[prev in list] [next in list] [prev in thread] [next in thread] 

List:       openjdk-openjfx-dev
Subject:    status of 3D support on Linux
From:       chien.yang () oracle ! com (Chien Yang)
Date:       2012-10-31 16:10:21
Message-ID: 50914D6D.6060009 () oracle ! com
[Download RAW message or body]

Hi Udo,

     Thank you for your understanding. You can bypass JavaFX GPU 
blacklist checking mechanism by specifying the following property at 
start up:
-Dprism.forceGPU=false

     It will be great if you have the resources, in-house, to evaluate a 
particular GPU that interest you. We may be able to assist you further 
by taking it off the blacklist, if it passes your internal evaluation 
process. Of course, this will include passing our internal SQE process too.

Thanks,
- Chien


On 10/31/12 8:51 AM, Udo Rader wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Yes, sure, nobody wants crashing applications.
>
> OTOH there is no way to know if an application crashes until someone
> can at least experimentally test it.
>
> For example we are currently trying to evaluate if JavaFX might be a
> good choice for a new application running on embedded scale devices,
> all driven by Intel 3600 GMA GPUs.
>
> On the one hand, none of our Linux developers are running at least the
> proprietary fglrx drivers for massive instability reasons (yes:
> artifacts, memory leaks, ...) and then additionally we find that Intel
> GPUs are completely unsupported in Linux.
>
> So is there a way to at least override the JavaFX GPU blacklist (like
> the system setting Kevin R mentioned)?
>
> And in order to not only complain and whine about how things are: what
> were be the required steps to improve (at least inofficial) 3D support
> for JavaFX in Linux? Can "interested parties" improve the situation?
>
> Thanks
>
> Udo
>
> On 10/31/2012 04:04 PM, Chien Yang wrote:
>> Hi Udo,
>>
>> We are monitoring the GPU driver progress, both open source and
>> proprietary versions, from the various vendors. That is why we
>> decided to add ATI's fglrx driver to our supported platform after
>> it has passed our internal evaluation. We want to ensure good user
>> experience of our JavaFX library. The goal is to avoid rendering
>> artifacts and hard crashes that are beyond our control.
>>
>> Thanks, - Chien
>>
>> On 10/31/12 1:42 AM, Udo Rader wrote: On 10/30/2012 11:54 PM, Kevin
>> Rushforth wrote:
>>>>>> #2 what about other GPUs, such as the extremely common
>>>>>> Intel HD GPUs?
>>>>> Unplanned for FX 8. The problem is that the Intel OpenGL
>>>>> drivers on Linux have been historically very bad in terms of
>>>>> quality and compliance. There is some anecdotal evidence that
>>>>> this is changing, but we don't want to support it. There is a
>>>>> system property you can set to force this for testing, but no
>>>>> way to deploy an application using it.
>> Yes, the intel drivers have a bad history in some areas. OTOH, you
>> might say the same about general usability of the proprietary
>> drivers under Linux (ie the fglrx driver).
>>
>>>>>> #3 any plans to support OpenSource/Gallium3D based drivers
>>>>>> (ie nouveau/radeon)?
>>>>> Not support, no. But it might "just work" and is something
>>>>> you could try.
>> I understand, "support" is a somewhat dangerous term, of course.
>>
>> How then do we force things on unsupported platforms? Right now,
>> 3D simply does not work and we did not find anything about the
>> system property you mention.
>>
>> Thanks
>>
>> Udo
>>
>> -- Udo Rader, CEO BestSolution.at EDV Systemhaus GmbH
>> Eduard-Bodem-Gasse 5-7, A-6020 Innsbruck
>> http://www.bestsolution.at/ Reg. Nr. FN 222302s am
>> Firmenbuchgericht Innsbruck
>>
> - -- 
> Udo Rader, CEO
> BestSolution.at EDV Systemhaus GmbH
> Eduard-Bodem-Gasse 5-7, A-6020 Innsbruck
> http://www.bestsolution.at/
> Reg. Nr. FN 222302s am Firmenbuchgericht Innsbruck
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.12 (GNU/Linux)
> Comment: Using GnuPG with Mozilla - http://www.enigmail.net/
>
> iQEcBAEBAgAGBQJQkUjlAAoJEJA9QoEqa9WBeBkH/3xCT4WC6HDSTgBm/s9UAEm9
> 3ko3SPol5QFJ4mehEmB0XOSQYjYlOQwG09E3sSlmNArMfmbdeDgwiK2hghYsUzvj
> Klwc1Qk5r+f4gcYlJJGpE1asnWdV1Ky7TiwTm1EbP3ZT/BERAFDo8wfcQwiHIsOG
> t/6MQzsq3mCjH+cIDMT5jxdAqqfMYWNcV6z1uviH5LAFs8iw4WHPh44g58Ve98Na
> ca+BXljRvKaSODOEQCtXxgVe6zBNzQB3oid15rPnBWJDnwpYhzFxpfCn3P2rBJVq
> bidcG/k5lAR19WKQTf99pbThFWmm84Vp7TbtmcOYSLOEGr9FmOFJtACoCmFi8aY=
> =WHw3
> -----END PGP SIGNATURE-----


[prev in list] [next in list] [prev in thread] [next in thread]