Results on Fujitsu VPP5000

From: Pierre Lagier (
Date: Wed Jan 31 2001 - 20:30:30 CST

  • Next message: Scott Carter: "STREAM result"

    Hi John,

    I just found no STREAM benchmark results were posted for the Fujitsu VPP5000 machine, right ?. As long as it is truely a supercomputer (peak > 1TFlops for 128 processors) with a pure distributed memory architecture I thought it could be interesting to measure it. So here are the results. I am not sure why Fujitsu Japan never released the numbers, but anyway anyone having access to such monster can easily run the test, so I did. Results are quite good I guess, feel free to publish them - well, hummm, seems to be one of the best single processor result so far...!


    - Results for STREAM on Fujitsu VPP5000 single processor, Thu Feb 1 11:22:45

    bash-2.03$ uname -p -s -v -m
    UNIX_System_V 3 5000 UXP/V
    bash-2.03$ frt -V
    frt : Fujitsu UXP/V Fortran V20L20 Driver L00091 (Nov 20 2000 18:54:41)
    bash-2.03$ frt -Of stream_d.f
    bash-2.03$ ./a.out -Wl,-Lr
    Double precision appears to have 16 digits of accuracy
    Assuming 8 bytes per DOUBLE PRECISION word ----------------------------------------------
    Array size = 8000000
    Offset = 0
    The total memory requirement is 183 MB
    You are running each test 10 times

    The *best* time for each test is used
    *EXCLUDING* the first and last iterations  ----------------------------------------------------
    Your clock granularity/precision appears to be      1 microseconds  ----------------------------------------------------
    Function     Rate (MB/s)  Avg time   Min time  Max time
    Copy:      37780.4014      0.0034      0.0034      0.0035
    Scale:     35724.2534      0.0036      0.0036      0.0036
    Add:       34594.5946      0.0056      0.0055      0.0056
    Triad:     37543.9969      0.0051      0.0051      0.0052  ----------------------------------------------------
    Solution Validates!
    ---------------------------------------------------- bash-2.03$


    This archive was generated by hypermail 2b29 : Mon Apr 23 2001 - 09:29:53 CDT