Session

DescriptionThis session highlights cutting-edge research on advancing next-gen computing architectures, focusing on processors, GPUs, and AI-enhanced systems. Papers explore incremental partitioning techniques on GPUs, fine-grain instruction analysis in out-of-order superscalar processors, and multi-core solutions with AI extensions for multimodal LLMs. Additional topics include efficient hardware resource sharing in GPU sub-cores, a multi-fidelity optimization framework for microarchitecture design space exploration (e.g., RISCV), new GPU kernel fusion techniques, and dataflow optimizations for tensor accelerators. These works showcase innovative strategies that push the boundaries of performance, scalability, and efficiency in modern computing systems.