the fundamental in-chip component in modern hardware accelerators for Artificial Intelligence, including Tensor Cores or Matrix Cores in GPUs, Neural Processing Units , and sophisticated automotive accelerators [5] [6].”
BayNNgine: Hardware-Enabled Bayesian Neural Network Support for Dependable AI Inference
Hoefer, J., Lesniak, F., Gutermann, A., Wöran, F., Harbaum, T., & Becker, J. (2025) · IEEE International System-on-Chip Conference (SOCC) · doi:10.1109/socc66126.2025.11235419 · Cited by 2
“…deployment of AI models in safety-critical applications with real-time constraints is now feasible due to advancements in AI hardware platforms integrating Neural Processing Units (NPUs) [6] [7], current state-of-the-art platforms lack inherent support for such architectural redundancy patterns.”
Emerging challenges of electric truck and truck platooning on flexible pavement responses
Shen, K., & Wang, H. (2025) · International Journal of Pavement Engineering · doi:10.1080/10298436.2025.2601712 · Cited by 2
“With sufficient electric supply, electric trucks also enable the integration of intelligent chips in vehicles, advancing autonomous driving technologies (Talpes et al., 2020).”
Establishing a Data-Efficient Witness Protocol for Connected Autonomous Vehicles
Chaisawat, S., Paik, H.-Y. H., & Kanhere, S. S. (2026) · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering · doi:10.1007/978-3-032-10554-7_25 · Cited by 2
Applications and Challenges of Multi-Core Scheduling in Intelligent Automotive Systems
Wei, Y., Li, N., An, Y., Liu, S. H., & Li, Y. Y. (2026) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2026.3670003 · Cited by 2
Jiao, L., Gao, L., Zheng, J., Yang, P., & Zhang, Z. (2026) · Computer Networks · doi:10.1016/j.comnet.2026.112283 · Cited by 2
Redundant dataflow applications on clustered manycore architectures
Kühbacher, C., Ungerer, T., & Altmeyer, S. (2022) · ACM Symposium on Applied Computing · doi:10.1145/3477314.3507272 · Cited by 2
Application of Cloud Robotics in Autonomous Vehicles
Raj, E. F. I., Balaji, M., Subathira, N. P., & Darwin, S. (2024) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3693-1914-7.ch020 · Cited by 2
Szabó, R. (2025) · Technologies · doi:10.3390/technologies13010027 · Cited by 2
Zhu, Y., Liu, K., Gong, Y., Liu, Y., Yanɡ, C., Luo, M., Zhang, H., Chen, D., Xiao, X., & Yu, S. (2025) · Journal of Lightwave Technology · doi:10.1109/jlt.2025.3570551 · Cited by 5
“…the large-scale rise of artificial intelligence technology, neural networks (NN) technology has been massively applied to automatic driving (e.g., Tesla FSD [1]), image generation (e.g., Stable Diffusion [2] and Midjourney [3]), natural language processing (e.g., Chatgpt [4]) and other scenarios.”
Mu, J., & Zhou, L. (2025) · Scientific Reports · doi:10.1038/s41598-025-15442-6 · Cited by 2
Design of (Semi-)Autonomous Vehicles: Perceptions of the People in Sweden
Rosenholm, L., Goswami, P., & Jagtap, S. (2022) · Proceedings of the Design Society · doi:10.1017/pds.2022.174 · Cited by 2
Falcão, G., & Ferreira, J. D. (2022) · Queue · doi:10.1145/3580503 · Cited by 2
Guiding a Harsh-Environments Robust Detector via RAW Data Characteristic Mining
Chen, H., Tai, H.-S., & Ma, K. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i2.27867 · Cited by 4
“Tesla Vision focuses on a solely camera-based approach, enabling a new so-lution through multi-camera and multi-sensor fusion (Talpes et al. 2020; Ajitha and Nagra 2021).”
BackgroundEnabling Level-4 Autonomous Driving on a Single $1k Off-the-Shelf CardPreprint
Sung, H.-H., Xu, Y., Guan, J., Niu, W., Liu, S., Ren, B., Wang, Y., & Shen, X. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2110.06373 · Cited by 3
“with redundant power supplies equipped on each device [38].”
