It's some debian release, compiled for amd-64 architecture, kernel
version is 2.6.21.4, writing on an nfs. hdf5 is serial.
It could be something specific to this latest compile (I've never seen
this before), rather than a bug. I'll try a few things out, and send
another message if I learn anything new. I can try the same thing out
on a different installation to make sure it's just a fluke.
Best,
Matt
On Tue, 29 Jan 2008, Steven G. Johnson wrote:
> On Tue, 29 Jan 2008, matt wrote:
>> I noticed some strange behavior in output-png+h5 with meep-mpi 0.10.1
>>
>> The png output is incorrect has white patches blocking the fields. If I run
>> the h5topng -Zc dkbluered on the h5 ouput, the resuling png appears correct
>> without the white patches. Non-parallel meep does not exhibit this behavior.
>
> The only thing I can think of is that there is some sort of race condition
> in the filesystem. The processors write their data to the h5 file in
> chunks, and maybe somehow the filesystem has not synchronized the file
> before h5topng reads it in?
>
> What operating system and filesystem are you using? Are you using HDF5
> compiled with MPI support? (Otherwise, Meep uses some hacks to get
> serialized parallel I/O.)
>
> Steven
>
> _______________________________________________
> meep-discuss mailing list
> [email protected]
> http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
>
_______________________________________________
meep-discuss mailing list
[email protected]
http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
