[Wrf-users] Problems with PGI 6.1-3
Brent Shaw
brent.shaw at wni.com
Mon Mar 6 10:22:05 MST 2006
All,
We installed the latest patch of PGI (6.1-3). While this patch appears
to have corrected the compiler bug affecting real.exe we originally
reported, I am now finding other problems. Specifically, the wrfbdy_d01
file is not correct, and the differences are severe. Running the Jun
2001 NCAR test case with WRF/real compiled with PGI 6.1-3 on an AMD
system results in CFL's beginning around 21 hours into the simulation,
and the results are drastically different.
On a positive note, whatever this "bug" is only seems to affect
real.exe, and within that, only the wrfbdy file. I was able to
successfully run the June 2001 case using the WRF code compiled with
6.1-3, but using the input files created using real.exe compiled with
the older version of PGI. The results were bit-for-bit identical to the
successful WRF simulation using the older compiler. Next, I wrote a
couple of quick programs to compare the contents of the wrfinput_d01 and
wrfbdy_d01 created with the two versions of PGI. I found that the
wrfinput_d01 file was identical in both cases, but the wrfbdy_d01 file
has signficant differences (I am including the print-out of the
wrfbdy_d01 differences below).
I have forwarded these results to PGI.
As with my last posts, if anyone else is able to verify my results, it
would be very useful!
Regards,
Brent
Differences
File 1: 6.0-5/wrfbdy_d01
File 2: 6.1-3/wrfbdy_d01
Differences (file2 - file1):
Variable Name Min Diff Max Diff Mean Diff
--------------- --------------- --------------- ---------------
U_BXS -2308123.500 1888424.000 -179814.391
U_BXE -1757943.250 2628634.750 251051.719
U_BYS -1866978.250 2710279.000 254768.547
U_BYE -2352703.500 1913744.500 -153818.188
U_BTXS -174.854 213.715 16.650
U_BTXE -243.392 162.773 -23.246
U_BTYS -250.952 172.868 -23.590
U_BTYE -177.199 217.843 14.242
V_BXS -3763931.500 1964166.250 -425418.594
V_BXE -3828386.250 1643290.750 -363339.000
V_BYS -1911013.000 953464.750 -283464.719
V_BYE -5068687.000 1407303.000 -661669.438
V_BTXS -181.867 348.512 39.390
V_BTXE -152.157 354.480 33.643
V_BTYS -88.284 176.946 26.247
V_BTYE -130.306 469.323 61.266
W_BXS 0.000 0.000 0.000
W_BXE 0.000 0.000 0.000
W_BYS 0.000 0.000 0.000
W_BYE 0.000 0.000 0.000
W_BTXS 0.000 0.000 0.000
W_BTXE 0.000 0.000 0.000
W_BTYS 0.000 0.000 0.000
W_BTYE 0.000 0.000 0.000
PH_BXS -112365440.000 63728560.000 -15131474.000
PH_BXE -189060464.000 5540736.000 -35758856.000
PH_BYS -111303744.000 8476784.000 -25023122.000
PH_BYE -187797792.000 63728560.000 -24498352.000
PH_BTXS -5900.792 10404.207 1401.057
PH_BTXE -513.031 17505.600 3310.984
PH_BTYS -784.887 10305.902 2316.964
PH_BTYE -5900.792 17388.684 2268.360
T_BXS -1258745.500 815136.000 -74470.648
T_BXE -1028473.625 1588185.500 -127469.625
T_BYS -1146616.875 931096.500 -124050.984
T_BYE -1007809.625 1647879.000 -92101.492
T_BTXS -75.476 116.551 6.895
T_BTXE -147.054 95.229 11.803
T_BTYS -86.213 106.168 11.486
T_BTYE -152.581 93.316 8.528
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/wrf-users/attachments/20060306/b3d4ca34/attachment.html
More information about the Wrf-users
mailing list