Stream results

From: Wayne Preston Allen (wallen@io.com)
Date: Sat Dec 11 2004 - 10:51:29 CST

  • Next message: Nahil Sobh: "New World Record 512p STREAM result on NCSA's Altix"

    Thanks for this great little benchmark. It helped me verify that I was
    actually getting dual-channel operation on my upgraded system.
    Wayne P. Allen <http://www.io.com/%7Ewallen/>

    Athlon XP 3000+ (400FSB) - $142
    Asus A7N8X-E (NVIDIA nForce2) - $104
    2x512 PC-3200 (DDR400) SpecTe <http://www.spectek.com/default.asp>c
    PD64M6408T27BD2T-6A <http://www.spectek.com/pdfs/64_128MX64JDDR.PDF>
    (in dual-channel mode) - $86x2
    Seagate Baracuda IV 80Gig ATA-100 - priceless! Uh, $85
    Linux, Fedora Core 3 - actually priceless!
    gcc 3.4.2 -O2 (-O3 not much different)
    ------------------------------------------------------------
    This system uses 8 bytes per DOUBLE PRECISION word.
    -------------------------------------------------------------
    Array size = 10000000, Offset = 0
    Total memory required = *228.9 MB.*
    Each test is run 10 times, but only
    the *best* time for each is used.
    -------------------------------------------------------------
    Your clock granularity/precision appears to be 2 microseconds.
    Each test below will take on the order of 86669 microseconds.
       (= 43334 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: 1377.5404 0.1167 0.1161 0.1175
    Scale: 1312.1628 0.1224 0.1219 0.1235
    Add: 1481.8574 0.1626 0.1620 0.1637
    Triad: 1612.6658 0.1495 0.1488 0.1505



    This archive was generated by hypermail 2.1.4 : Sat Dec 25 2004 - 08:35:14 CST