chromium-users — Mailing list for all users of the Chromuim rendering system

Re: [Chromium-users] Spew SPU

[Stephen F. <ste...@gm...>]

Broadcast is not performing as expected...
I've got the config:

App[Tilesort - Broadcast] --> Server[Render - toAppWindow]
   |
  V
Server[Tilesort - Test all Tiles] --> 2x Server[Render - toVideoWall]

But the program never starts.  And if I set any render or tile size larger
than 640x480 it complains that it can't "fit" it in the tile.  I've tried
setting everything under the Config.py program, but I can't get this to
work.

-S!

2009/9/8 Sean Ahern <ah...@or...>

> Stephen Furlani wrote:
> > Okay, so I want to make a "Spew" SPU that takes the commands and sends
> > identical commands to all of it's nodes.
> >
> > ... I have to write a function for each individual OpenGL command!?
> >  SPUs like Passthrough and Pack have 600+ commands...
>
> Why not just run the existing Tilesort SPU in 'Broadcast' mode?  It
> sends all primitives to all downstream servers.
>
> See "bucket_mode" at
> http://chromium.sourceforge.net/doc/configoptions.html#tilesort
>
> -Sean
>
> __
> Sean Ahern
> Oak Ridge National Laboratory
> 865-241-3748
> AIM: ornlsean
>



-- 
-Stephen Furlani
*I'm moving to a new email*
ste...@gm...
Please update your address books, thanks!

Re: [Chromium-users] Spew SPU

[Sean A. <ah...@or...>]

Stephen Furlani wrote:
> Okay, so I want to make a "Spew" SPU that takes the commands and sends 
> identical commands to all of it's nodes.
> 
> ... I have to write a function for each individual OpenGL command!? 
>  SPUs like Passthrough and Pack have 600+ commands...

Why not just run the existing Tilesort SPU in 'Broadcast' mode?  It
sends all primitives to all downstream servers.

See "bucket_mode" at
http://chromium.sourceforge.net/doc/configoptions.html#tilesort

-Sean

__
Sean Ahern
Oak Ridge National Laboratory
865-241-3748
AIM: ornlsean

[Chromium-users] Spew SPU

[Stephen F. <ste...@gm...>]

Okay, so I want to make a "Spew" SPU that takes the commands and sends
identical commands to all of it's nodes.
... I have to write a function for each individual OpenGL command!?  SPUs
like Passthrough and Pack have 600+ commands...

-S!

2009/9/6 daniel <zd...@16...>

>  I have paid some time to search such SPU, but failed. I think maybe you
> have to write one by yourself.
>
>  ------------------------------
>  daniel
> 2009-09-07
>  ------------------------------
>  *发件人：* Stephen Furlani
> *发送时间：* 2009-09-05 06:27:04
> *收件人：* chr...@li...
> *抄送：*
> *主题：* Re: [Chromium-users] Echo Graphics
>
> Hrmmm... it's not on the sourceforge documentation.  The only that's close
> to it in meaning is Replicate SPU but that's for VNC.
> -S!
>
> On Fri, Sep 4, 2009 at 5:07 PM, Chris Niederauer <cc...@cc...> wrote:
>
>> Although I haven't looked at this recently, I believe there is a split spu
>> or something similar that you can stick in before the tilesort spu to
>> achieve this.  Reading back from a video wall, however, would present other
>> challenges if there is anything occurring post-tilesort.
>>
>> Chris Niederauer
>>
>>
>> On Sep 4, 2009, at 6:06 AM, Stephen Furlani wrote:
>>
>> Hello,
>>>
>>> Can I get Chromium to display the OpenGl in the App's native window in
>>> addition to sort-first on my video wall?
>>>
>>
>>
>
>
> --
> -Stephen Furlani
> *I'm moving to a new email*
> ste...@gm...
> Please update your address books, thanks!
>



-- 
-Stephen Furlani
*I'm moving to a new email*
ste...@gm...
Please update your address books, thanks!

