Task granularity studies on a many-processor CRAY X-MP

نویسنده

  • Donald A. Calahan
چکیده

A hybrid granularity model is proposed for general concurrent solution. It is applied to the triangular factorization of a dense matrix ranging in size from 4 to 1024. Concurrency is achieved at two levels: (!) with small (micro) task granularity and (2) with large (blocked) task granularity. Rdevano~ to a many-proccssax CRAY X-MP is demonstrated by simulation.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Multiprocessing linear algebra algorithms on the CRAY X-MP-2: Experiences with small granularity

“Multiprocessor” is a term that has been used for years. Our definition follows those of [8], [9], and [lo]. The CRAY X-MP is a follow-up to the CRAY-1S system offered by CRAY Research, Inc. The CRAY X-MP family is a general-purpose multiprocessor systeln. It inherits the basic vector functions of CRAY-lS, with major architectural improvements for each individual processor. The interprocessor c...

متن کامل

Reduce Conflicts between Vector Streams in Complex Memory Systems

The performance of a vector processor accessing vectors is strongly dependent on the conflicts produced in the memory subsystem. These conflicts delay the job of the functional units. Conflicts named linked conflict and complex conflicts interfere the steady-state performance in complex memory systems (CRAY X-MP).This paper proposes a method to access vector streams that reduces the average mem...

متن کامل

Implementation of the LISP-Arbitrary Precision Arithmetic for a Vector Processor

Portable Standard LISP (PSL, Version 3.4) and REDUCE 3 were implemented for Cray 1 and Cray X-MP computers at the Konrad Zuse-Zentrum Berlin in 1986. As an special aspect of the implementation of PSL, an interface to the vector hardware of Cray processors was defined. With that interface and mostly driven by the needs of REDUCE applications (e.g. extensive calculations of Gröbner bases), the ar...

متن کامل

A New Algorithm for the SLALOMBenchmark 1

We apply a preconditioned conjugate gradient algorithm to the SLALOM benchmark, proving that the solution phase of the benchmark is reduced to O(N 2). The algorithm therefore performs well over the full range of input parameters speciied by the benchmark. This has a dramatic impact on the benchmark for all computers. We illustrate this with new implementations on an 8192 processor MP-1 system a...

متن کامل

Neun Portable Standard LISP for Cray X - MP Computers Preprint SC 86 - 2 ( Dezember 1986 )

Portable Standard LISP (PSL) is a portable implementation of the programming language LISP constructed at the University of Utah. The version 3.4 of PSL was implemented for Cray X-MP computers by KonradZuse-Zentrum Berlin; this implementation is based to an important part on the earlier implementation of PSL 3.2 at the University of Utah, Los Alamos National Laboratories and Cray Research Inc. ...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:
  • Parallel Computing

دوره 2  شماره 

صفحات  -

تاریخ انتشار 1985