Locality-oblivious cache organization leveraging single-cycle multi-hop NoCs

Abstract

Locality has always been a critical factor in on-chip data placement on CMPs as accessing further-away caches has in the past been more costly than accessing nearby ones. Substantial research on locality-aware designs have thus focused on keeping a copy of the data private. However, this complicates the problem of data tracking and search/invalidation; tracking the state of a line at all on-chip caches at a directory or performing full-chip broadcasts are both non-scalable and extremely expensive solutions. In this paper, we make the case for Locality-Oblivious Cache Organization (LOCO), a CMP cache organization that leverages the on-chip network to create virtual single-cycle paths between distant caches, thus redefining the notion of locality. LOCO is a clustered cache organization, supporting both homogeneous and heterogeneous cluster sizes, and provides near single-cycle accesses to data anywhere within the cluster, just like a private cache. Globally, LOCO dynamically creates a virtual mesh connecting all the clusters, and performs an efficient global data search and migration over this virtual mesh, without having to resort to full-chip broadcasts or perform expensive directory lookups. Trace-driven and full system simulations running SPLASH-2 and PARSEC benchmarks show that LOCO improves application run time by up to 44.5% over baseline private and shared cache.

Adaptive memory-side last-level GPU caching

“…Jigsaw [24] uses a hardware/software approach to partition cache banks and manage the data placement in 'shares' of sub-partitions to address scalability and interference issues in shared caches. Kwon et al [56] cluster cores and LLC banks and use a fast interconnect to provide cores with a fast inter-and intra-cluster LLC access. Abousamra et al [57] propose a symbiotic NoC/LLC design to reduce latency and improve cache utilization.…”

Scalability of Broadcast Performance in Wireless Network-on-Chip

“…The main reason is that these functions would greatly benefit from global communication schemes [7], [29], implying that in spite of the recent coherence-aware proposals [12], [30], multicast support will become a key design point for NoCs. Also, and in spite of the recent NoC-aware works in architecture [31], [32], this implies that avoiding multicast may become unaffordable at some point. Regardless of whether multicast is avoided or intensively used, scaling parallel architectures and applications increases their multicast requirements.…”