Memory Systems and Interconnects for Scale-Out Servers

The information revolution of the last decade has been fueled by the digitization of almost all human activities through a wide range of Internet services. The backbone of this information age are scale-out datacenters that need to collect, store, and process massive amounts of data. These datacenters distribute vast datasets across a large number of servers, typically into memory-resident shards so as to maintain strict quality-of-service guarantees. While data is driving the skyrocketing demands for scale-out servers, processor and memory manufacturers have reached fundamental efficiency limits, no longer able to increase server energy efficiency at a sufficient pace. As a result, energy has emerged as the main obstacle to the scalability of information technology (IT) with huge economic implications. Delivering sustainable IT calls for a paradigm shift in computer system design. As memory has taken a central role in IT infrastructure, memory-centric architectures are required to fully utilize the IT's costly memory investment. In response, processor architects are resorting to manycore architectures to leverage the abundant request-level parallelism found in data-centric applications. Manycore processors fully utilize available memory resources, thereby increasing IT efficiency by almost an order of magnitude. Because manycore server chips execute a large number of concurrent requests, they exhibit high incidence of accesses to the last-level-cache for fetching instructions (due to large instruction footprints), and off-chip memory (due to lack of temporal reuse in on-chip caches) for accessing dataset objects. As a result, on-chip interconnects and the memory system are emerging as major performance and energy-efficiency bottlenecks in servers. This thesis seeks to architect on-chip interconnects and memory systems that are tuned for the requirements of memory-centric scale-out servers. By studying a wide range of data-centric applications, we uncover application phenomena common in data-centric applications, and examine their implications on on-chip network and off-chip memory traffic. Finally, we propose specialized on-chip interconnects and memory systems that leverage common traffic characteristics, thereby improving server throughput and energy efficiency.

Processeur à haut degré de parallélisme basé sur des composantes sérielles

Ralph Hoffmann

Nowaday, the world of processors is still dominated by the RISC architectures, which foundations have been laid down in the 70's. The RISC concept may be summarized by one word : simplicity. With this concept, much simpler architectures are born, in particular from their instruction sets point of view. As a result, these architectures have a more efficient instruction decoding and smaller computing units which in turn provides the ability to have more of these units for the same price, a reduced clock per instruction ratio — a single cycle per instruction ideally — and increased clock frequencies. As time goes by, RISC architectures have shown outstanding performances, on one hand with the spectacular improvements of semiconductor technologies, and on the other hand with the use of large cache memories and with the development of new ways of execution such as speculative execution. Later developments have greatly altered the original spirit of the simplicity of the RISC concept. Current high performance processors are extremely complex and difficult to design. Moreover, semiconductor technologies begin to show serious problems ignored or neglected for a long time. In this thesis, we would like to introduce a new architecture. Two complementary aspects have been taken in consideration : the design of the processor's units (arithmetic operators, control logic, and so on) and the architectural model, the way the processing units are used. The first aspect has been addressed with an unusual and original approach which involves the use of the serial arithmetic principles. Serial arithmetic tends to reduce the hardware cost, enables higher clock rates, and shows interesting properties which will be developped later on. The research of an architectural model is driven by the same wish of simplicity, recurrent in this field, and will inevitably shift the complexity from the processor to the compiler. This simplification relies on two things : first, the way we think of the problem is inspired by the EPIC philosophy (Explicitly Parallel Instruction Computer), the latest offspring of the VLIW model which advocates the delegation of decisions from the hardware to the software. In one sentence, "the compiler decides, the hardware executes". Second, the way we act has been given by the TTA model (Transport Triggered Architecture), the last offspring of the MOVE model which not only removes from the processor the responsability to decide which processing unit to be used for each instruction (it's already the case in a VLIW processor), but also releases the processor from scheduling the data flows. In doing so, the processor is confined to provide a set of computing units, storage units and means of communication, and the compiler decides how to move the data. In this way, a part of the control logic is removed, communication resources may be fine-tuned and the compiler's code-optimization can be enhanced. With the combination of those elements, we wish to get an architecture which exhibits a better computing power/consumption ratio, which is more flexible, easier to developp and to evolve.

EPFL2004

Scalable Simulation Methodologies for Many-Core Heterogeneous Systems

Shivani Raghav

With increasing complexity and performance demands of emerging compute-intensive data-parallel workloads, many-core computing systems are becoming a popular trend in computer design. Fast and scalable simulation methods are needed to make meaningful predictions of design alternatives to prepare early software development and to assess the performance of a system before the real hardware is available. However, simulation technology for large-scale, many-core systems is lagging behind. Most of the existing simulation techniques are slow and complex or have poor scalability and high cost of development, which leads to an unacceptable performance when simulating a large number of cores. New challenges brought by emerging many-core systems demand new methods for simulating these platforms. This thesis investigates the techniques for improving the performance of parallel simulators to model many-core heterogeneous architectures. With the recent advances in programmability and processing power of graphical hardware, General Purpose Graphical Processing Units (GPGPUs) are becoming a popular host platform for solving broad range of computationally expensive problems. The easy availability of low cost, highly parallel GPGPU platforms presents an opportunity for accelerating architectural simulations. In this thesis, I propose and investigate a novel idea of using GPGPUs as a host platform for accelerating simulation of many-core systems. GPGPUs are most suitable for certain specific highly parallel workloads with fine-grained multi-threading and high computational intensity. Architectural simulation, on the other hand, is an extremely complex task with numerous dependencies and rules. To this end, I first determine the feasibility of accelerating various target many-core architectures at different levels of abstraction details, considering the particularities of prevalent GPU architectures and their programming models. Based on this study, I target a heterogeneous architecture of a multi-core CPU connected to many-core coprocessors for GPGPU acceleration. I propose ways to exploit ample parallelism in many-core simulation and present a comprehensive framework to develop an architecture-independent, fast, scalable, easy-to-use, full-system simulator. In particular, I outline some of the challenges faced and insights gained using our proposed approach. I present specific methods and approaches that maximize the throughput of GPU power and memory bandwidth allowing optimized parallelization with minimal synchronization. Our results show high scalability and high speed up in performance for simulating up to eight thousand cores.

EPFL2014

TurboTag: Lookup Filtering to Reduce Coherence Directory Power

Daniel Crisan, Babak Falsafi, Michael Ferdman, Pejman Lotfi Kamran

On-chip coherence directories of today’s multi-core systems are not energy efficient. Coherence directories dissipate a significant fraction of their power on unnecessary lookups when running commercial server and scientific workloads. These workloads have large working sets that are beyond the reach of on-chip caches of modern processors. Limited to capturing a small part of the working set, private caches retain cache blocks only for a short period of time before replacing them with new blocks. Moreover, coherence enforcement is a known performance bottleneck of multi-threaded software, hence data-sharing in optimized high-performance software is minimal. Consequently, the majority of the accesses to the coherence directory find no sharers in the directory because the data are not available in the on-chip private caches, effectively wasting power on the coherence checks. To improve energy-efficiency for future many-core systems, we propose TurboTag, a filtering mechanism to eliminate needless directory lookups. We analyze full-system traces of server and scientific workloads and find that over 69% of accesses to the directory find no sharers and can be entirely avoided. Taking advantage of this behavior, TurboTag achieves a 58% reduction in the directory’s dynamic power consumption.