BackgroundEfficient Deployment of Deep Neural Networks on Hardware Devices for Edge AI
Bouzidi, H. (2024) · theses.fr (ABES) · doi:10.70675/a25bbd6cz8a5bz4d93za232z46df005062c4 · Cited by 8
Temperature-Aware Memory Mapping and Active Cooling of Neural Processing Units
Moghaddas, V., Kattan, H., Bücher, T., Yayla, M., Chen, J.-J., & Amrouch, H. (2023) · International Symposium on Low Power Electronics and Design (ISLPED) · doi:10.1109/islped58423.2023.10244458 · Cited by 3
Chen, Y., Chiu, Y.-T., Tu, H.-J., & Huang, C.-T. (2024) · IEEE Transactions on Very Large Scale Integration (VLSI) Systems · doi:10.1109/tvlsi.2024.3488042 · Cited by 2
“The recent trend in CNN accelerators is to incorporate a wide variety of CNN models and to efficiently support models with large numbers of weights, such as TPU [29] and Tesla NPU [30].”
About the Functional In-Field Self-Testing of AI Accelerators
Guerrero-Balaguera, J.-D., Condia, J. E. R., Sierra, R. L., & Reorda, M. S. (2025) · IEEE Latin American Test Symposium (LATS) · doi:10.1109/lats65346.2025.10963960 · Cited by 1
“In fact, equivalent structures are commonly found in the modern DLA generations, including the NVDLA accelerator [17], the NVIDIA Tensor Cores (TCU) [19], [23], the Tesla Neural Processor [21], [24], and specialized transformer neural network accelerators (e.g., MAGNet [18]).”
Investigating the Automotive HPC Safety Architecture for Software Defined Vehicle
Park, J., Jeong, J., Kim, B., Park, S., Lee, M., Yun, D., & Park, N. (2025) · IEEE International Reliability Physics Symposium proceedings · doi:10.1109/irps48204.2025.10982897 · Cited by 1
“In our approach, we have taken power redundancy a step further than conventional designs, which typically only provide independent power connections [12].”
Early Reliability Assessment of AI-based Automotive Systems
Hegde, S. S., Selvaraj, D. C., Condia, J. E. R., Amati, N., Chiasserini, C. F., Deflorio, F. P., & Reorda, M. S. (2025) · ACM Transactions on Internet of Things · doi:10.1145/3758327 · Cited by 1
“…acceleration platforms with extended capabilities, including embedded FPGAs, Vector Processors in CPUs, Neural Processing Units ( NPUs ), Tensor Processing Units ( TPUs ), Graphics Processing Units ( GPUs ) or customized chips (e.g., Tesla’s Full Self-Driving or “ FSD ”) [11, 16, 17, 43, 57].”
Data-Driven Vehicle Autonomy: A Comprehensive Review of Sensor Fusion, Localisation, and Control
Borra, P. (2025) · Lecture notes in electrical engineering · doi:10.1007/978-981-95-3616-0_25 · Cited by 4
A Predictive Framework for Fault Vulnerabilities in Edge Hardware Accelerators
Castillo, E. C. V., Condia, J. E. R., Guerrero-Balaguera, J.-D., da Silva, F. A., & Glaß, M. (2026) · IEEE Latin American Symposium on Circuits and Systems (LASCAS) · doi:10.1109/lascas67804.2026.11457081 · Cited by 1
“On the other hand, Tensor Cores (TCs) [14], [15] accelerate matrix multiplication workloads.”
Neuromorphic Denoising with Fully Analog Memristive In‐Memory Computing
Shi, D., Zhang, T., Li, Y., Tao, Y., Yan, B. Y., & Yang, Y. (2026) · Advanced Intelligent Systems · doi:10.1002/aisy.70419 · Cited by 1
Functional Test Sequences for Inducing Voltage Droops in a Multi-Threaded Processor
Kalyanam, V. K., Mahurin, E., Spence, M., & Abraham, J. A. (2020) · IEEE International Test Conference (ITC) · doi:10.1109/itc44778.2020.9325271 · Cited by 1
“Processor systems consume increasingly higher amounts of power as a result of advances in processing capabilities, increased clock frequencies (F CLK ) and high-performance requirements driven by the needs of technological developments [1-4].”
