Research
A paper is listed when one of its authors is affiliated with the lab. Data: OpenAlex.
Cambricon-FlexLLM: A Flexible Chiplet-Based Hybrid Architecture for On-Device 70B LLM InferenceCambricon Technologies · ACM Transactions on Architecture and Code OptimizationAug 29Cambricon-QM: A Hybrid Architecture for Microscaling Format TrainingCambricon Technologies · IEEE Transactions on ComputersJul 8QiMeng-Tensify: Scaling Up Tensor Computation Optimization via Architecture-Aware LLM-Guided MCTSCambricon Technologies · ACM/IEEE International Symposium on Computer Architecture (ISCA)Jun 1Efficient Diffusion Planning with Temporal DiffusionCambricon Technologies · Proceedings of the AAAI Conference on Artificial IntelligenceMar 14Cambricon-GS: An Accelerator for 3D Gaussian Splatting Training With Gaussian-Pixel Hybrid ParallelismCambricon Technologies · Proceedings - International Symposium on High-Performance Computer Architecture/ProceedingsJan 31FlashAttention-T: Towards Fully Tensorized Attention by Exploiting Tensor-Vector ParallelismCambricon Technologies · ACM SIGPLAN Symposium on Principles and Practice of Parallel ProgrammingJan 28QiMeng-CPU-v2: Automated Superscalar Processor Design by Learning Data DependenciesCambricon Technologies · IEEE Transactions on Computer-Aided Design of Integrated Circuits and SystemsJan 1Efficient Diffusion Planning with Temporal DiffusionCambricon Technologies · arXiv (Cornell University)Nov 26, 2025TensorFHE+: Fully Homomorphic Encryption Acceleration Based on Linear AlgebraCambricon Technologies · IEEE Transactions on Computers1 citationNov 5, 2025DASA: Distribution-Aware Sparse Attention for Accelerating Diffusion TransformerCambricon Technologies · IEEE Transactions on Computer-Aided Design of Integrated Circuits and SystemsOct 31, 20253D Non-Uniform Fast Fourier Transform Program OptimizationCambricon Technologies · Applied SciencesSep 30, 2025Automated Superscalar Processor Design by Learning Data DependenciesCambricon Technologies · International Joint Conference on Artificial Intelligence1 citationSep 1, 2025Swift: High Parallelism Program Generation of Tensor Operators for Accelerating Deep Learning InferenceCambricon Technologies · ACM Transactions on Architecture and Code Optimization1 citationAug 25, 2025Cambricon-SR: An Accelerator for Neural Scene Representation with Sparse Encoding TableCambricon Technologies · Proceedings - International Symposium on Computer Architecture/Proceedings1 citationJun 20, 2025VariPar: Variation-Aware Workload Partitioning in Chiplet-Based DNN AcceleratorsCambricon Technologies · IEEE Transactions on Computer-Aided Design of Integrated Circuits and SystemsMay 9, 2025Mosaic: Exploiting Instruction-Level Parallelism on Deep Learning Accelerators with iTex TessellationCambricon Technologies · International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS)2 citationsMar 27, 2025SaaP: Rearchitect SoC-as-a-Processor to Orchestrate Hardware HeterogeneityCambricon Technologies · IEEE Transactions on Computer-Aided Design of Integrated Circuits and SystemsMar 21, 2025Cambricon-DG: An Accelerator for Redundant-Free Dynamic Graph Neural Networks Based on Nonlinear IsolationCambricon Technologies · Proceedings - International Symposium on High-Performance Computer Architecture/Proceedings2 citationsMar 1, 2025AI Computing Systems for Large Language Models TrainingCambricon Technologies · Journal of Computer Science and Technology17 citationsJan 1, 2025