Re: [Chromium-users] Echo Graphics

[daniel <zd...@16...>]

I have paid some time to search such SPU, but failed. I think maybe you have to write one by yourself.




daniel
2009-09-07



发件人： Stephen Furlani
发送时间： 2009-09-05 06:27:04
收件人： chr...@li...
抄送： 
主题： Re: [Chromium-users] Echo Graphics

Hrmmm... it's not on the sourceforge documentation.  The only that's close to it in meaning is Replicate SPU but that's for VNC.


-S!


On Fri, Sep 4, 2009 at 5:07 PM, Chris Niederauer <cc...@cc...> wrote:

Although I haven't looked at this recently, I believe there is a split spu or something similar that you can stick in before the tilesort spu to achieve this.  Reading back from a video wall, however, would present other challenges if there is anything occurring post-tilesort.

Chris Niederauer


On Sep 4, 2009, at 6:06 AM, Stephen Furlani wrote:


Hello,

Can I get Chromium to display the OpenGl in the App's native window in addition to sort-first on my video wall?






-- 
-Stephen Furlani
*I'm moving to a new email*
ste...@gm...
Please update your address books, thanks!

Re: [Chromium-users] Echo Graphics

[Stephen F. <ste...@gm...>]

Hrmmm... it's not on the sourceforge documentation.  The only that's close
to it in meaning is Replicate SPU but that's for VNC.
-S!

On Fri, Sep 4, 2009 at 5:07 PM, Chris Niederauer <cc...@cc...> wrote:

> Although I haven't looked at this recently, I believe there is a split spu
> or something similar that you can stick in before the tilesort spu to
> achieve this.  Reading back from a video wall, however, would present other
> challenges if there is anything occurring post-tilesort.
>
> Chris Niederauer
>
>
> On Sep 4, 2009, at 6:06 AM, Stephen Furlani wrote:
>
>  Hello,
>>
>> Can I get Chromium to display the OpenGl in the App's native window in
>> addition to sort-first on my video wall?
>>
>
>


-- 
-Stephen Furlani
*I'm moving to a new email*
ste...@gm...
Please update your address books, thanks!

[Chromium-users] Multi - TileSort

[Stephen F. <ste...@gm...>]

Hello,
I'm trying to get the MultiTileSort example to work on my machine.

I've got the simplemural example running on my video wall.

The MultiTileSort example displays the stereocube app in the App's Window on
Localhost - but does not display anything on my wall.

there is also no error information, and the crServers on my Tiles are giving
me the CR Info() total output dims = x,y.  So it looks like it's hooking up
correctly.

Help!

-- 
-Stephen Furlani
*I'm moving to a new email*
ste...@gm...
Please update your address books, thanks!

[Chromium-users] Echo Graphics

[Stephen F. <ste...@gm...>]

Hello,
Can I get Chromium to display the OpenGl in the App's native window *in
addition* to sort-first on my video wall?

-- 
-Stephen Furlani
*I'm moving to a new email*
ste...@gm...
Please update your address books, thanks!

[Chromium-users] Fw: RE: crappfaker hangs, please help

[Donhu R. <don...@ya...>]

libGL error: open DRM failed (Operation not permitted)
libGL error: reverting to (slow) indirect rendering
libGL error: open DRM failed (Operation not permitted)
libGL error: reverting to (slow) indirect rendering
libGL error: open DRM failed (Operation not permitted)
libGL error: reverting to (slow) indirect rendering

Re: [Chromium-users] crappfaker hangs, please help

[Donhu R. <don...@ya...>]

libGL error: open DRM failed (Operation not permitted)
libGL error: reverting to (slow) indirect rendering
libGL error: open DRM failed (Operation not permitted)
libGL error: reverting to (slow) indirect rendering
libGL error: open DRM failed (Operation not permitted)
libGL error: reverting to (slow) indirect rendering

Re: [Chromium-users] IRC

[Andrew J. <kd...@gm...>]

> When you say you "only get one window" you mean that the render is being
> done on a single Head, or that you only see the render (for instance,
> Atlantis demo) and no other windows?

