Abstract

The results of benchmark tests on the superscalar IBM RISC System/6000 Model 590 are presented. A set of well‐characterized Fortran benchmarks spanning a range of computational characteristics was used for the study. The data from the 590 system are compared with those from a single‐processor CRAY C90 system as well as with other microprocessor‐based systems, such as the Digital Equipment Corporation AXP 3000/500X and the Hewlett‐Packard HP/735.

Highlights

  • The IBM RISC System/6000, first introduced in 1990, was an important step in the development of high-performance microprocessor systems

  • The Digital Equipment Corp. (DEC), HP, and CRAY systems included in this article for comparison are all older than the RIOS-2, and it is expected that follow-on products to each will be made available soon

  • The superscalar nature of the RIOS-2 allows it to perform better than microprocessors operating at three times its central processing unit (CPC) clock speed

Read more

Summary

INTRODUCTION

The IBM RISC System/6000, first introduced in 1990, was an important step in the development of high-performance microprocessor systems. The data bus that feeds both floating point pipelines from the data cache is four 64-bit words wide and the compiler can issue "loadquadword" instructions that will cause two 64-bit operands to be loaded from memorv. Each of the two floating point units in the RIOS-2 performs IEEE standard arithmetic and each has thirty-two 64-bit general purpose registers. The Alpha microprocessor is a full 64-bit implementation and is dualissue superscalar, but it cannot overlap floating point additions and multiplies, so its theoretical peak performance is 200 MFLOPS. The Hewlett Packard HP735 workstation, at least the second generation of the HP "Precision Architecture,'' is superscalar and can overlap adds and multiplies [6] It is currently available at a CPU speed of 99 MHz, which al~o corresponds ·to about 200 MFLOPS peak. The DEC, HP, and CRAY systems included in this article for comparison are all older than the RIOS-2, and it is expected that follow-on products to each will be made available soon

Benchmark Codes
Performance on Elementary Vector Operations
Store Float Double 2 Load Float Double 2 Branch to LABEL
Store Float Quad 2 Branch to LABEL
Results Using Application Benchmarks
Comparison of RIOS-2 and RIOS-1
Comparison with Other Workstations
CP 30 ns 2 CP 20 ns 2 CP 30 ns 6 CP 25 ns
Comparison with a Single Processor of the CRAY C90
DISCUSSION
Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call

Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.