Message-Driven Processor Architecture: Verson 11

The Message-Driven Processor is a node of a large-scale multiprocessor being developed by the Concurrent VLSI Architecture Group. It is intended to support fine-grained, message passing, parallel computation. It contains several novel architectural features, such as a low-latency network interface, extensive type-checking hardware, and on-chip memory that can be used as an associative lookup table. This document is a programmer's guide to the MDP. It describes the processor's register architecture, instruction set, and the data types supported by the processor. It also details the MDP's message sending and exception handling facilities.

An Operating Environment for the Jellybean Machine

The Jellybean Machine is a scalable MIMD concurrent processor consisting of special purpose RISC processors loosely coupled into a low latency network. I have developed an operating system to provide the supportive environment required to efficiently coordinate the collective power of the distributed processing elements. The system services are developed in detail, and may be of interest to other designers of fine grain, distributed memory processing networks.

Fault-Tolerant Design for Multistage Routing Networks

As the size of digital systems increases, the mean time between single component failures diminishes. To avoid component related failures, large computers must be fault-tolerant. In this paper, we focus on methods for achieving a high degree of fault-tolerance in multistage routing networks. We describe a multipath scheme for providing end-to-end fault-tolerance on large networks. The scheme improves routing performance while keeping network latency low. We also describe the novel routing component, RN1, which implements this scheme, showing how it can be the basic building block for fault-tolerant multistage routing networks.

Robust, High-Speed Network Design for Large-Scale Multiprocessing

As multiprocessor system size scales upward, two important aspects of multiprocessor systems will generally get worse rather than better: (1) interprocessor communication latency will increase and (2) the probability that some component in the system will fail will increase. These problems can prevent us from realizing the potential benefits of large-scale multiprocessing. In this report we consider the problem of designing networks which simultaneously minimize communication latency while maximizing fault tolerance. Using a synergy of techniques including connection topologies, routing protocols, signalling techniques, and packaging technologies we assemble integrated, system-level solutions to this network design problem.

A Parallel Crossbar Routing Chip for a Shared Memory Multiprocessor

This thesis describes the design and implementation of an integrated circuit and associated packaging to be used as the building block for the data routing network of a large scale shared memory multiprocessor system. A general purpose multiprocessor depends on high-bandwidth, low-latency communications between computing elements. This thesis describes the design and construction of RN1, a novel self-routing, enhanced crossbar switch as a CMOS VLSI chip. This chip provides the basic building block for a scalable pipelined routing network with byte-wide data channels. A series of RN1 chips can be cascaded with no additional internal network components to form a multistage fault-tolerant routing switch. The chip is designed to operate at clock frequencies up to 100Mhz using Hewlett-Packard's HP34 $1.2\\mu$ process. This aggressive performance goal demands that special attention be paid to optimization of the logic architecture and circuit design.

Design and Evaluation of the Hamal Parallel Computer

Parallel shared-memory machines with hundreds or thousands of processor-memory nodes have been built; in the future we will see machines with millions or even billions of nodes. Associated with such large systems is a new set of design challenges. Many problems must be addressed by an architecture in order for it to be successful; of these, we focus on three in particular. First, a scalable memory system is required. Second, the network messaging protocol must be fault-tolerant. Third, the overheads of thread creation, thread management and synchronization must be extremely low. This thesis presents the complete system design for Hamal, a shared-memory architecture which addresses these concerns and is directly scalable to one million nodes. Virtual memory and distributed objects are implemented in a manner that requires neither inter-node synchronization nor the storage of globally coherent translations at each node. We develop a lightweight fault-tolerant messaging protocol that guarantees message delivery and idempotence across a discarding network. A number of hardware mechanisms provide efficient support for massive multithreading and fine-grained synchronization. Experiments are conducted in simulation, using a trace-driven network simulator to investigate the messaging protocol and a cycle-accurate simulator to evaluate the Hamal architecture. We determine implementation parameters for the messaging protocol which optimize performance. A discarding network is easier to design and can be clocked at a higher rate, and we find that with this protocol its performance can approach that of a non-discarding network. Our simulations of Hamal demonstrate the effectiveness of its thread management and synchronization primitives. In particular, we find register-based synchronization to be an extremely efficient mechanism which can be used to implement a software barrier with a latency of only 523 cycles on a 512 node machine.

