Maia 200 is introduced, an advanced AI accelerator delivering high performance-10 145 Tflop/s FP4 and 5072 Tflop/s FP8 within a 750W TDP and 7 TB/s HBM bandwidth, and a taxonomy of data management, inspired by Flynn's classification, highlights how SDLA addresses challenges in modern AI computing.
Abstract
We introduce Maia 200, an advanced AI accelerator delivering high performance-10 145 Tflop/s FP4 and 5072 Tflop/s FP8 within a 750W TDP and 7 TB/s HBM bandwidth. Maia exemplifies a new class of Software Defined Locally Accessed Dataflow Architectures (SDLA), which explicitly program dataflow engines to orchestrate highly specialized memories and data movement engines. This approach shifts the focus from today's thread-centric to data-movement-centric architecture, improving efficiency and scalability. Our taxonomy of data management, inspired by Flynn's classification, highlights how SDLA addresses challenges in modern AI computing. Maia 200 achieves significant cost and energy savings while supporting massive parallelism for AI inference workloads, making it a compelling solution for next-generation high-performance computing systems.
This survey classifies industrial AI accelerators into four architectural categories and compares their compute and memory organizations, and examines how node-, rack-, and pod-scale interconnects support collective communication, and traces architectural evolution across accelerator generations.
CMD, an integrated CGRA framework featuring cluster-based distributed memory design with a co-designed compilation toolchain, is proposed, which includes a novel memory-aware mapper and a design space exploration mechanism that identifies the optimal memory architecture design for specific kernels.
Shang-Kun Li, Cheng Tan, Ze-Yu Li et al.· 0 citations
FlashGPU-sim is presented, an open-source, execution-driven, cycle-accurate GPU simulator for modern AI workloads that faithfully models modern hardware features such as asynchronous data movement, fine-grained synchronization, tensor-core execution, and distributed shared memory.
Si-Ying Yu, Yi-Xun Hong, Guo-Zhi Qiu et al.· 0 citations
Preliminary evidence that resolving context-dependent control through centralized generation can reduce software overhead and enable simpler edge hardware for efficient, adaptive computation is provided.
Peng-Ju Ren, Ge-Lin Fu, Sheng-Huan Liu et al.· AI Plus· 0 citations
The study demonstrates that mid-range GPU infrastructure can effectively support modern AI applications, and provides practical deployment guidelines and demonstrate that mid-scale HPC systems can serve as a foundation for scalable and cost-effective AI infrastructure.
A. Temirbayev, Azati Yakufujiang· Вестник Алматинского универс...· 0 citations
(English) High-performance computing (HPC) platforms are evolving towards increasingly complex architectures: many-core CPUs with multi-level NUMA hierarchies, heterogeneity with multiple classes of accelerators and higher-capacity interconnects. The increasing complexity and variety of resources in these machines make...
David Álvarez Robert· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.