HBM Shifts Testing Left To Preserve AI Chip Yield


Key Takeaways: A high-yield, known-good stack requires multiple test insertions. Known good stack testing poses challenges for power delivery and thermal management. The shift to HBM4 and HBM5 will increase the pressure for shift-left test flows. Taller high-bandwidth memory (HBM) stacks and tighter TSV pitch are impacting AI module yields. The solution is to push test furth... » read more

Replacing GPU Compute Dies With PNM-Enabled HBM Cubes For Long-Context Decode Attention (UCSD, Columbia, Yonsei U., NVIDIA, Samsung)


A new technical paper, "AMMA: A Multi-Chiplet Memory-Centric Architecture for Low-Latency 1M Context Attention Serving," was published by researchers at UC San Diego, Columbia University, Yonsei University, NVIDIA, and Samsung. Abstract "All current LLM serving systems place the GPU at the center, from production-level attention-FFN disaggregation to NVIDIA's Rubin GPU-LPU heterogeneous p... » read more

When Semiconductor Materials Misbehave


Key Takeaways Material behavior in production depends on the process context that no development environment can fully replicate. In advanced packaging, the interactions that cross domain boundaries are increasingly where failures originate. The most accurate materials data is also the most commercially sensitive, leaving simulation models calibrated against generic inputs rather tha... » read more

TSV Complexity Leads To Manufacturing Bottleneck


Key Takeaways: Through-silicon vias are the biggest enabler of 3D chip stacking and chip-to-PCB connections through silicon interposers. The AI boom is causing HBM and advanced assembly shortages, straining the supply chain. Optimization around etch, fill and reveal help reduce TSV cost. Through-silicon vias (TSVs) provide essential interconnects between DRAM dies inside hig... » read more

Panel-Level Packaging’s Second Wave Meets Engineering Reality


Key Takeaways Panel-level packaging is arriving not because the engineering is ready, but because wafer-level economics are breaking down. Glass improves the warpage and dimensional stability problems of organic substrates but introduces a different class of failure modes that require materials solutions, not process adjustments. The central challenges of panel-level processing are m... » read more

PDN Challenges In DRAM-Based Compute-In-Memory Systems (UT Austin)


A new technical paper, "A comparative study on power delivery aspects of compute-in/near-memory approaches using DRAM," was published by researchers at UT Austin. Abstract "Compute-in-memory (PIM) mitigates the memory wall by performing computation within memory, reducing data movement and improving energy efficiency. DRAM-based PIM is particularly attractive due to its high density, matu... » read more

Early HBM4 Validation Points The Way For Next Generation AI And HPC Systems


As AI and high‑performance computing systems continue to scale, memory bandwidth has emerged as a primary system‑level constraint. Larger models, higher compute density, and increasingly complex multi‑die designs are driving the need for memory interfaces that can deliver extreme bandwidth while operating within tight power and signal‑integrity margins. High‑Bandwidth Memory (HBM) has... » read more

AI Accelerators Usher In New Era For IC Test


Key Takeaways The parallelism in AI accelerators enables low latency but complicates failure isolation. HBM can account for 50% of package cost, so known-good stack assurance is critical. DFT and test cooperate to solve final test, singulated die test, SLT, and in-system test for data centers. AI accelerators are used for everything from training large language models to mak... » read more

HBM4E Raises The Bar For AI Memory Bandwidth


The pace of AI innovation continues to expose a painful reality. Compute keeps scaling, but memory bandwidth remains one of the hardest bottlenecks to remove. As AI models grow larger and more complex, feeding data fast enough into accelerators has become just as critical as raw compute capability. High Bandwidth Memory (HBM) has been central to solving this challenge, and the next step in that... » read more

RPU: A Chiplet-Based Architecture To Address The Challenges of the Modern Memory Wall (Harvard University)


Researchers from Harvard University have released “RPU -- A Reasoning Processing Unit”. Abstract “Large language model (LLM) inference performance is increasingly bottlenecked by the memory wall. While GPUs continue to scale raw compute throughput, they struggle to deliver scalable performance for memory bandwidth bound workloads. This challenge is amplified by emerging reasonin... » read more

← Older posts Newer posts →