IC-260DeepGate2 achieves an average NDCG@3 of 0.334 and top-10% commonality of 0.226 on QOR prediction across 10 circuit designs

Haoyuan WU, Haisheng Zheng, Yuan Pu, Bei Yu

SourceCircuit Representation Learning with Masked Gate Modeling and Verilog-AIG Alignment

The paper evaluates DeepGate2's ability to predict the quality of logic synthesis optimization sequences. Using circuit embeddings extracted from AIGs, DeepGate2 ranks 1500 optimization sequences per design. Across 10 evaluation designs (bc0, apex1, div, k2, i10, mainpla, or1200 cpu, picorv32, rocket, sqrt), DeepGate2 achieves an average NDCG@3 of 0.334, NDCG@5 of 0.324, top-3% commonality of 0.116, top-5% of 0.157, and top-10% of 0.226.

Evidence
correlational
Key metric
average NDCG@3 = 0.334, NDCG@5 = 0.324, top-3% = 0.116, top-5% = 0.157, top-10% = 0.226
Model
DeepGate2
Datasets
ITC'99 [train], PicoRV32 [eval]
Methods
ABC [eval]
Related work
DeepGate2 [compared-to]
Related findings
IC-261, IC-262, IC-263
Extraction
automatic-extraction