Join us virtually at the TSMC Symposium & OIP Ecosystem Forum and watch Frank Ferro, Sr. Director of Product Marketing at Rambus, give a presentation titled, “Memory Systems for AI Applications Using TSMC Advanced Process Nodes.” See abstract details below.
Abstract: Exponential data growth is driving accelerating performance requirements in data centers and networking infrastructure. The rapid rise of AI contributes greatly to the enormous growth in data which the industry has responded to with new hardware platforms being developed for both Training and Inference applications in TSMC’s advanced process nodes. But even these new designs are increasingly limited by memory bandwidth, thus underutilizing the processing power available. To remove this bottleneck, TSMC customers need new memory subsystems that offer 4x or greater the performance of existing solutions. GDDR6 and HBM2E have emerged specifically to address the AI memory bottleneck. This talk will discuss the new memory systems available in TSMC’s advanced process nodes, along with the benefits and trade-offs of different memory systems. Rambus will also discuss the need for close collaboration between SMC and IP companies to develop these advanced PHY solutions. To learn more, visit the event page here.
For the full agenda, click here.