Powerful commodity systems and networks o#er a promising direction for high performance computing because they are inexpensive and they closely track technology progress. However, high, raw--hardware performance is rarely delivered to the end user. Previous work has shown that the bottleneck in these architectures is the overheads imposed by the software communication layer. To reduce these overheads, researchers have proposed a number of user-space communication models. The common feature of these models is that applications have direct access to the network, bypassing the operating system in the common case and thus avoiding the cost of send/receive system calls. In this paper we examine five user--space communication layers, that represent di#erent points in the configuration space: Generic AM, BIP-0.92, FM-2.02, PM-1.2, and VMMC-2. Although these systems support di#erent communication paradigms and employ a variety of di#erent implementation tradeo#s, we are able to quantitatively...

An important aspect of a high-speed network system is the ability to transfer data directly between the network interface and application buffers. Such a direct data path requires the network interface to “know ” the virtual-to-physical address translation of a user buffer, i.e., the physical memory location of the buffer. This paper presents an efficient address translation architecture, User-managed TLB (UTLB), which eliminates system calls and device interrupts from the common communication path. UTLB also supports application-specific policies to pin and unpin application memory. We report micro-benchmark results for an implementation on Myrinet PC clusters. A trace-driven analysis is used to compare the UTLB approach with the interrupt-based approach. It is also used to study the effects of UTLB cache size, associativity, and prefetching. Our results show that the UTLB approach delivers robust performance with relatively small translation cache sizes. 1

Device drivers are well known to be one of the prime sources of unreliability in today's computer systems. We argue that this need not be, as drivers can be run as user-level tasks, allowing them to be encapsulated by hardware protection. In contrast to prior work on user-level drivers, we show that on present hardware it is possible to prevent DMA from undermining this encapsulation.

Abstract not found

Rights to individual papers remain with the author or the author's employer. Permission is granted for noncommercial reproduction of the work for educational or research purposes. This copyright notice must be included in the reproduced paper. USENIX acknowledges all trademarks herein.

Traditional message-passing (TMP) is characterized by multiple, individual, point-to-point, typed messages, inducing unintended temporal synchronization. Asynchronous communication can hide some of the overhead and latency from an application, but doing so changes neither the behavior of nor the demands placed upon the communication substrate; messages of the same size are sent at approximately the same times whether synchronous or asynchronous communication is used.