Pipeline Machine Research Articles

Maximum Pipelining Of Array Computation: A Pipelined Code Mapping Scheme For Dataflow Computers

AbstractDespite the attractive features of data flow computers, skepticism exists concerning their efficiency in handling arrays (vectors) in high performance scientific computation. This paper argues that massive parallelism in vector computation can be exploited effectively utilizing data flow principles. The key is to organize the data flow machine program graph such that array operations can be fully pipelined. Unlike in conventional pipelined vector processors, there is no requirement that the activities belonging to the same vector operation be continuously processed by one or a group of physically tightly-coupled function units in the processor. The applicative nature of the data flow graph model allows flexible scheduling of the execution of enabled instructions in the pipeline data flow programs. Accordingly, program transformation can be performed on the basis of both the global and local data flow analysis to generate efficient pipelined data flow machine code. As a result of such fine-grain pa...

Parallel Algorithms for Fluid Dynamics Problems

The recent trend toward increased parallelism in supercomputer architectures, together with the need to use existing supercomputer resources more efficiently, has motivated this study of parallel algorithms designed specifically for implementation on parallel machines. Users of parallel machines like the ICL Distributed Array Processor (DAP) are all too familiar with the problem of having to reformulate their algorithms to incorporate the inherent parallelism of the DAP architecture. The individual processing elements of the DAP are comparatively slow, and the full potential of such a machine is only realized if the algorithm used is highly parallel. The very fast processing capabilities of pipeline machines like the Cray-1 or the Cyber 205 tend to obscure this point somewhat, and in fact most users of these machines are content with using existing codes, vectorized where possible. Such a strategy, however, is not appropriate for the parallel processors like the ICL DAP, AMT DAP, the Goodyear MPP, GEC GRID, etc.The application of parallel algorithms to the solution of fluid dynamics problems is considered. Most of the work concerns the solution of the two- and three-dimensional incompressible Navier-Stokes equations, steady and unsteady. The results are relevant for the above-mentioned parallel machines, but the methods can be adapted for use on vector machines.Concerning parallel algorithms, the easiest and possibly those with the widest potential application are relaxation techniques based on red-black (R-B) ordering. The use of R-B successive overrelaxation, R-B LSOR, etc., on two- and three-dimensional cavity flows, unsteady channel, and boundary layer flows at high Reynolds numbers is considered. While these methods are extremely problem dependent, in the favorable cases they can be as competitive as more sophisticated serial “vectorized” algorithms. As far as parallel algorithms are concerned, the most important criteria are what is the cost per iteration, and how fast do they converge. For application on serial machines, the latter factor seems to dominate.

Fine-grain compilation for pipelined machines

Computer architecture design requires careful attention to the balance between the complexity of code scheduling problems and the cost and feasibility of building a machine. In this paper, we show that recently developed software pipelining algorithms produce optimal or near-optimal code for a large class of loops when the target architecture is a clean pipelined parallel machine. The important feature of these machines is the absence of structural hazards. We argue that the robustness of the scheduling algorithms and relatively simple hardware make these machines realistic and cost-effective. To illustrate the delicate balance between architecture and scheduling complexity, we show that scheduling with structural hazards is NP-hard, and that there are machines with simple structural hazards for which vectorization and the software pipelining techniques generate poor code.

A program for predicting significant RNA secondary structures

We describe a program for the analysis of RNA secondary structure. There are two new features in this program. (i) To get vector speeds on a vector pipeline machine (such as Cray X-MP/24) we have vectorized the secondary structure dynamic algorithm. (ii) The statistical significance of a locally 'optimal' secondary structure is assessed by a Monte Carlo method. The results can be depicted graphically including profiles of the stability of local secondary structures and the distribution of the potentially significant secondary structures in the RNA molecules. Interesting regions where both the potentially significant secondary structures and 'open' structures (single-stranded coils) occur can be identified by the plots mentioned above. Furthermore, the speed of the vectorized code allows repeated Monte Carlo simulations with different overlapping window sizes. Thus, the optimal size of the significant secondary structure occurring in the interesting region can be assessed by repeating the Monte Carlo simulation. The power of the program is demonstrated in the analysis of local secondary structures of human T-cell lymphotrophic virus type III (HIV).

Parallelism description on target architectures: an approach

In an earlier work we have introduced an Architecture Description Language (ADL for short) and its support system implemented in a machine-independent fashion [5]. This system provides facilities for teaching and research in computer architecture: target machines described in ADL are fed into the support system, and the corresponding interpreter programs are automatically generated. These programs reproduced the run-time environment of the target, as specified by the ADL description, including the interpretation of the target code. This paper presents a brief review of the language and the implementation of its support system. Next, the new constructs that have been incorporated into ADL to give an extended version (ADL-X) are described. Through these mechanisms, designers can specify and synchronize target machine activities occuring in parallel. Finally, we show how the new support system is being implemented, and give a description of a pipeline machine, using the new language features.

Implementing Linear Algebra Algorithms for Dense Matrices on a Vector Pipeline Machine

This paper examines common implementations of linear algebra algorithms, such as matrix-vector multiplication, matrix-matrix multiplication and the solution of linear equations. The different versions are examined for efficiency on a computer architecture which uses vector processing and has pipelined instruction execution. By using the advanced architectural features of such machines, one can usually achieve maximum performance, and tremendous improvements in terms of execution speed can be seen over conventional computers.

