Casper: Accelerating Stencil Computation using Near-cache Processing

Abstract

Stencil computation is one of the most used kernels in a wide variety of scientific applications, ranging from large-scale weather prediction to solving partial differential equations. Stencil computations are characterized by three unique properties: (1) low arithmetic intensity, (2) limited temporal data reuse, and (3) regular and predictable data access pattern. As a result, stencil computations are typically bandwidth-bound workloads, which only experience limited benefits from the deep cache hierarchy of modern CPUs.In this work, we propose Casper, a near-cache accelerator consisting of specialized stencil compute units connected to the lastlevel cache (LLC) of a traditional CPU. Casper is based on two key ideas: (1) avoiding the cost of moving rarely reused data through the cache hierarchy, and (2) exploiting the regularity of the data accesses and the inherent parallelism of the stencil computation to increase the overall performance. With minimal changes in LLC address decoding logic and data placement, Casper performs stencil computations at the peak bandwidth of the LLC. We show that, by tightly coupling lightweight stencil compute units near to LLC, Casper improves performance of stencil kernels by 1.65× on average, while reducing the energy consumption by 35% compared to a commercial high-performance multi-core processor. Moreover, Casper provides a 37× improvement in performance-per-area compared to a state-of-the-art GPU.

Methodologies, Workloads, and Tools for Processing-in-Memory: Enabling the Adoption of Data-Centric Architectures

“…PrIM is opensource and publicly available at [168]. Unlike these prior works, DAMOV is applicable to and can be used to study other PIM architectures than processing-in/-near DRAM, including processing-in/-near cache [68,[93][94][95][169][170][171], processing-in/-near storage [40,[172][173][174][175][176][177][178][179][180][181], and processing-in/-near emerging NVMs [81,82,90,91,100,182,183]. This is possible since DAMOV's methodology and benchmarks are mainly concerned with broadly characterizing data movement bottlenecks in an application, independent of the underlying PIM architecture.…”

Heterogeneous Data-Centric Architectures for Modern Data-Intensive Applications: Case Studies in Machine Learning and Databases

“…Naively employing PIM to accelerate data-intensive workloads can lead to sub-optimal performance due to the many design constraints PIM substrates impose (e.g., limited area and power budget available inside 3D-stacked memories [6] or manufacturing limitations of combining memory and logic elements [6,13]). Therefore, many recent works co-design specialized PIM accelerators and algorithms to improve performance and reduce the energy consumption of (i) applications from various application domains, such as graph processing , machine learning [1,, bioinformatics , high-performance computing [95,[101][102][103][104][105][106][107][108][109][110][111][112], databases [18,19,29,46,60,[113][114][115][116][117][118][119][120][121][122][123][124][125][126][127][128][129][130], security [131][132][133][134][135...…”

Gamify Stencil Dwarf on Cloud for Democratizing Scientific Computing

“…Though the applications are diverse in scientific computing, analogy to AI, there are several common and performance-critical operations in scientific computing, named Dwarf, defined by the Berkeley View [63]. As one of the seven computational Dwarfs, Stencil is ubiquitously involved in various scientific computing [14], which lies at the heart of thermal diffusion (∼100%), earth system model (>90%), and earthquake prediction model (>90%), etc [34,59,8,40,60,15,12].…”