TT-Edge: A Hardware-Software Co-Design for Energy-Efficient Tensor-Train Decomposition on Edge AI
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866908660204044288 |
|---|---|
| author | Kwak, Hyunseok Lee, Kyeongwon Min, Kyeongpil Jung, Chaebin Lee, Woojoo |
| author_facet | Kwak, Hyunseok Lee, Kyeongwon Min, Kyeongpil Jung, Chaebin Lee, Woojoo |
| contents | The growing demands of distributed learning on resource constrained edge devices underscore the importance of efficient on device model compression. Tensor Train Decomposition (TTD) offers high compression ratios with minimal accuracy loss, yet repeated singular value decompositions (SVDs) and matrix multiplications can impose significant latency and energy costs on low power processors. In this work, we present TT-Edge, a hardware software co designed framework aimed at overcoming these challenges. By splitting SVD into two phases--bidiagonalization and diagonalization--TT-Edge offloads the most compute intensive tasks to a specialized TTD Engine. This engine integrates tightly with an existing GEMM accelerator, thereby curtailing the frequent matrix vector transfers that often undermine system performance and energy efficiency. Implemented on a RISC-V-based edge AI processor, TT-Edge achieves a 1.7x speedup compared to a GEMM only baseline when compressing a ResNet 32 model via TTD, while reducing overall energy usage by 40.2 percent. These gains come with only a 4 percent increase in total power and minimal hardware overhead, enabled by a lightweight design that reuses GEMM resources and employs a shared floating point unit. Our experimental results on both FPGA prototypes and post-synthesis power analysis at 45 nm demonstrate that TT-Edge effectively addresses the latency and energy bottlenecks of TTD based compression in edge environments. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_13738 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | TT-Edge: A Hardware-Software Co-Design for Energy-Efficient Tensor-Train Decomposition on Edge AI Kwak, Hyunseok Lee, Kyeongwon Min, Kyeongpil Jung, Chaebin Lee, Woojoo Distributed, Parallel, and Cluster Computing Hardware Architecture The growing demands of distributed learning on resource constrained edge devices underscore the importance of efficient on device model compression. Tensor Train Decomposition (TTD) offers high compression ratios with minimal accuracy loss, yet repeated singular value decompositions (SVDs) and matrix multiplications can impose significant latency and energy costs on low power processors. In this work, we present TT-Edge, a hardware software co designed framework aimed at overcoming these challenges. By splitting SVD into two phases--bidiagonalization and diagonalization--TT-Edge offloads the most compute intensive tasks to a specialized TTD Engine. This engine integrates tightly with an existing GEMM accelerator, thereby curtailing the frequent matrix vector transfers that often undermine system performance and energy efficiency. Implemented on a RISC-V-based edge AI processor, TT-Edge achieves a 1.7x speedup compared to a GEMM only baseline when compressing a ResNet 32 model via TTD, while reducing overall energy usage by 40.2 percent. These gains come with only a 4 percent increase in total power and minimal hardware overhead, enabled by a lightweight design that reuses GEMM resources and employs a shared floating point unit. Our experimental results on both FPGA prototypes and post-synthesis power analysis at 45 nm demonstrate that TT-Edge effectively addresses the latency and energy bottlenecks of TTD based compression in edge environments. |
| title | TT-Edge: A Hardware-Software Co-Design for Energy-Efficient Tensor-Train Decomposition on Edge AI |
| topic | Distributed, Parallel, and Cluster Computing Hardware Architecture |
| url | https://arxiv.org/abs/2511.13738 |