SÜT SIĞIRCILIĞINDA SÜT SAĞIMININ MEKANİZASYON OLANAKLARI ÜZERİNDE BİR ARAŞTIRMA

OZET Bu arastirmanin amaci, ulkemizde kullamlan yerli ve yabanci yapim sut sagim makineleri ile el ve saltiinda sut verimi i'e sag/m zamal1l arasindaki iliskinin; sut sagim makineleri ile sagimda tuketilen ene~ji miktannin, sut sagim makineleri ile sa.~im slYasll1da yapTIan cesitli isler icin gerekli zaman dilimlerinin saptanmasi, benzerlik ve ayriliklann ortaya konulmasi; cesitli. iip sut sagim makil1t!leri icin ekonomik suru buyuklugunun ortaya cikarilmasidir. Denemeler sirasll7da yerli yapim iki, yabanci yapim yedi sut sagini makinesi ile yukanda siralanan konularda karsilastirmali bir calisma yapilmistir. SUMMARY A Research on the Possibilities for the Mechanisation of Milking for Dairy ttle Farms This study is done to determine the possibilities for the mechanisation of milking and for purpose milking test are conducted with foreign and domestic n;ianufactured milking machines in tbe farms having different number of eattles. This milking tests were done in several trials and it continued at last LO day' periods. In the same farms cows also were milked with hand to compare with machine trials. As a result of trials, energy and time requirements were detetmined and. economieal berd size were cakulated., This study in one of firsi studies conducted in this field in Turkey. The results obtained from this study is summarized below: - Time for the hand milking changes with the person doing the job and tberefore sometimes it is shorter or longier than machine milking time. If milk yield Is high 'milking time is also higb. -When milk yield raises the unit energy requirement decreases Annual energy consumption for machine milking is 21.9-28.5 kWh per cow for directtocan milking machines, 15.6 kWh per cow for. bucket milking machine and 18.855. i kWh per cow for milking pipeline machines epending on· the cluster of ass embly. - When machine milking is used, the. number of milked per hour for the stal1ed barns is 31-56 for direct-to-can milking machines depending on the size of the machines used.

An Algorithm for Solving Linear Recurrence Systems on Parallel and Pipelined Machines

A new algorithm for the solution of linear recurrence systems on parallel or pipelined computers is described. Time bounds, speed-up and efficiency for SIMD and MIMD computers with fixed number of arithmetic elements (AE's), as well as for pipelined computers with fixed number of stages per operation, are obtained. The model of each computer is discussed in detail to explain better performance of the pipelined model. A simple modification in the design of AE's for parallel computers makes parallel model superior.

Pipeline Architecture

Pipelined computer architecture has received considerable attention since the 1960s when the need for faster and more cost-effective systems became critical. The merit of pipelining is that it can help to match the speeds of various subsystems without duplicating the cost of the entire system involved. As technology evolves, faster and cheaper LSI circuits become available, and the future of pipelining, either in a simple or complex form, becomes more promising. This paper reviews the many theoretical considerations and problems behind pipelining, surveying and comparing various representative pipeline machines that operate in either sequential or vector pipeline mode, the practical solutions adopted, and the tradeoffs involved. The performance of a simple pipe, the physical speed limitation, and the control structures for penalty-incurring events are analyzed separately. The problems faced by the system designers are tackled, including buffering, busing structur, branching, and interrupt handling. Aspects of sequential and vector processing are studied. Fundamental advantages of vector processing are unveiled, and additional requirements (costs) are discussed to establish a criterion for the tradeoff between sequential and vector pipeline processing. Finally, two recent machines (the CRAY-1 and the Amdahl 470 V/6 systems) are presented to demonstrate how complex pipeline techniques can be used and how simple but advantageous pipeline concepts can be exploited.

Automatic recognition of vector and parallel operations in a higher level language

A compiler for recognizing statements of a FORTRAN program which are suited for fast execution on a parallel or pipeline machine such as ILLIAC-IV, STAR or ASC is described. The technique employs "interval analysis" to provide flow information to the vector/parallel recognizer. Where profitable the compiler changes scalar variables to subscripted variables. The output of the compiler is an extension to FORTRAN which shows parallel and vector operations explicitly.

Factors affecting the rate of milk flow in pipeline milking machines

A factorial experiment has been carried out to determine the effects of varying the level of vacuum, air admission, and reserve air on milking rate and the amount of strippings, in high pipeline, low pipeline, and large diameter low pipeline milking machines. The low pipeline and the large diameter low pipeline machines were found to milk 10.1 and 12.1 per cent faster, respectively, than the high pipeline machine. The large diameter low pipeline operating without air admission was 6.3-7.5 per cent faster than the low pipeline machine operating with and without air admission. Milking rate was 16.1 per cent faster at 15 than at 12 inches Hg vacuum and about 6 per cent faster at 17 and 5 cubic feet per minute (c.f.m.) reserve air than at no reserve air. The admission of air at the claw piece resulted in 9 per cent faster milking with the high pipeline and 9.5 per cent slower milking with the large diameter low pipeline machine. There was no difference in milking rate in the low pipeline machine with or without air admission. The fastest milking rate was observed with the large diameter low pipeline without air admission. However, this system probably has no real advantage over the low pipeline providing the diameter of milk pipe of the latter is sufficient to cope with high rates of milk flow.