News · 新闻动态

🎉 Congratulations to Dr. Fangxin Liu on receiving the 2026 ACM SIGHPC China Rising Star Award — Award results announcement [Announcement]

🎉 Fangxin Liu: JQK: Joint Quantization and Kernel Co-Optimization for Hardware-Efficient Neural Network Inference — accepted by IEEE TCAD 2026 CCF-A — Huawei collaboration; joint quantization / kernel co-optimization [Paper]

🎉 Junjie Wang, Can Cui (co-first): HiLoViT: Dataflow-Aware Vision Transformer Compression via Hilbert Guided Spatial Coding — accepted by ICCD 2026 CCF-B — Special Session

Feature: HyperParallel at PyTorch Conference China 2026, with HyperOffload shown in the architecture — Official PyTorch recap, authoritative coverage, and conference photos [Feature]

🎉 Jialin Zhan: DELTA: Decoupling Latent Heterogeneity in Asymmetric Low-Rank Compression — accepted by EMNLP 2026 CCF-B

🎉 Ziyan Gan: PCoMoE: Shifting MoE Inference from Monolithic Expert Selection to Fine-Grained Path Composition — accepted by EMNLP 2026 CCF-B

🎉 Yiwei Hu: Seer: Efficient KV Cache Management for LLM Acceleration via Mitigating Semantic Search Inefficiency — accepted by EMNLP 2026 CCF-B

🎉 Xuwen Zhou: Calibrated Speculative Decoding: Frequency-Guided Candidate Selection for Efficient Inference — accepted by ACL 2026 CCF-A

🎉 Ning Yang: NICE: Deep Neural Network Acceleration via Hardware-Friendly Index Assisted Compression — accepted by ACM TACO 2026 CCF-A

🎉 Haomin Li: GEMM-GS: Accelerating 3D Gaussian Splatting on Tensor Cores with GEMM-Compatible Blending — accepted by DAC 2026 CCF-A

🎉 Yilong Zhao: PUSHtap: PIM-based In-Memory HTAP with Unified Data Storage Format — accepted by ASPLOS 2025 CCF-A

🎉 Haomin Li, Fangxin Liu (co-first): Look Once, Compute Less: A Spatio-Temporal Redundancy-Aware CIM Accelerator for Adaptive BEV Representation — accepted by IEEE TCAD 2026 CCF-A

🎉 Fangxin Liu, Xin Ju (co-first): SqzAct: Taming Activation Outliers for Efficient 4-bit LLM Inference via Block-Level Squeezing — accepted by IEEE TCAD 2026 CCF-A

🎉 Ning Yang, Chenyang Guan (co-first): LOAD: Enabling Low-Bit Group Quantization for Vision Transformers via Adaptive Data Formats — accepted by ACM TACO 2026 CCF-A

🎉 Junjie Wang, Can Cui (co-first): DNA-ViT: Developmental Neural Archiving for Storage-Efficient Vision Transformers — accepted by ACM MM 2026 CCF-A — Main Track [Paper]

🎉 Fangxin Liu: MOSAIC: Exploiting Structured Tolerance for Adaptive LLM Inference Mapping in Heterogeneous PIM Accelerators — accepted by MICRO 2026 CCF-A

🎉 Zongwu Wang: ULTRA: Bridging the VQ Gap in Anisotropic LLM Quantization via a Unified LUT-Based Transformer Architecture — accepted by MICRO 2026 CCF-A

🎉 Yilong Zhao: COMET: A Cooperative Scheduling Framework for Concurrent PIM/CPU Execution on Mobile Devices — accepted by ISCA 2026 CCF-A — Best Paper Award Finalist (one of five finalists)

🎉 Jingkui Yang: Harmonia: A Unified Hierarchical Scheduling Framework for Sparse Matrix Multiplication — accepted by ISCA 2026 CCF-A

🎉 Fangxin Liu: STEP: Adaptive Spatio-Temporal Expert Prefetching for Low-Latency and Memory-Efficient MoE Inference — accepted by ISCA 2026 CCF-A