Kushnir et al.
Agentic copilot accelerates ML accelerator co-design by grounding insights in fresh simulation data
Coco deploys a four-layer platform that automates experiment setup, simulation sweeping, and analysis, with early evidence of halving time-to-insight for TPU architects.
Big Tech
Samuel Kushnir · Kavya Sreedhar · Yeshwanth Reddy Pogula · Amir Yazdanbakhsh · Narges Shahidi · Ming Liu · +5 more
Google · Google DeepMind · MIT
Research Digest··2 min read
The authors present Coco, an agentic platform for hardware-software co-design of ML accelerators, deployed with TPU architects.
Why this paper
From Google and 2 others
In one line
Coco automates simulation experiments and data analysis to accelerate hardware-software co-design for ML accelerators.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§