Balancing Computation Loads and Optimizing Input Vector Loading in LSTM Accelerators
| العنوان: | Balancing Computation Loads and Optimizing Input Vector Loading in LSTM Accelerators |
|---|---|
| المؤلفون: | Daehyun Ahn, Jae-Joon Kim, Wooseok Yi, Jaeha Kung, Junki Park |
| المصدر: | IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems. 39:1889-1901 |
| بيانات النشر: | Institute of Electrical and Electronics Engineers (IEEE), 2020. |
| سنة النشر: | 2020 |
| مصطلحات موضوعية: | Hardware architecture, Artificial neural network, Computer science, Computation, Pipeline (computing), 02 engineering and technology, Parallel computing, Computer Graphics and Computer-Aided Design, Column (database), 020202 computer hardware & architecture, Matrix (mathematics), 0202 electrical engineering, electronic engineering, information engineering, Hardware acceleration, Pruning (decision trees), Electrical and Electronic Engineering, Software, Sparse matrix |
| الوصف: | The long short-term memory (LSTM) is a widely used neural network model for dealing with time-varying data. To reduce the memory requirement, pruning is often applied to the weight matrix of the LSTM, which makes the matrix sparse. In this paper, we present a new sparse matrix format, named rearranged compressed sparse column (RCSC), to maximize the inference speed of the LSTM hardware accelerator. The RCSC format speeds up the inference by: 1) evenly distributing the computation loads to processing elements (PEs) and 2) reducing the input vector load miss within the local buffer. We also propose a hardware architecture adopting hierarchical input buffer to further reduce the pipeline stalls which cannot be handled by the RCSC format alone. The simulation results for various datasets show that combined use of the RSCS format and the proposed hardware requires $2\times $ smaller inference runtime on average compared to the previous work. |
| تدمد: | 1937-4151 0278-0070 |
| URL الوصول: | https://explore.openaire.eu/search/publication?articleId=doi_________::706894e1c30e1eebe3498efa465fe881 https://doi.org/10.1109/tcad.2019.2926482 |
| حقوق: | CLOSED |
| رقم الأكسشن: | edsair.doi...........706894e1c30e1eebe3498efa465fe881 |
| قاعدة البيانات: | OpenAIRE |
| تدمد: | 19374151 02780070 |
|---|