Proteus: Achieving High-Performance Processing-Using-DRAM via Dynamic Precision Bit-Serial Arithmetic
Proteus: Achieving High-Performance Processing-Using-DRAM via Dynamic Precision Bit-Serial Arithmetic
Geraldo F. Oliveira,Mayank Kabra,8 Authors,Onur Mutlu
2025 · DOI: 10.48550/arXiv.2501.17466
arXiv.org · 5 Citations
TLDR
Proteus parallelizes the execution of independent primitives in a PUD operation by leveraging DRAM’s internal parallelism and reduces the bit-precision for PUD operations by leveraging narrow values (i.e., values with many leading zeros).