BackgroundA Note on Big Data and Value Creation
Moreno, M. Á., & Carou, D. (2022) · Management and industrial engineering · doi:10.1007/978-3-030-91006-8_1 · Cited by 2
Arthanto, Y. F., Ojika, D., & Kim, J.-Y. (2022) · International Conference on Field Programmable Logic and Applications (FPL) · doi:10.1109/fpl57034.2022.00042 · Cited by 3
“Recently, as artificial intelligence (AI) and machine learning (ML) technology are transforming major industries with highly beneficial applications such as image captioning [4], [5], virtual assistant [6], stock market prediction [7], and self-driving cars [8], HPC systems have evolved to accommodate AI workloads [9]…”
BackgroundTCRing: A Novel Single-Channel Deadlock-Free Concentrated on-Chip Network
Zhenjiang, G., Wang, H., Longbing, Z., & Junhua, X. (2022) · IEEE International Conference on High Performance Computing and Communications; IEEE International Conference on Smart City; IEEE International Conference on Data Science and Systems (HPCC/SmartCity/DSS) · doi:10.1109/hpcc-dss-smartcity-dependsys57074.2022.00216 · Cited by 1
Li, X., Ye, Z., Wei, S., Chen, Z., Chen, X., Tian, Y., Dang, J., Fu, S., & Zhao, Y. (2023) · Journal of Image and Graphics · doi:10.11834/jig.230036 · Cited by 2
Kothari, A., Gerdes, R., Billington, T., & Jaggannagari, R. (2023) · IEEE International Automated Vehicle Validation Conference (IAVVC) · doi:10.1109/iavvc57316.2023.10328073 · Cited by 1
“Tesla was one of the first OEM to switch over to centralized compute architecture because it reduced the number of semiconductor chips within the car [7].”
BackgroundACCO: Automated Causal CNN Scheduling Optimizer for Real-Time Edge Accelerators
Yin, J., Mei, L., Guntoro, A., & Verhelst, M. (2023) · IEEE International Conference on Computer Design: VLSI in Computers and Processors (ICCD) · doi:10.1109/iccd58817.2023.00065 · Cited by 1
“As the DSE hardware target, we select three well-known DNN accelerator architectures: Meta’s architecture [22], the Edge TPU [23], and the Tesla NPU [24].”
MethodMemScape: Sculpting Tiered Memory Management for Autonomous Vehicles
Seo, D., Chen, P.-X., Sung, C., Hoang, Q. A., Manzanares, A., & Dutt, N. D. (2024) · ACM Transactions on Embedded Computing Systems · doi:10.1145/3701621 · Cited by 1
“To evaluate the eicacy of MemScape in an autonomous vehicle scenario, we built a pipeline application conigured using the Chaufeur [27] benchmark suite to emulate the Tesla pipeline scenario [38].”
InfluentialEstimating the Impact of Soft Errors on AI-Based Perception in Automotive
Hegde, S. S., Selvaraj, D. C., Condia, J. E. R., Amati, N., Chiasserini, C. F., Deflorio, F. P., & Reorda, M. S. (2025) · IEEE Latin American Test Symposium (LATS) · doi:10.1109/lats65346.2025.10963959 · Cited by 1
“…AI applications often require high computational power and data-intensive operations, which need to be handled by advanced hardware accelerators with extended capabilities, such as embedded Neural Processing Units (NPUs), Tensor Processing Units (TPUs), or Graphics Processing Units (GPUs) [4]–[6].”
