josh@polaris.UUCP (Josh Knight) (08/21/86)
%A S.G. Tucker %T The IBM 3090 System: An Overview %J IBM Sys. J. %V 25 %N 1 %P 4-19 %D 1986 JUN %X The first part of this paper places the IBM 3090 system in historical perspective with respect to its predecessors. Treated briefly are the technology and the design process, both of which were critical to the development of the 3090. Presented in detail is the 3090 system itself, with emphasis on its features that differ from those of prior systems. %A Y. Singh %A G.M. King %A J.W. Anderson %T IBM 3090 Performance: A Balanced System Approach %J IBM Sys. J. %V 25 %N 1 %P 20-35 %D 1986 JUN %X The IBM 3090 system represents the highest level of system performance offered by IBM to date. To realize the full performance potential of this system, it is essential to maintain a balance among its various components. The major components of the system are the processor(s), storage, I/O and the software that manages the system resources. Their performance attributes are discussed and their effect on system performance illustrated by laboratory benchmark measurements for the MVS and VM operating systems. %A D.H. Gibson %A D.W. Rain %A H.F. Walsh %T Engineering and Scientific Processing on the IBM 3090 %J IBM Sys. J. %V 25 %N 1 %P 36-50 %D 1986 JUN %X The IBM 3090 processor implementation of the System/370 Vector Architecture represents a major new system design for engineering and scientific processing, featuring both scalar and vector capability in a uniprocessor and in a dyadic and four-way parallel processing environment. The history of large-scale scientific processing is reviewed, leading to a statement of current requirements. The design objectives for scalar, parallel, and vector capabilities are identified, followed by a summary of the resulting 3090 features. Selected highlights of the vector hardware are given, followed by a summary of the supporting software. The paper concludes with a discussion of performance, beginning with the identification of suitable applications. An example is given of one application utilizing each of the three capabilities: scalar, parallel and vector. Several of the most important performance parameters are identified. %A W. Bucholz %T The IBM System/370 Vector Architecture %J IBM Sys. J. %V 25 %N 1 %P 51-62 %D 1986 JUN %X Discussed is the instruction-set architecture of the IBM System/370 vector facility, a compatible extension of the System/370 architecture. Both the base system, which is a general-purpose System/370 processor, and the optional vector facility employ a register type of organization. Data formats are the same, arithmetic operations produce exactly the same results, arithmetic exceptions are handled in the same way, and instructions are precisely interruptible for page faults and other causes in the same manner as those of the base system. This approach permits substantially increased performance on vectorizable programs with only a modest increase in hardware and software, while retaining the ability to run existing nonvector programs unchanged. %A R.S. Clark %A T.L. Wilson %T Vector System Performance of the IBM 3090 %J IBM Sys. J. %V 25 %N 1 %P 63-82 %D 1986 JUN %X Performance of the Vector Facility of the IBM 3090 processor is discussed. The paper has two parts, the first presenting factors affecting performance measurement of the Vector Facility and the criteria for its design. In the second part, use of the 3090 storage hierarchy to support the vector processing implementation is the main aspect of the discussion. %A K.D. Gottschalk %T The System Usability Process for Network Management Products %J IBM Sys. J. %V 25 %N 1 %P 83-91 %D 1986 JUN %X This paper presents an overview of a process for system usability. The process is a systematic series of activities and procedures designed to improve the usability of software network management products. The elements of the process are given and future directions for evaluating usability described. %A L.C. Percival %A S.K. Johnson %T Network Management Software Usability Design and Test Implementation %J IBM Sys. J. %V 25 %N 1 %P 92-104 %D 1986 JUN %X The approach used at one of IBM's development sites for usability testing is somewhat different from methods used elsewhere. The approach was developed specifically for testing of software communications products as one aspect of the System Usability Process. The test design and implementations are described. %A L. Koven %A G. Waldbaum %T Improving Availability of Software Subsystems through On-line Error Detection %J IBM Sys. J. %V 25 %N 1 %P 105-115 %D 1986 JUN %X A VM/370 program called Auditor detects faults in the operation of computer software subsystems and attempts to restore service as quickly as possible. Through a series of periodic tests, Auditor diagnoses whether these subsystems are operating properly. When faults are detected, service restoration procedures are automatically called, and the persons responsible for the subsystems are notified. The various types of faults are recorded for subsequent analysis. %A J.K. Kravitz %A D. Lieber %A F.H. Robbins %A J.M. Palermo %T Workstations and Mainframe Computers Working Together %J IBM Sys. J. %V 25 %N 1 %P 116-128 %D 1986 JUN %X The history and design philosophy of a research project that produced a prototype software package are described. The package provides cooperation between large-scale mainframe computing systems and personal desktop workstations. Also discussed are the features of PC/VM Bond, the product that grew out of the research project. -- Josh Knight, IBM T.J. Watson Research josh@ibm.com, josh@yktvmh.bitnet, ...!philabs!polaris!josh