China's LineShine Supercomputer Shows HBM Moving Beyond GPU Clusters

China's new LineShine supercomputer is a notable high-performance computing memory design. TechPowerUp reports that the system uses CPU nodes with HBM attached for high-bandwidth access, plus a much larger pool of system DDR memory per processor. For RamTrend, the point is that HBM demand is not limited to GPU-centric AI clusters. Scientific HPC systems can also use HBM as a performance tier next to conventional system memory. The payload does not identify the memory supplier or procurement volume, but the configuration underscores how high-end compute systems are widening the use cases for bandwidth-rich memory.