19990909: Sun SC5.0 problems (cont.)

>From: weather <address@hidden>
>Organization: NMSU/NSBF
>Keywords: 199909091440.IAA28776 McIDAS-X GRDDISP Sun SC5.0 f77

Robert,

>Thanks, I am going there now...whoops I can't change to the
>SC5.0 directory, permissions are set to 770.

Oops, sorry.  I just changed the permissions to 775.

>I am also installing the compiler patches which I needed to
>anyway, maybe that is why mine compiles slowly.  I noticed
>there is a bug fix for slow compiling for the C++ compiler.

OK.

>22 minutes to compile McIDAS-X and XCD on the 400Mhz PII
>with SC4.2 and 54 minutes to compile on the 333Mhz Ultrasparc
>with SC5.0.

This is a huge difference.

>The PII is probably faster for this sort of stuff 
>but I suspect the difference in compilers is part of
>it as well.  Actually running, decoding, web images, etc
>are all faster on the Ultrasparc.

OK.

>That is good news about prices on PC's, I did not realize
>they have dropped so much.  Building one from parts is the
>way to go with Solaris because of the limited hardware
>support.  I did put RedHat 6.0 (along with NT, Solaris, and DOS)
>on my machine at home, but it sounds like 6.0 is not
>compatible yet with GEMPAK and McIDAS.  I figure I need to 
>learn more about it, even though I much prefer Solaris.

My experience with RedHat 6.0 was that:

o I had to update the netCDF files; SSEC had included a beta release
  of netCDF 3.4 which wouldn't compile on RedHat 6.0.  The official
  3.4 release compiles fine

o Tcl code (in tcl/generic) wouldn't compile.  The error was:

/usr/bin/cc -c -O  -fPIC -I./../generic -I. -DHAVE_GETCWD=1 -DHAVE_UNISTD_H=1 
-DUSE_TERMIOS=1 -DHAVE_SYS_TIME_H=1 -DTIME_WITH_SYS_TIME=1 -DHAVE_TM_ZONE=1 
-DHAVE_TM_GMTOFF=1 -DHAVE_TIMEZONE_VAR=1 -DHAVE_ST_BLKSIZE=1 -DSTDC_HEADERS=1 
-DNEED_MATHERR=1 -DRETSIGTYPE=void -DHAVE_SIGNED_CHAR=1 -DHAVE_SYS_IOCTL_H=1    
    -DTCL_SHLIB_EXT=\".so\" ./../generic/tclPosixStr.c
./../generic/tclPosixStr.c: In function `Tcl_ErrnoId':
./../generic/tclPosixStr.c:340: duplicate case value
./../generic/tclPosixStr.c:328: this is the first entry for that value
./../generic/tclPosixStr.c: In function `Tcl_ErrnoMsg':
./../generic/tclPosixStr.c:787: duplicate case value
./../generic/tclPosixStr.c:775: this is the first entry for that value
make[1]: *** [tclPosixStr.o] Error 1
make[1]: Leaving directory `/home/mcidas/mcidas7.6/tcl/unix'
make: *** [.built_tcl] Error 2

I havn't had time to figure this one out yet.  Perhaps it is an
easy one.  If so, and if the rest compiles, then I can give my blessing
for McIDAS at least.  Chiz has created a binary for GEMPAK and GARP
so 6.0 users can get running.

>From address@hidden  Thu Sep  9 10:52:37 1999

>Since as I noted in my previous message I could not get to
>the SC5.0 dir, I tried to remake grddisp.k after applying all of
>the latest patches.  The way I test it is to go to my
>Ottumwa web page and run the model data CGI script..it 
>worked.  I checked the CGI log this AM when it did not work
>and it said segmentation violation,,there is no error
>message now and an image is produced. GRDCOPY works now as well.

OK, I will have to relay this information to our system adminstrator.

>Here are the patches I installed:
>107596-02
>107357-04
>107311-05
>107989-01
>107289-03

>I don't know which one did the trick, but one apparently
>did.  I suspect it was 107357-04.

>Hope this helps.

As always, thanks for the information!

Tom

>From address@hidden  Thu Sep  9 13:26:58 1999

I recompiled all of McIDAS-X and XCD after applying the patches.

I looked back at the compile times...28 minutes for Intel/SC4.2
49 minutes for Ultrasparc/SC5.0 pre-patches and 42 minutes
for Ultrasparc/SC5.0 post-patches!

Robert
NSBF Meteorology