CAPE: A Content-Addressable Processing Engine

Caminal, Helena; Yang, Kun; Srinivasa, Srivatsa; Ramanathan, Akshay Krishna; Al-Hawaj, Khalid; Wu, Tianshu; Narayanan, Vijaykrishnan; Batten, Christopher; Martínez, José F.

doi:10.1109/hpca51647.2021.00054

Cited by 13 publications

(29 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A Content-Addressable Processing Engine (CAPE) [15] is a generalpurpose AP core that adopts a contemporary ISA abstraction, namely RISC-V with standard vector extensions, and can be readily integrated into a tiled architecture. CAPE has been shown to accelerate a broad range of data-parallel workloads [15]. CAPE's architecture comprises four main blocks (Figure 3).…”

Section: A Content-addressable Processing Enginementioning

confidence: 99%

“…CAPE's associative arrays are built out of SRAM technology. A CAPE with a 4 MB CSB is comparable in area to a conventional out-of-order core [15].…”

Section: A Content-addressable Processing Enginementioning

confidence: 99%

“…As described in [15], the CSB is organized in chains of 32×32 (rows×columns) memory subarrays. Each of these subarrays 𝑖 propagates the tag bits to their next 𝑖 + 1 neighbor, hence, forming a chain.…”

Section: A Content-addressable Processing Enginementioning

confidence: 99%

“…Associative Processors (AP), first introduced in the 1970s [22], have recently re-emerged as an attractive concept for general purpose applications that exhibit data-level parallelism [15,46,66,68]. In addition to their very long vector-style compute capabilities, APs provide efficient multi-row search and update operations.…”

Section: Introductionmentioning

confidence: 99%

“…In this paper, we investigate how amenable APs are to delivering high performance on analytic query workloads. We build a research prototype database system called Castle that executes end-to-end analytic query workloads on the recently proposed Content-Addressable Processing Engine (CAPE) [15]. CAPE is an associative processor core that is highly programmable via the RISC-V ISA with standard vector extensions.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Accelerating database analytic query workloads using an associative processor

Caminal

Chronis

et al. 2022

Proceedings of the 49th Annual International Symposium on Computer Architecture

Self Cite

View full text Add to dashboard Cite

Database analytic query workloads are heavy consumers of datacenter cycles, and there is constant demand to improve their performance. Associative processors (AP) have re-emerged as an attractive architecture that offers very large data-level parallelism that can be used to implement a wide range of general-purpose operations. Associative processing is based primarily on efficient search and bulk update operations. Analytic query workloads benefit from data parallel execution and often feature both search and bulk update operations. In this paper, we investigate how amenable APs are to improving the performance of analytic query workloads. For this study, we use the recently proposed Content-Addressable Processing Engine (CAPE) framework. CAPE is an AP core that is highly programmable via the RISC-V ISA with standard vector extensions. By mapping key database operators to CAPE and introducing APaware changes to the query optimizer, we show that CAPE is a good match for database analytic workloads. We also propose a set of database-aware microarchitectural changes to CAPE to further improve performance. Overall, CAPE achieves a 10.8× speedup on average (up to 61.1×) on the SSB benchmark (a suite of 13 queries) compared to an iso-area aggressive out-of-order processor with AVX-512 SIMD support.

show abstract

Section: A Content-addressable Processing Enginementioning

confidence: 99%

“…CAPE's associative arrays are built out of SRAM technology. A CAPE with a 4 MB CSB is comparable in area to a conventional out-of-order core [15].…”

Section: A Content-addressable Processing Enginementioning

confidence: 99%

Section: A Content-addressable Processing Enginementioning

confidence: 99%