The Case for In-Memory Inferencing of Quantized CNNs at the EdgePreprint
Falcão, G., & Ferreira, J. D. (2022) · doi:10.36227/techrxiv.19249511.v1 · Cited by 1
Pérez, A. G. (2022) · Digital Repository (Polytechnic University of Cartagena) · doi:10.31428/10317/11206 · Cited by 3
Alcaide, S. (2023) · UPCommons institutional repository (Universitat Politècnica de Catalunya) · doi:10.5821/dissertation-2117-402844 · Cited by 3
The Outline of Deception: Physical Adversarial Attacks on Traffic Signs Using Edge PatchesPreprint
Jia, H., Hu, T., Li, H., Jin, L., Xin, C., Yao, Y., & Xiao, J. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2512.00765 · Cited by 1
Real-time and efficient control for autonomous racing
Li, N. J. (2023) · theses.fr (ABES) · doi:10.70675/e86ce678z3839z4d90z8b8ezda5eda81dd72 · Cited by 1
Architectures to Ensure the Functional Safety of Neural Network based Systems
Burel, S. (2022) · theses.fr (ABES) · doi:10.70675/08b0620bzfafaz401fz8e91z6856bff62a00 · Cited by 1
Case study of hardware implementation
Takano, S. (2021) · Elsevier eBooks · doi:10.1016/b978-0-12-818279-6.00017-7
A Transfer Learning Approach to Object Detection Acceleration for Embedded Applications
Vance, L. M. (2021) · Purdue e-Pubs (Purdue University System) · doi:10.25394/pgs.14798727.v1 · Cited by 1
Cui, S. (2021) · Illinois Digital Environment for Access to Learning and Scholarship (University of Illinois at Urbana-Champaign) · Cited by 3
(2021) · Elsevier eBooks · doi:10.1016/b978-0-12-818279-6.00025-6
Tang, D. (2022) · International Conference on Smart Systems and Inventive Technology (ICSSIT) · doi:10.1109/icssit53264.2022.9716445
(2021) · Elsevier eBooks · doi:10.1016/b978-0-12-818279-6.00005-0
The Case for In-Memory Inferencing of Quantized CNNs at the EdgePreprint
Falcão, G., & Ferreira, J. D. (2022) · doi:10.36227/techrxiv.19249511
Design of Lightweight Intelligent Vehicle System Based on Hybrid Depth Model
Yan, Z., Lan, B., Chen, S., Yu, S., Wang, X., Fang, Z., Chen, W., Wang, C., & Shi, X. (2021) · IEEE International Conference on Trust, Security and Privacy in Computing and Communications (TrustCom) · doi:10.1109/trustcom53373.2021.00180
A Many-ported and Shared Memory Architecture for High-Performance ADAS SoCsPreprint
Luan, H., Yao, Y., & Huang, C. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2209.05731
“Tesla’s full self-driving (FSD) SoC uses two instances of neural-network accelerator (NNA), where each NNA has a 32 MB of on-chip SRAM memory providing a high throughput data feed to enable and sustain a total of 72 TOPS processing power [2].”
InfluentialWomen in Circuits [Guest Editorial]
Gu, Q. J. (2022) · IEEE Solid-State Circuits Magazine · doi:10.1109/mssc.2022.3207334
A SoC Booting Acceleration Strategy Based on Self-Migration Technology
Man, H., Zhou, L., & Wang, S. (2022) · International Conference on Networking, Communications and Information Technology (NetCIT) · doi:10.1109/netcit57419.2022.00123
“Digital electronic devices such as smartphones, automatic drive equipments, Internet of Things (IoT) end nodes, servers and storages of data center contain one or more integrated circuit chips [1-4].”
BackgroundA Survey of Motion Planning Algorithms
Guo, Y., Ge, G., Xu, Q., & Yang, S. (2023) · International Conference on Mathematics and Machine Learning (ICMML) · doi:10.1145/3653724.3653730
Comparative analysis of silicon and steel technologies: Shaping societal evolution and prospects
Zhang, P., Yang, J., & Zhai, Y. (2024) · Applied and Computational Engineering · doi:10.54254/2755-2721/84/20240893
In-Field Fault Detection Framework for Edge Accelerator Using Autoencoder
Haripriya, R. S., & Tudu, J. (2024) · IEEE International Test Conference India (ITC India) · doi:10.1109/itcindia62949.2024.10651981
“Tesla’s Full Self Driving (FSD) performs the comparison of the results after executing the same AI task in two redundant hardware [9], [10].”
