Artifact Evaluation Repository for the FPGA'26 Paper: TeLLMe: An Efficient End-to-End Ternary LLM Prefill and Decode Accelerator with Table-Lookup Matmul on Edge FPGAs
Year: 2026
Type: software
Abstract: Please download and extract the provided .7z archives before running the evaluation. If you encounter any issues with direct downloads from Zenodo, the same .7z package is also available through the following Google_Drive.txt. Make sure that all files are fully extracted and placed in the expected directory structure before proceeding with the evaluation steps. If you are using this code in your own project, please cite the following work. The citation for the later conference proceeding will be added once it becomes available. @article{qiao2025tellme, title={TeLLMe: An Energy-Efficient Ternary LLM Accelerator for Prefilling and Decoding on Edge FPGAs}, author={Qiao, Ye and Chen, Zhiheng and Zhang, Yifan and Wang, Yian and Huang, Sitao}, journal={arXiv preprint arXiv:2504.16266}, year={2025}, doi={10.48550/arXiv.2504.16266}} @article{qiao2025tellmev2, title={TeLLMe v2: An Efficient End-to-End Ternary LLM Prefill and Decode Accelerator with Table-Lookup Matmul on Edge FPGAs}, author={Qiao, Ye and Chen, Zhiheng and Zhang, Yifan and Wang, Yian and Huang, Sitao}, journal={arXiv preprint arXiv:2510.15926}, year={2025}, doi={10.48550/arXiv.2510.15926}}
Institution: University of California, Irvine
Language: English
Subfield: Hardware and Architecture
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7124977832