Performance Evaluation of the Scheme 86 and HP Precision Architecture

The Scheme86 and the HP Precision Architectures represent different trends in computer processor design. The former uses wide micro-instructions, parallel hardware, and a low latency memory interface. The latter encourages pipelined implementation and visible interlocks. To compare the merits of these approaches, algorithms frequently encountered in numerical and symbolic computation were hand-coded for each architecture. Timings were done in simulators and the results were evaluated to determine the speed of each design. Based on these measurements, conclusions were drawn as to which aspects of each architecture are suitable for a high- performance computer.

Fat-Tree Routing for Transit

The Transit network provides high-speed, low-latency, fault-tolerant interconnect for high-performance, multiprocessor computers. The basic connection scheme for Transit uses bidelta style, multistage networks to support up to 256 processors. Scaling to larger machines by simply extending the bidelta network topology will result in a uniform degradation of network latency between all processors. By employing a fat-tree network structure in larger systems, the network provides locality and universality properties which can help minimize the impact of scaling on network latency. This report details the topology and construction issues associated with integrating Transit routing technology into fat-tree interconnect topologies.

Pi: A Parallel Architecture Interface for Multi-Model Execution

This thesis defines Pi, a parallel architecture interface that separates model and machine issues, allowing them to be addressed independently. This provides greater flexibility for both the model and machine builder. Pi addresses a set of common parallel model requirements including low latency communication, fast task switching, low cost synchronization, efficient storage management, the ability to exploit locality, and efficient support for sequential code. Since Pi provides generic parallel operations, it can efficiently support many parallel programming models including hybrids of existing models. Pi also forms a basis of comparison for architectural components.

Thread Scheduling Mechanisms for Multiple-Context Parallel Processors

Scheduling tasks to efficiently use the available processor resources is crucial to minimizing the runtime of applications on shared-memory parallel processors. One factor that contributes to poor processor utilization is the idle time caused by long latency operations, such as remote memory references or processor synchronization operations. One way of tolerating this latency is to use a processor with multiple hardware contexts that can rapidly switch to executing another thread of computation whenever a long latency operation occurs, thus increasing processor utilization by overlapping computation with communication. Although multiple contexts are effective for tolerating latency, this effectiveness can be limited by memory and network bandwidth, by cache interference effects among the multiple contexts, and by critical tasks sharing processor resources with less critical tasks. This thesis presents techniques that increase the effectiveness of multiple contexts by intelligently scheduling threads to make more efficient use of processor pipeline, bandwidth, and cache resources. This thesis proposes thread prioritization as a fundamental mechanism for directing the thread schedule on a multiple-context processor. A priority is assigned to each thread either statically or dynamically and is used by the thread scheduler to decide which threads to load in the contexts, and to decide which context to switch to on a context switch. We develop a multiple-context model that integrates both cache and network effects, and shows how thread prioritization can both maintain high processor utilization, and limit increases in critical path runtime caused by multithreading. The model also shows that in order to be effective in bandwidth limited applications, thread prioritization must be extended to prioritize memory requests. We show how simple hardware can prioritize the running of threads in the multiple contexts, and the issuing of requests to both the local memory and the network. Simulation experiments show how thread prioritization is used in a variety of applications. Thread prioritization can improve the performance of synchronization primitives by minimizing the number of processor cycles wasted in spinning and devoting more cycles to critical threads. Thread prioritization can be used in combination with other techniques to improve cache performance and minimize cache interference between different working sets in the cache. For applications that are critical path limited, thread prioritization can improve performance by allowing processor resources to be devoted preferentially to critical threads. These experimental results show that thread prioritization is a mechanism that can be used to implement a wide range of scheduling policies.