MethodThe Impact of Redundancy on Resource Utilization in Vehicular Clouds
Ghazizadeh, P., Narine, O., & Ghazizadeh, A. (2024) · IEEE Global Conference on Artificial Intelligence and Internet of Things (GCAIoT) · doi:10.1109/gcaiot63427.2024.10833587
“The computing power of these vehicles, which can be many times faster than most modern consoles, phones, and even desktop computers, often remains underutilized while they are parked [4], [5].”
RISC-V Neural Network Instruction Design and Simulation with Cache Scheduling via ROCC Interface
Dai, S., Zhou, Z., & Han, J. S. (2024) · IEEE International Conference on Solid-State and Integrated Circuit Technology · doi:10.1109/icsict62049.2024.10831182
“For example, in the Inception network, convolution operations constitute more than 98% of the computations [1].”
FusionFrame: A Fusion Dataflow Scheduling Framework for DNN Accelerators via Analytical Modeling
Zheng, L., Lan, H., Liu, X., Jiang, L., & Zhou, X. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-1551-3_21
Classifying paint colour using acoustic data from laser ablation
Grant‐Jacob, J. A., Zervas, M. N., & Mills, B. (2025) · Optical Materials Express · doi:10.1364/ome.558459
“These networks have been applied to various tasks, including image classification of bird songs [20], labelling cars [21], and analysing magnetic resonance imaging (MRI) scans [22].”
RoTA: Rotational Torus Accelerator for Wear Leveling of Neural Processing Elements
Lim, T., Kim, H., Park, J., Kim, B., & Song, W. J. (2025) · Design, Automation, and Test in Europe · doi:10.23919/date64628.2025.10992787
“…the RWL+RO scheme improves lifetime reliability by 1.69x on average for various DNN workloads [6], [7], [9], [18], [22], [24], [4], [12], [26] with only 0.3% design overhead and no performance degradation, compared to typical accelerator designs without wear-leveling [2], [19], [20], [23], [25].”
InfluentialTahoori, M. B., Becker, J., Henkel, J., Kunz, W., Schlichtmann, U., Sigl, G., Teich, J., & Wehn, N. (2025) · Design, Automation, and Test in Europe · doi:10.23919/date64628.2025.10992986
“For example, Tesla’s Full Self-Driving (FSD) chip incorporates functional safety through redundancy, with two independent FSD chip instances running separate operating systems [TAL20].”
SPEED: Scalable and Predictable EnhancEments for Data Handling in Autonomous Systems
Seo, D., Sung, C., Park, J., Chen, P.-X., Donyanavard, B., & Dutt, N. D. (2025) · International Symposium on Quality Electronic Design (ISQED) · doi:10.1109/isqed65160.2025.11014394
“However, the current rosbag architecture’s single-queue buffer system introduces significant performance degradation as the number of sensors and concurrent datastreams grow [3], limiting scalability to meet the increasing data throughput demands in modern autonomous systems.”
InfluentialMulti-DOF Fusion: A Flexible Fusion Strategy for Reducing Redundancy in CNN Workloads
Chen, Y., Zhou, Z., Dai, S., Duan, X., & Han, J. S. (2025) · ACM Great Lakes Symposium on VLSI (GLSVLSI) · doi:10.1145/3716368.3735173
Hoefer, J., Schmidt, P., Kiesa, H. T., Hoefer, S., Schewior, G., Engelke, D., Eickel, K.-H., Grantz, D., Harbaum, T., & Becker, J. (2025) · ACM Great Lakes Symposium on VLSI (GLSVLSI) · doi:10.1145/3716368.3735208
Ren, Y., Cui, Y., & Su, J. (2025) · International Conference on Neural Networks, Information and Communication Engineering (NNICE) · doi:10.1109/nnice64954.2025.11064723
“The V12 version launched by the Full Self-Driving system[6] of Tesla automobiles in the United States supports the execution of the entire process by AI and can directly generate vehicle control commands.”
