[Sidefx-houdini-list] bgeo.gz and gzip with windows

paul simpson paul at realisestudio.com
Fri Jun 20 10:54:24 EDT 2008


hey chris - a very useful table for sure!  thx for sharing.  i hope
that sesi are listening in and maybe can include lzma in the default
houdini distro.  btw, how did they compare with (de)compression times?

2008/6/20 Edward Lam <edward at sidefx.com>:
> 7zip is more like a "container format" where the default compression is
> LZMA.
>
> -Edward
>
> jiversen wrote:
>> Hi there,
>>
>> How does 7zip perform?  I think it's pretty slow on creating one but not
>> bad on decompression (where it's most important).
>>
>>
>> Chris AtLee wrote:
>>> Pablo Giménez wrote:
>>>
>>>> 2008/6/17 Nicholas Pliatsikas <nick_p102 at hotmail.com>:
>>>>
>>>>
>>>>> Does anyone know the full benefits of using bgeo.gz? I.e I know the files
>>>>> sizes are smaller, meaning less network traffic and uses less space on our
>>>>> servers. But would complex setups with geos writing in houdini or delayed
>>>>> read archives rendering through mantra require more time to process?
>>>>> Just so we can access if its worth looking into this further at the moment.
>>>>> Cheers
>>>>> Nicholas Pliatsikas
>>>>>
>>>> Yes, if you are dealing with big geometry files using gz is really good, but
>>>> in my experience bzip2 is really good, I have seen ratios very clos to 1:8
>>>> in the size, bgeo is binary so bzip2 works much better on this kind of
>>>> files.
>>>> I think there is a version of bzip2  for windows.
>>>> The only problem I have experienced using compressed files is rendering from
>>>> file, but this was on version 8, I don´t know if this have been solved in
>>>> version 9.
>>>>
>>> LZMA is also an option.  It performs significantly better than bz2, and
>>> is also faster to decompress than bz2.
>>>
>>> I ran a test against about 7000 bgeo files here, here's a summary of my
>>> results:
>>>
>>> Compression Method      gzip     bzip2     lzma
>>> Average Compression
>>>   (% of original size)  45.74    47.78     34.76
>>> Best compression ratio
>>>   (% of original size)   7.22     4.50      2.20
>>> Worst compression ratio
>>>   (% of original size) 136.00   137.14     93.50
>>> Standard Deviation       0.18     0.23      0.17
>>>
>>> So, for the files that I was testing, bz2 is in fact not a good choice
>>> to use.  It performs worse on average than gzip and has more varied
>>> behaviour.  lzma is a clear winner here.  I didn't measure the
>>> compression times for the different methods here, but other reports
>>> online indicate that lzma is slower than gzip, but faster than bz2.
>>>
>>> I know lzma is available for Linux, and it looks like the LZMA SDK
>>> includes lzma.exe for Windows, but I haven't tested it.
>>>
>>> Cheers,
>>> Chris
>>> _______________________________________________
>>> Sidefx-houdini-list mailing list
>>> Sidefx-houdini-list at sidefx.com
>>> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>>>
>> _______________________________________________
>> Sidefx-houdini-list mailing list
>> Sidefx-houdini-list at sidefx.com
>> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>
> _______________________________________________
> Sidefx-houdini-list mailing list
> Sidefx-houdini-list at sidefx.com
> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>



More information about the Sidefx-houdini-list mailing list