Manycore network interfaces for in-memory rack-scale computing
Open Access
- 13 June 2015
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 567-579
Abstract
Datacenter operators rely on low-cost, high-density technologies to maximize throughput for data-intensive services with tight tail latencies. In-memory rack-scale computing is emerging as a promising paradigm in scale-out datacenters capitalizing on commodity SoCs, low-latency and high-bandwidth communication fabrics and a remote memory access model to enable aggregation of a rack's memory for critical data-intensive applications such as graph processing or key-value stores. Low latency and high bandwidth not only dictate eliminating communication bottlenecks in the software protocols and off-chip fabrics but also a careful on-chip integration of network interfaces. The latter is a key challenge especially in architectures with RDMA-inspired one-sided operations that aim to achieve low latency and high bandwidth through on-chip Network Interface (NI) support. This paper proposes and evaluates network interface architectures for tiled manycore SoCs for in-memory rack-scale computing. Our results indicate that a careful splitting of NI functionality per chip tile and at the chip's edge along a NOC dimension enables a rack-scale architecture to optimize for both latency and bandwidth. Our best manycore NI architecture achieves latencies within 3% of an idealized hardware NUMA and efficiently uses the full bisection bandwidth of the NOC, without changing the on-chip coherence protocol or the core's microarchitecture.Keywords
Funding Information
- Schweizerische Nationalfonds zur Förderung der Wissenschaftlichen Forschung (200021_140550)
- Microsoft Research (2014-006 (DP1061309))
- Nano-Tera (530665)
This publication has 32 references indexed in Scilit:
- Scale-out NUMA2014
- Workload analysis of a large-scale key-value storeACM SIGMETRICS Performance Evaluation Review, 2012
- Achieving predictable performance through better memory controller placement in many-core CMPsACM SIGARCH Computer Architecture News, 2009
- Integrated network interfaces for high-bandwidth TCP/IP2006
- Near-Optimal Worst-Case Throughput Routing for Two-Dimensional Mesh NetworksACM SIGARCH Computer Architecture News, 2005
- Modeling object characteristics of dynamic Web contentJournal of Parallel and Distributed Computing, 2003
- The SGI Origin1997
- Coherent network interfaces for fine-grain communication1996
- Hive1995
- Integration of message passing and shared memory in the Stanford FLASH multiprocessorACM SIGPLAN Notices, 1994