3D reconstruction simulation method based on adaptive generalized orthogonal projection
Zhou, M., & Ye, D. (2025) · Engineering Research Express · doi:10.1088/2631-8695/adf947
TimelyNet: Adaptive Neural Architecture for Autonomous Driving with Dynamic Deadline
Chen, J., Le, D. V., Li, Y., Liu, Y., & Tan, R. (2025) · ACM Transactions on Embedded Computing Systems · doi:10.1145/3762652
“The computation capability of the NVIDIA Quadro RTX 8000 GPU is 16.3 tera floating-point operations per second ( TFLOPS ), which is equivalent to commercial autonomous vehicles computing platforms, such as Tesla FSD with 18.45 TFLOPS [21].”
InfluentialAdaptive orthogonal projection high speed multi-vision pose measurement based on CUDA
Zhou, M., & Dong, Y. (2025) · Measurement · doi:10.1016/j.measurement.2025.119565
Resource-Efficient Model Deployment for Enterprise AI
Chen, T., Yu, J., & Yin, H. (2025) · Springer eBooks · doi:10.1007/978-3-032-01940-0_1
Wang, X., Li, C., Li, Y., Sun, L., Xu, C., Hou, X., Wang, J., Guo, M., & Zhao, Y. (2025) · IEEE Transactions on Computers · doi:10.1109/tc.2025.3644184
“The capability of AES to alleviate human labor has spurred substantial investments from prominent industrial entities, including Nvidia [3] and Tesla [4], into technology advancements.”
LOCO: A Low-Cost SNU-Self-Resilient Latch Using an Output-Split C-ElementPreprint
Ma, R., Chen, X., Wen, X., Xu, H., Ye, S., Chuanjian, Z., & Wang, S. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2512.19292
CH2AI: Dynamic Checkpoint Scheduling of Anytime Neural Networks on Heterogeneous Units
Stammler, M., Sotiropoulos, G., Hoefer, J., Harbaum, T., Guha, K., Chakrabarti, A., & Becker, J. (2025) · IEEE International Symposium on Embedded Multicore/Many-core Systems-on-Chip (MCSoC) · doi:10.1109/mcsoc67473.2025.00036
“Lastly, some embedded execution platforms include multiple execution units, allowing parallel execution of multiple tasks [8].”
Fisheye Camera and Lidar Fusion for Autonomous Following Vehicles
Wan, J., & Wu, G. (2025) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc60802.2025.11423758
“And Tesla's FSD processes 2300 frames per second for road safety [9].”
Secure AI Runtime System Based on the M3 Platform
Hasler, M., Raitza, M., Pauls, F., & Fettweis, G. P. (2025) · Asilomar Conference on Signals, Systems and Computers (ACSSC) · doi:10.1109/ieeeconf67917.2025.11443879
“The accelerator features a systolic array of computation units that utilize scalable precision MAC units to by [3], does have a security module.”
A Mapping Strategy Optimization Framework for Systolic Array Accelerators
Yuan, X., Wang, C., Wu, Q., Yue, H., Xu, H., & Wei, X. (2026) · Lecture notes in computer science · doi:10.1007/978-981-95-8399-7_24
Nguyen, V. T. (2022) · theses.fr (ABES) · doi:10.70675/67883d2dz362cz4ca1z88f2z6a392d349bbb · Cited by 1
Contributions to safety assurance of autonomous trains
Chelouati, M. (2024) · theses.fr (ABES) · doi:10.70675/370d5308z351bz48baz8627zbce71f5b44ac · Cited by 1
Shi, Y., Ma, S., Li, T., Wang, B., Luo, S., Zhao, Y., Zhang, J., & Wu, L. (2026) · ACM Transactions on Design Automation of Electronic Systems · doi:10.1145/3801976
“Existing industrial practices for soft error handling are mainly based on double or TMR, as exemplified by Matsubara [12] and Tesla [13].”
