stencil 计算在 Tenstorrent Wormhole 上的实现
Stencil Computations on Tenstorrent Wormhole
AI总结:
本文研究了将二维五点 stencil 计算映射到 Tenstorrent Wormhole 上的效率,提出两种异构实现方法,并分析了平台的限制与改进方向。
摘要:
随着人工智能加速器投资的增长及其在超级计算设施中的部署扩展,理解这些架构是否能高效支持传统科学内核对高性能计算的未来至关重要。我们研究了将二维五点 stencil 计算映射到 Tenstorrent Wormhole,一种 RISC-V AI 数据流加速器。我们开发了两种异构实现:Axpy,将 stencil 分解为元素级子矩阵操作,以及 MatMul,将其重新表述为矩阵乘法。尽管 CPU 基线仍保持 3 倍更快的端到端性能,但分析显示孤立的 Wormhole 内核与 CPU 执行竞争,差距由 PCIe 传输、设备初始化和主机侧预处理驱动。尽管运行时间较慢,Axpy 在大输入情况下比 CPU 基线消耗更少的能量。通过详细分析和理论分析,我们识别了当前平台的关键架构和软件限制,并概述了可能使 AI 加速器在 HPC 工作负载中具有竞争力的硬件和软件改进方向。
英文摘要:
As investment in AI-focused accelerators grows and their deployment in supercomputing facilities expands, understanding whether these architectures can efficiently support traditional scientific kernels is critical for the future of High-Performance Computing. We investigate the mapping of 2D 5-point stencil computations onto the Tenstorrent Wormhole, a RISC-V AI dataflow accelerator. We develop two heterogeneous implementations: Axpy, which decomposes the stencil into element-wise submatrix operations, and MatMul, which reformulates it as a matrix multiplication. While the CPU baseline remains 3x faster end-to-end, profiling reveals that the isolated Wormhole kernel is competitive with CPU execution, with the gap driven by PCIe transfers, device initialization, and host-side preprocessing. Despite slower runtime, Axpy achieves lower energy consumption than the CPU baseline for large inputs. Through detailed profiling and theoretical analysis, we identify key architectural and software limitations of the current platform and outline concrete hardware and software directions that could make AI accelerators competitive for HPC workloads.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。