Optical Bus Control for Distributed Multiprocessors
نویسندگان
چکیده
Optical interconnections offer a potential for gigahertz transfer rates in an environment free from capacitive bus loading, crosstalk, and electromagnetic interference. Device technology in electrooptics has matured to a point where small, low power, and low cost devices exist which are suitable for use in bus level implementations. Therefore, the realization of physically distributed, bus-interconnected multiprocessors is now possible. In this paper we propose a bus arbitration mechanism suitable for a large optical bus structure heavily populated with asynchronous bus masters. It is a two-level arbitration system with incoming requests batched on a demand basis and serviced in a linear priority order within each batch. Low priority requests cannot be starved in high contention environments. For an optical bus of fixed length, propagation delay is bounded by the end-to-end transit time and is independent of the number of devices attached. In addition, as contention for the bus increases and batch sizes become larger, the time overhead paid for bus control decreases. Thus, a performance improvement is achieved dynamically under high contention conditions. o 1990 Academic Press, IN.
منابع مشابه
Predicting the performance measures of an optical distributed shared memory multiprocessor by using support vector regression
Recent advances in the development of optical technologies suggest the possible emergence of optical interconnects within distributed shared memory (DSM) multiprocessors. The performance of these DSM architectures must be evaluated under varying values of DSM parameters. In this paper, we develop a Support Vector Regression (SVR) model for predicting the performance measures (i.e. average netwo...
متن کاملHardware Prefetching in Bus-Based Multiprocessors: Pattern Characterization and Cost-Effective Hardware
Data prefetching has been widely studied as a technique to hide memory access latency in multiprocessors. Most recent research on hardware prefetching focuses either on uniprocessors, or on distributed shared memory (DSM) and other non bus-based organizations. However, in the context of bus-based SMPs, prefetching poses a number of problems related to the lack of scalability and limited bus ban...
متن کاملType Data tra c Replacement tra c Coherence tra cUMA
Shared-bus multiprocessors represent a mainstream of accepted and commercially viable computer systems. However, as microprocessors become faster and demand more bandwidth, the already limited scalability of shared-bus decreases even further. As an eeort, not a mutually exclusive but rather a complementary to developing better backplane bus, this paper considers adapting distributed shared-memo...
متن کاملDesign of Large-Scale Symmetric Multiprocessors (SMPs) using Parallel Optical Interconnects
In this paper, we address the primary limitation of bandwidth demands for address transaction in future cache coherent symmetric multiprocessors (SMPs). As a solution, we propose a scalable optical address sub-network called Symmetric Multiprocessor Network (SYMNET). SYMNET, not only has the ability to pipeline address requests, but also multiple address requests from different processors can p...
متن کاملReconfiguration and optimal placement of distributed generations in distribution networks in the presence of remote voltage controlled bus using exchange market algorithm
Abstract: Since distribution networks have a large share of the losses in power systems, reducing losses in these networks is one of the key issues in reducing the costs of global networks, including issues Which has always been considered. In this thesis, the reconfiguration of the distribution network in the presence of distributed generation sources (DGs) with respect to two types of bus, P ...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید
ثبت ناماگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید
ورودعنوان ژورنال:
- J. Parallel Distrib. Comput.
دوره 10 شماره
صفحات -
تاریخ انتشار 1990