Scale-Out Processors | EPFL Graph Search

Scale-out datacenters mandate high per-server throughput to get the maximum benefit from the large TCO investment. Emerging applications (e.g., data serving and web search) that run in these datacenters operate on vast datasets that are not accommodated by on-die caches of existing server chips. Large caches reduce the die area available for cores and lower performance through long access latency when instructions are fetched. Performance on scale-out workloads is maximized through a modestly-sized last-level cache that captures the instruction footprint at the lowest possible access latency. In this work, we introduce a methodology for designing scalable and efficient scale-out server processors. Based on a metric of performance-density, we facilitate the design of optimal multi-core configurations, called pods. Each pod is a complete server that tightly couples a number of cores to a small last-level cache using a fast interconnect. Replicating the pod to fill the die area yields processors which have optimal performance density, leading to maximum per-chip throughput. Moreover, as each pod is a stand-alone server, scale-out processors avoid the expense of global (i.e., interpod) interconnect and coherence. These features synergistically maximize throughput, lower design complexity, and improve technology scalability. In 20nm technology, scale-out chips improve throughput by 5x-6.5x over conventional and by 1.6x-1.9x over emerging tiled organizations.

Transactions Chasing Scalability and Instruction Locality on Multicores

Pinar Tözün

For several decades, online transaction processing (OLTP) has been one of the main server applications that drives innovations in the data management ecosystem, and in turn the database and computer architecture communities. Recent hardware trends oblige software to overcome two major challenges against systems scalability on modern multicore processors: (1) exploiting the abundant thread-level parallelism across cores and (2) taking advantage of the implicit parallelism within a core. The traditional design of the OLTP systems, however, faces inherent scalability problems due to its tightly coupled components. In addition, OLTP cannot exploit the full capability of the micro-architectural resources of modern processors because of the conventional scheduling decisions that ignore the cache locality for transactions. As a result, today’s commonly used server hardware remains largely underutilized leading to a huge waste of hardware resources and energy. .... In this thesis, we first identify the unbounded critical sections of traditional OLTP systems as the main enemy of thread-level parallelism. We design an alternative shared-everything system based on physiological partitioning (PLP) to eliminate the unbounded critical sections while providing an infrastructure for low-cost dynamic repartitioning and without introducing high-cost distributed transactions. Then, we demonstrate that L1 instruction cache stalls are the dominant factor leading to underutilization in the commodity servers. However, we also observe that independently of their high-level functionality, transactions running in parallel on a multicore system share significant amount of common instructions. By adaptively spreading the execution of a transaction over multiple cores through thread migration or multiplexing transactions on one core, we enable both an ample L1 instruction cache capacity for a transaction and reuse of common instructions across concurrent transactions. .... As the hardware demands more from the software to exploit the complexity and parallelism it offers in the multicore era, this work would change the way we traditionally schedule transactions. Instead of viewing a transaction as a single big task, we split it into smaller parts that can exploit data and instruction locality through careful dynamic scheduling decisions. The methods this thesis presents are not only specific to OLTP systems, but they can also benefit other types of applications that have concurrent requests executing a series of actions from a predefined set and face similar scalability problems on emerging hardware.

EPFL2014

NOC-Out: Microarchitecting a Scale-Out Processor

Babak Falsafi, Boris Robert Grot, Pejman Lotfi Kamran

Scale-out server workloads benefit from many-core processor organizations that enable high throughput thanks to abundant request-level parallelism. A key characteristic of these workloads is the large instruction footprint that exceeds the capacity of private caches. While a shared last-level cache (LLC) can capture the instruction working set, it necessitates a low-latency interconnect fabric to minimize the core stall time on instruction fetches serviced by the LLC. Many-core processors with a mesh interconnect sacrifice performance on scale-out workloads due to NOC-induced delays. Low diameter topologies can overcome the performance limitations of meshes through rich inter-node connectivity, but at a high area expense. To address the drawbacks of existing designs, this work introduces NOC-Out – a many-core processor organization that affords low LLC access delays at a small area cost. NOC-Out is tuned to accommodate the bilateral core-to-cache access pattern, characterized by minimal coherence activity and lack of inter-core communication, that is dominant in scale-out workloads. Optimizing for the bilateral access pattern, NOC-Out segregates cores and LLC banks into distinct network regions and reduces costly network connectivity by eliminating the majority of inter-core links. NOC-Out further simplifies the interconnect through the use of low-complexity tree based topologies. A detailed evaluation targeting a 64-core CMP and a set of scale-out workloads reveals that NOC-Out improves system performance by 17% and reduces network area by 28% over a tiled mesh-based design. Compared to a design with a richly-connected flattened butterfly topology, NOC-Out reduces network area by 9x while matching the performance.

2012

Scale-Out Processors

Pejman Lotfi Kamran

Global-scale online services, such as Google’s Web search and Facebook’s social networking, run in large-scale datacenters. Due to their massive scale, these services are designed to scale out (or distribute) their respective loads and datasets across thousands of servers in datacenters. The growing demand for online services forced service providers to build networks of datacenters, which require an enormous capital outlay for infrastructure, hardware, and power consumption. Consequently, efficiency has become a major concern in the design and operation of such datacenters, with processor efficiency being of, utmost importance, due to the significant contribution of processors to the overall datacenter performance and cost. Scale-out workloads, which are behind today’s online services, serve independent requests, and have large instruction footprints and little data locality. As such, they benefit from processor designs that feature many cores and a modestly sized Last-Level Cache (LLC), a fast access path to the LLC, and high-bandwidth interfaces to memory. Existing server-class processors with large LLCs and a handful of aggressive out-of-order cores are inefficient in executing scale-out workloads. Moreover, the scaling trajectory for these processors leads to even lower efficiency in future technology nodes. This thesis presents a family of throughput-optimal processors, called Scale-Out Processors, for the efficient execution of scale-out workloads. A unique feature of Scale-Out Processors is that they consist of multiple stand-alone modules, called pods, wherein each module is a server running an operating system and a full software stack. To design a throughput-optimal processor, we developed a methodology based on performance density, defined as throughput per unit area, to quantify how effectively an architecture uses the silicon real estate. The proposed methodology derives a performance-density optimal processor building block (i.e., pod), which tightly couples a number of cores to a small LLC via a fast interconnect. Scale-Out Processors simply consist of multiple pods with no inter-pod connectivity or coherence. Moreover, they deliver the highest throughput in today’s technology and afford near-ideal scalability as process technology advances. We demonstrate that Scale-Out Processors improve datacenters’ efficiency by 4.4x-7.1x over datacenters designed using existing server-class processors.