2002.03.14_lanucara

来自「大名鼎鼎的海洋数值计算模型-pom模型源代码及说明书和示例」· 14_LANUCARA 代码 · 共 49 行

Dear Dr. Ezer,

here there are some benchmark data about pom2k on different serial machines; 
We hope that these data can be useful for the Pom's community.

Sincerely

Piero Lanucara (Caspur www.caspur.it)
Giorgio Amati (Caspur)
M.Pia Busico (Caspur)
Vincenzo Artale (Enea www.enea.it)
Gianmaria Sannino (Enea)
*********************************************************************
Total elapsed time (in sec) for 3 different data set of different
processors

Simulation parameters:

days : 4.0
prtd1 : 1.0
iisplit : 30
dte : 6.0
iproblem: 1 = seamount
-----------------------------------------------

Test case A: 65*49*21 grid
IBM Power3@375 Mhz 849.44
Compaq Alpha EV6@500 Mhz 632.75
Intel PIV@1300 Mhz 674.07

Test case B: 65*49*31 grid
IBM Power3@375 Mhz 1338.02
Compaq Alpha EV67@667 Mhz 959.21
Intel PIV@1300 Mhz 799.44

Test case C: 130*98*21 grid
IBM Power3@375 Mhz 4290.47
Compaq Alpha EV67@667 Mhz 2949.07
Intel PIV@1300 Mhz 2485.82


Compiler options:
IBM: xlf -qnosave -qautodbl=dbl4 -O3 -qtune=pwr3 -qarch=pwr3
Intel: pgf77 -i4 -r8 -fast -O2 -tp p6
Compaq: f77 -real_size 64 -O5 -tune host -arch host
-math_library fast -convert big_endian
-align dcommon -transform_loop -inline speed