Reproduce “Memory-Immersed Collaborative Digitization for Area-Efficient Compute-in-Memory Deep Learning” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Shamma Nasrin, Maeesha Binte Hashem, Nastaran Darabi, Benjamin Parpillon, Farah Fahim, Wilfred Gomes, Amit Ranjan Trivedi

The authors demonstrate a memory-immersed analog-to-digital converter that uses a neighboring compute-in-memory array’s bit-line capacitance to generate conversion references, avoiding a separate capacitive converter. Their 65-nanometer test chip supports successive-approximation, Flash, and hybrid operation. At 10 MHz, the 5-bit design uses 207.8 μm² and 74.23 pJ, versus 5,235.20 μm² and 105 pJ for a 40-nanometer SAR ADC; these are converter-level comparisons, not system-level results.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.