Open Benchmark for AI Thermal Models in 2.5D and 3D ICs
Researchers from UTS, ShanghaiTech, and TU Munich published IC-ThermBench, an open progressive benchmark for generalizable 2.5D/3D-IC thermal learning.
According to Semiconductor Engineering, researchers from the University of Technology Sydney, ShanghaiTech University, and Technical University of Munich published a technical paper titled 'IC-ThermBench: An Open, Progressive Benchmark for Generalizable 2.5D/3D-IC Thermal Learning.'
The mechanism is the benchmark: IC-ThermBench combines established 3D-IC steady-state, transient, and industrial package tasks with a new 50,000-sample 2.5D chiplet dataset. The 'progressive' aspect means the benchmark scales in difficulty.
What changed is the thermal learning evaluation: rather than ad-hoc benchmarks, IC-ThermBench is an open, progressive benchmark that tests generalizability across 2.5D and 3D IC thermal tasks. The 50,000-sample 2.5D chiplet dataset is the new contribution.
What is unconfirmed is whether the benchmark is widely adopted or remains a research tool. The paper is a research contribution, so the benchmark's value depends on adoption.
What to watch is whether IC-ThermBench becomes the standard for evaluating AI thermal models in 2.5D/3D ICs, and whether the 2.5D chiplet dataset is extended. The open, progressive design is the key value.
Desk take
IC-ThermBench is an open, progressive benchmark for AI thermal models in 2.5D/3D ICs. The 50,000-sample 2.5D chiplet dataset is the new contribution. Watch for adoption as the standard for thermal model evaluation.
A standard benchmark for AI thermal models in 2.5D/3D ICs matters for the evaluation of thermal learning in advanced packaging.
Source dispatch
Researchers at the University of Technology Sydney, ShanghaiTech University, and Technical University of Munich published a technical paper titled “IC-ThermBench: An Open, Progressive Benchmark for Generalizable 2.5D/3D-IC Thermal Learning.” Abstract Excerpt: “We introduce IC-ThermBench, an open and progressive benchmark that combines established 3D-IC steady-state, transient, and industrial package tasks with a new 50,000-sample 2.5D chiplet... » read more The post Open Benchmark Evaluates AI Thermal Models for 2.5D and 3D ICs (UTS, TU Munich, ShanghaiTech) appeared first on Semiconductor Engineering .
