TensorCache: Reconstructing Memory Architecture With SRAM-Based In-Cache Computing for Efficient Tensor Computations in GPGPUs.
Yicong ZhangMingyu WangYangzhan MaiZhiyi YuPublished in: IEEE Trans. Very Large Scale Integr. Syst. (2023)
Keyphrases
- memory hierarchy
- efficient computation
- memory access
- multithreading
- memory subsystem
- higher order
- memory management
- cache conscious
- highly efficient
- management system
- random access memory
- garbage collection
- limited memory
- main memory
- high order
- memory space
- block size
- computing power
- real time
- dynamic random access memory