A Hierarchical Cache Coherent Protocol

As the number of processors in distributed-memory multiprocessors grows, efficiently supporting a shared-memory programming model becomes difficult. We have designed the Protocol for Hierarchical Directories (PHD) to allow shared-memory support for systems containing massive numbers of processors. PHD eliminates bandwidth problems by using a scalable network, decreases hot-spots by not relying on a single point to distribute blocks, and uses a scalable amount of space for its directories. PHD provides a shared-memory model by synthesizing a global shared memory from the local memories of processors. PHD supports sequentially consistent read, write, and test- and-set operations. This thesis also introduces a method of describing locality for hierarchical protocols and employs this method in the derivation of an abstract model of the protocol behavior. An embedded model, based on the work of Johnson[ISCA19], describes the protocol behavior when mapped to a k-ary n-cube. The thesis uses these two models to study the average height in the hierarchy that operations reach, the longest path messages travel, the number of messages that operations generate, the inter-transaction issue time, and the protocol overhead for different locality parameters, degrees of multithreading, and machine sizes. We determine that multithreading is only useful for approximately two to four threads; any additional interleaving does not decrease the overall latency. For small machines and high locality applications, this limitation is due mainly to the length of the running threads. For large machines with medium to low locality, this limitation is due mainly to the protocol overhead being too large. Our study using the embedded model shows that in situations where the run length between references to shared memory is at least an order of magnitude longer than the time to process a single state transition in the protocol, applications exhibit good performance. If separate controllers for processing protocol requests are included, the protocol scales to 32k processor machines as long as the application exhibits hierarchical locality: at least 22% of the global references must be able to be satisfied locally; at most 35% of the global references are allowed to reach the top level of the hierarchy.

Concurrent Smalltalk on the Message-Driven Processor

Concurrent Smalltalk is the primary language used for programming the J- Machine, a MIMD message-passing computer containing thousands of 36-bit processors connected by a very low latency network. This thesis describes in detail Concurrent Smalltalk and its implementation on the J-Machine, including the Optimist II global optimizing compiler and Cosmos fine-grain parallel operating system. Quantitative and qualitative results are presented.

ADAM: A Decentralized Parallel Computer Architecture Featuring Fast Thread and Data Migration and a Uniform Hardware Abstraction

The furious pace of Moore's Law is driving computer architecture into a realm where the the speed of light is the dominant factor in system latencies. The number of clock cycles to span a chip are increasing, while the number of bits that can be accessed within a clock cycle is decreasing. Hence, it is becoming more difficult to hide latency. One alternative solution is to reduce latency by migrating threads and data, but the overhead of existing implementations has previously made migration an unserviceable solution so far. I present an architecture, implementation, and mechanisms that reduces the overhead of migration to the point where migration is a viable supplement to other latency hiding mechanisms, such as multithreading. The architecture is abstract, and presents programmers with a simple, uniform fine-grained multithreaded parallel programming model with implicit memory management. In other words, the spatial nature and implementation details (such as the number of processors) of a parallel machine are entirely hidden from the programmer. Compiler writers are encouraged to devise programming languages for the machine that guide a programmer to express their ideas in terms of objects, since objects exhibit an inherent physical locality of data and code. The machine implementation can then leverage this locality to automatically distribute data and threads across the physical machine by using a set of high performance migration mechanisms. An implementation of this architecture could migrate a null thread in 66 cycles -- over a factor of 1000 improvement over previous work. Performance also scales well; the time required to move a typical thread is only 4 to 5 times that of a null thread. Data migration performance is similar, and scales linearly with data block size. Since the performance of the migration mechanism is on par with that of an L2 cache, the implementation simulated in my work has no data caches and relies instead on multithreading and the migration mechanism to hide and reduce access latencies.