[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

20040602: asynchronous read ahead and write behind on NetCDF

Giang,

>Date: Wed, 2 Jun 2004 14:27:38 -0400
>From: Giang Nong <address@hidden>
>Organization: GFDL/NOAA
>To: Steve Emmerson <address@hidden>
>Subject: Re: asynchronous read ahead and write behind on NetCDF
> Keywords: 200405242114.i4OLENtK012695 netCDF

The above message contained the following:

> I did what you suggested to the file ncio.c and here is the result when I 
> type "make":
> ---------------------------------------------------------------------------
> lsc6> make
> 
> Making `all' in directory /home/gtn/netcdf/netcdf-3.5.1/src/libsrc
> 
>         c89 -c -O -I.  -DNDEBUG attr.c
>         c89 -c -O -I.  -DNDEBUG dim.c
>         c89 -c -O -I.  -DNDEBUG error.c
>         c89 -c -O -I.  -DNDEBUG -DVERSION=`cat ../VERSION` libvers.c
>         c89 -c -O -I.  -DNDEBUG nc.c
>         c89 -c -O -I.  -DNDEBUG ncio.c
> cc-1565 c89: ERROR File = ffio.c, Line = 51
>   The struct "ffc_stat_s" has no field "st_oblksize".
> 
>                 if(sb.st_oblksize > 0)
>                       ^
> 
> cc-1565 c89: ERROR File = ffio.c, Line = 52
>   The struct "ffc_stat_s" has no field "st_oblksize".
> 
>                         return (size_t) sb.st_oblksize;
>                                            ^
> 
> cc-1162 c89: ERROR File = ffio.c, Line = 359
>   Too few arguments in function call.
> 
>         if(ffflush(nciop->fd) < 0)
>                             ^
> 
> cc-1552 c89: WARNING File = ffio.c, Line = 551
>   The variable "ControlString" is set but never used.
> 
>         const char *ControlString;
>                     ^
> 
> cc-1552 c89: WARNING File = ffio.c, Line = 648
>   The variable "ControlString" is set but never used.
> 
>         const char *ControlString;
>                     ^
> 
> 3 errors detected in the compilation of "ncio.c".

This means that the programming interface to the FFIO subsystem is
different under IRIX than it is under UNICOS.  This is not surprising.

It means, however, that a new module (e.g., irix_ffio.c) would have to
be written in order to interface the netCDF library to the IRIX FFIO
subsystem.  The existing "ffio" module could probably be used as a
starting-point.  Regretably, we don't have the resources to do this.

You might try posting a message to the netCDF mailing list
<address@hidden> to see if anyone has written such a
module.

Good luck.

Regards,
Steve Emmerson