[FLASH-USERS] suggestions for FLASH format modification

[Tomasz Plewa]

Sounds good to me. The FLASH format is pretty much established and I 
certainly did not mean a major I/O overhaul (unless there are serious 
performance problems, for example, or limited usability, as in my case).

Tomek
--
Anshu Dubey wrote:
> Tomek and the community,
>
> It is certainly a possibility, and not at all difficult to do. Randy
> tells me that VisIt interface will also need a minor modification.
> What I would like to do is take this opportunity to get a general
> input from the community about what features they'd like to see in
> FLASH IO. From those we can implement and include the feasible ones in
> the next release.
>
> Anshu
>
> On Wed, Apr 29, 2009 at 3:46 PM, Tomasz Plewa <tplewa at fsu.edu> wrote:
>   
>> Dear FLASH Developers -
>>
>> I would like to suggest modifying plotfile format to preserve full (8
>> byte/double precision) accuracy for all floating point data except
>> state/scratch/unk-like variables.
>>
>> This will not add to the size of output files in any significant way but
>> will allow using plotfiles for visualizing models with > 1e6 spatial dynamic
>> range.
>>
>> Clearly, such a modification would require some (hopefully equally minor)
>> changes to data analysis/visualization tools  (FLASH plugin reader for
>> VisIt, parts of fidlr). To allow for format compatibility, the accuracy of
>> non-state data in the plotfile output could be controlled by either a
>> configuration flag or additional runtime parameter.
>>
>> Tomek
>> --
>>
>>
>>     
-------------- next part --------------
A non-text attachment was scrubbed...
Name: tplewa.vcf
Type: text/x-vcard
Size: 338 bytes
Desc: not available
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20090501/9d4d0aad/attachment.vcf>