Analysis of fault injection strategies targeting neural networks accelerators
Guinebert, I. (2024) · theses.fr (ABES) · doi:10.70675/94409412z5984z426cza50ez2d7556f89b49
HW-aware Neural Architecture Search Using Genetic Optimisation: In-Memory Compute vs NPU
Hendrickx, L., Colleman, S., & Goedemé, T. (2025) · International Conference on Electrical, Computer and Energy Technologies (ICECET) · doi:10.1109/icecet63943.2025.11472144
“We compare the Tesla Neural Processing Unit (NPU) [18] provided by the ZigZag framework with the analog and digital in-memory compute accelerators provided by the same framework [9].”
Reliability Assessment of Deep Neural Networks and Accelerators Across Design Stages
Ahmadilivani, M. H., Cherezova, N., Glaß, M., Guerrero-Balaguera, J.-D., Jenihhin, M., Kritikakou, A. S., Sierra, R. L., Poelhs, L. B., Raik, J., Roquet, L., et al. (2026) · IEEE Latin American Test Symposium (LATS) · doi:10.1109/lats70329.2026.11480274 · Cited by 1
“DPU-based-array cores are crucial elements commonly found in several platforms to train and deploy AI-based workloads, including Tensor cores (TCs) in GPUs [13], engines in NPUs [14], matrix core extensions in CPUs [15], [16], and on-chip accelerators in automotive devices [17].”
bashi, F. N. K., Khairullah, S. S., & Dawwd, S. A. (2026) · Digital · doi:10.3390/digital6030054
“An example is [78], which follows a lifecycle-driven safety framework, and automotive-grade designs like [81,82] that implement a fail-operational behavior with redundant system designs.”
InfluentialSparsity-Aware Fault Tolerance for Reliable CNN Inference on GPUs
Byeon, G., Lee, S., Kim, S., Kim, Y. W., Nair, P. J., & Hong, S. (2026) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2026.3715635
“T HE rapid development of Convolutional Neural Networks (CNNs)hasmadethemessentialforsafety-criticalapplica-tions, such as self-driving cars [1], [2], autonomous drones [3], and industrial robotics [4], [5].”
Evaluating tensor-related metrics for early fault detection in CNNs
Turco, V., Ruospo, A., Sánchez, E., & Reorda, M. S. (2026) · Applied Soft Computing · doi:10.1016/j.asoc.2026.116079
RHODES: Robust Optimization for Uncertainty-Aware Design of CO 2 -Efficient Computing Systems
Elgamal, M., Mahmoud, A., Wei, G.-Y., Brooks, D. M., & Hills, G. (2026) · Proceedings - International Symposium on Computer Architecture/Proceedings · doi:10.1109/isca66397.2026.00028
“This demonstrates that even when operational carbon dominates, uncertainties in embodied carbon parameters can still shift the robust design.”
Metadata
- Authors
- Talpes, E., Sarma, D. D., Venkataramanan, G., Bannon, P., McGee, B., Floering, B., Jalote, A., Hsiong, C., Arora, S., Gorti, A., & Sachdev, G. S.
- Affiliations
- Tesla (United States) (US)
- Venue
- IEEE Micro
- Published
- 2020-03-01
- Volume
- 40
- Issue
- 2
- Pages
- 25-35
- ISSN
- 0272-1732 (ISSN-L)
- 1937-4143
- OpenAlex
- W3007788310
- Cited by
- 193
- References
- 4
- Citations by year
- 2026: 18 · 2025: 51 · 2024: 31 · 2023: 32 · 2022: 29 · 2021: 22 · 2020: 10
- FWCI
- 8.60
- Type
- journal article
- Language
- en
- Open access
- closed
- Licence
- Locations
- IEEE Micro (publishedVersion)
- Topics
- Physical Sciences › Engineering › Automotive Engineering › Autonomous Vehicle Technology and Safety
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Hardware and Architecture › Embedded Systems Design Techniques
- Keywords
- system-on-chip, neural network accelerators, graphics processing units, image sensor interface
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.