P4_1500 results

From: Herbert Huber (Herbert.Huber@lrz-muenchen.de)
Date: Mon Feb 26 2001 - 03:30:39 CST

  • Next message: Winfrid Tschiedel: "Stream V5 - OS Redhat 7.0/Windows 2000, PGI pgf90 3.2"

    Hello John,

    I have run your benchmark on our Dell 8100 (P4 1500 MHz, 1 GByte
    Dual-Channel Rambus-DRAM RIMM 800) linux box, running kernel 2.4.0.

    PGI FORTRAN-77 compiler (pgf77 3.2-3):
    pgf77 -fast -Mvect=sse -Mcache_aline -Mdalign -Munroll -Minline

    No P4 optimizations! --pgf77 version 3.2-3 supports only PIII and Athlon
    spezific optimizations.

      Double precision appears to have 16 digits of accuracy
      Assuming 8 bytes per DOUBLE PRECISION word
     ----------------------------------------------
      Array size = 25012345
      Offset = 0
      The total memory requirement is 572 MB
      You are running each test 10 times
      The *best* time for each test is used
      ----------------------------------------------------
      Your clock granularity/precision appears to be 10000 microseconds
      The tests below will each take a time on the order
      of 200000 microseconds
         (= 20 clock ticks)
      Increase the size of the arrays if this shows that
      you are not getting at least 20 clock ticks per test.
      ----------------------------------------------------
      WARNING -- The above is only a rough guideline.
      For best results, please be sure you know the
      precision of your system timer.
      ----------------------------------------------------
     Function Rate (MB/s) RMS time Min time Max time
     Copy: 2106.3028 0.1941 0.1900 0.2000
     Scale: 2106.3028 0.2203 0.1900 0.3700
     Add: 2143.9153 0.2850 0.2800 0.2900
     Triad: 2143.9153 0.2830 0.2800 0.2900
      Sum of a is = 2.8846757034495990E+019
      Sum of b is = 5.7693514069685250E+018
      Sum of c is = 7.6924685388864070E+018

    Regards

    /Herbert



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