From the Book: Who Needs This Book It is not MindShares intention to convince the reader that the InfiniBand networking technology is superior to other networking technologies. As with all of our books, the goal was to provide a detailed description of the InfiniBand technology for those tasked with designing InfiniBand hardware andor software, as well as those responsible for supporting the technology once it has been deployed. The MindShare Architecture Series The MindShare System Architecture book series is published by Addison-Wesley and includes these publications: 80486 System Architecture,Third Edition Pentium Processor System Architecture, Second Edition Pentium Pro and Pentium II System Architecture, Second Edition Power PC System Architecture PCI System Architecture, Fourth Edition EISA System Architecture, Second Edition FireWire® System Architecture: IEEE 1394 ISA System Architecture, Third Edition Universal Serial Bus System Architecture PCI-X System Architecture InfiniBand System Architecture PCMCIA System Architecture, Second Edition: 16-Bit PC Cards CardBus System Architecture Plug and Play System Architecture Protected Mode Software Architecture AGP System Architecture Cautionary Note The reader should keep in mind that MindShares book series often deals with rapidly evolving technologies. This being the case, it should be recognized that the book is a snapshot of the state of the targeted technology at the time that the book was completed. We attempt tochanges in the targeted technology, but, due to various factors (waiting for the next version of the spec to be frozen, the time necessary to make the changes, and the time to produce the books and get them out to the distribution channels), there will always be a delay. Specifications This Book Is Based On This book is based on Volumes 1 and 2 of the version 1.0a InfiniBand specification. Specification Is the Final Word As with all of our books, this book represents the authors interpretation of the IBA 1.0a specification. When in doubt, the specification is the final word. Organization of This Book The book is organized as follows: Part 1: Introductory Topics. A comprehensive overview of the InfiniBand technology can be gained from this part of the book. The reader who does not need a comprehensive understanding of all the tortuous detail can just read this part. Part 2: QP Creation and Operation. Part 3: Protection Mechanisms. Part 4: Detailed Description of the Transport Services. Part 5: Link and Physical Layer Descriptions. Part 6: The SM and the SA. Part 7: General Services. Appendix: Contains the glossary of terms. Documentation Conventions This document contains conventions for numeric values as follows: Hexadecimal Notation. This section defines the typographical convention used throughout this book. All hex numbers are followed by an h. Examples: 9A4Eh, 0100h Binary Notation. All binary numbers are followed by a b. Examples& 01b Decimal Notation. Numbers without any suffix are decimal. When required for clarity, decimal numbers are followed by a d. The following examples each represent a decimal number: 16, 255, 256d, 128d Bits Versus Bytes Notation. All abbreviations for bits use lower case. For example: 2.5Gbs = 2.5 Gigabits per second; 2Mb = 2 Megabits. All references to bytes are specified in upper case. For example: 10MBs = 10 Megabytes per second; 2KB = 2 Kilobits Other designations: lsb refers to the least-significant bit; LSB refers to the least-signicant byte; msb refers to the most-significant bit; MSB refers to the most-significant byte. Bit Fields (Logical Groups of Bits or Signals). In many cases, bit fields are documented as 15:8, with this example referring to bits 8-through-15. Visit Our Web Site In addition to listing all of our courses and books, our Web site ( ) contains: Errata for a number of the books. Information on the courses that we teach at your site as well as live over the Web to your site(s). Short courses available for viewing. Technical papers. Our publishers web site contains a listing or our currently available books and includes pricing and ordering information. Their home page is accessible at We Want Your Feedback MindShare values your comments and suggestions. You can contact us via mail, phone, fax or Internet email. Phone: (719) 487-1417 and in the U.S. (800) 633-1440.Fax: (719) 487-1434.Technical seminars questions: E-mail , or Address: MindShare, Inc., 4285 Slash Pine Drive, Colorado Springs, CO 80908
Cited By
- Li S and Hoefler T Near-optimal sparse allreduce for distributed deep learning Proceedings of the 27th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, (135-149)
- Li S and Hoefler T Chimera Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, (1-14)
- Montazeri B, Li Y, Alizadeh M and Ousterhout J Homa Proceedings of the 2018 Conference of the ACM Special Interest Group on Data Communication, (221-235)
- Rao N, Liu Q, Sen S, Towlsey D, Vardoyan G, Kettimuthu R and Foster I TCP Throughput Profiles Using Measurements over Dedicated Connections Proceedings of the 26th International Symposium on High-Performance Parallel and Distributed Computing, (193-204)
- Liu Q and Russell R A performance study of InfiniBand fourteen data rate (FDR) Proceedings of the High Performance Computing Symposium, (1-10)
- Claver J, Canseco M, Agustí P and León G A hardware NIC scheduler to guarantee qos on high performance servers Proceedings of the 4th international conference on Parallel and Distributed Processing and Applications, (86-97)
- Bowers K, Chow E, Xu H, Dror R, Eastwood M, Gregersen B, Klepeis J, Kolossvary I, Moraes M, Sacerdoti F, Salmon J, Shan Y and Shaw D Scalable algorithms for molecular dynamics simulations on commodity clusters Proceedings of the 2006 ACM/IEEE conference on Supercomputing, (84-es)
- Fine S, Freund A, Jaeger I, Mansour Y, Naveh Y and Ziv A (2006). Harnessing Machine Learning to Improve the Success Rate of Stimuli Generation, IEEE Transactions on Computers, 55:11, (1344-1355), Online publication date: 1-Nov-2006.
- Bermudez A and Casado R (2006). Fast Routing Computation on InfiniBand Networks, IEEE Transactions on Parallel and Distributed Systems, 17:3, (215-226), Online publication date: 1-Mar-2006.
- Sarkar P, Uttamchandani S and Voruganti K Storage over IP Proceedings of the 2nd USENIX conference on File and storage technologies, (17-17)
Recommendations
PM/InfiniBand-FJ: A High Performance Communication Facility Using InfiniBand for Large Scale PC Clusters
HPCASIA '04: Proceedings of the High Performance Computing and Grid in Asia Pacific Region, Seventh International ConferenceThis paper describes a design of high performance communication facility called the PM/InfiniBand-FJ using InfiniBand interconnect for large scale PC clusters. The PM/InfiniBand-FJ has developed to realize higher application performance than commercial ...
Topologies for improved InfiniBand latency
SC '06: Proceedings of the 2006 ACM/IEEE conference on SupercomputingInfiniBand continues to become more and more important in High Performance Computing world. This talk discusses the impact of DDR (i.e., 20 GigE) switches and HCA's on the creation of low latency, high bandwidth InfiniBand fabrics. When combined with ...
Performance evaluation of InfiniBand with PCI Express
HOTI '04: Proceedings of the High Performance Interconnects, 2004. on Proceedings. 12th Annual IEEE SymposiumWe present an initial performance evaluation of InfiniBand HCAs (host channel adapters) from Mellanox with PCI Express interfaces. We compare the performance with HCAs using PCI-X interfaces. Our results show that InfiniBand HCAs with PCI Express can ...