I am getting a completely unmodified atlantis window.  The window size
is unchanged (as would not be the case for crdemo.conf) and the blank
window which is supposed to be there according to
[http://chromium.sourceforge.net/doc/chromium2.html#CompilingCr] does
not appear.  Moreover, as I've been testing more this morning, none of
the client spus are taking effect.  For example, the following would
render something very different from the original atlantis:

$python ~/cr-1.9/mothership/configs/crdemo.conf 'atlantis -s 100' hiddenline
$python ~/cr-1.9/mothership/configs/crdemo.conf 'atlantis -s 100' print

Nevertheless, I am still getting the same old atlantis window.

> It's been a while since I messed around with the configuration, but have you
> set the appropriate parameters on each render node? If I'm not mistaken,
> there must be CRMOTHERSHIP env variable set in order for chromium to work
> properly.

I've been leaving the mothership variable out as localhost is correct.
 I'll try it just to remove the errors.  I'll report a change if there
is one.  Right now, my PC has capacity for 6 monitors over three
nvidia graphics cards.  I have no need to add extra network nodes at
this time.  Would my hardware configuration influence the results?

Messages from the apps are as follows:

MOTHERSHIP:

jackman@charliebrown ~/cr-1.9/mothership/configs $ python
crdemo_full.conf 'atlantis -s 100' hiddenline
This is Chromium, Version 1.9
Start a crappfaker on charliebrown.jacknet
Start a crserver on charliebrown.jacknet

Thank you for using Chromium!

CRSERVER:

jackman@charliebrown ~ $ crserver
CR Warning(charliebrown:5184): Couldn't find the CRMOTHERSHIP
environment variable, defaulting to localhost
CR Warning(charliebrown:5184): Couldn't find the CRMOTHERSHIP
environment variable, defaulting to localhost
NVIDIA: could not open the device file /dev/nvidiactl (Permission denied).
NVIDIA: Direct rendering failed; attempting indirect rendering.
CR Warning(charliebrown:5184): Couldn't find the CRMOTHERSHIP
environment variable, defaulting to localhost
CR Warning(charliebrown:5184): Couldn't find the CRMOTHERSHIP
environment variable, defaulting to localhost
CR Error(charliebrown:5184): Mothership didn't like my accept request

CRAPPFAKER:

jackman@charliebrown ~ $ crappfaker
CR Warning(charliebrown:5192): Couldn't find the CRMOTHERSHIP
environment variable, defaulting to localhost
NVIDIA: could not open the device file /dev/nvidiactl (Permission denied).
NVIDIA: Direct rendering failed; attempting indirect rendering.
freeglut (/home/jackman/cr-1.9/bin/Linux/atlantis): Unable to create
direct context rendering for window 'GLUT Atlantis Demo'
This may hurt performance.
NVIDIA: could not open the device file /dev/nvidiactl (Permission denied).
NVIDIA: Direct rendering failed; attempting indirect rendering.
NVIDIA: could not open the device file /dev/nvidiactl (Permission denied).
NVIDIA: Direct rendering failed; attempting indirect rendering.
freeglut (/home/jackman/cr-1.9/bin/Linux/atlantis): Unable to create
direct context rendering for window 'freeglut menu'
This may hurt performance.
122.6 fps
123.8 fps
123.5 fps
122.1 fps
124.0 fps

Would you like me to post screenshots somewhere?

I'm going to add my user to the video group (get permissions to the
nvidia device) and I'll try it again.

[Time passes.]

I also added the '-mothership localhost' arguments to the crserver,
which removed all of the relevant errors.

Aside from removing the permissions errors and the framerate tripling,
there was no effect on the test with regards to the nvidia
persmissions.

387.2 fps
317.4 fps

Still, the lack of spu handling and the lack of an extra window persist.

Jackman.

Re: [Chromium-users] IRC

[Andrew J. <kd...@gm...>]

Well, by all means, I'd like to talk to /someone/.  Thank you for your response!

If anyone is willing to talk on a more spontaneous leveI am on gTalk
(using this address) or other protocols as necessary.

On Sat, Apr 18, 2009 at 2:31 PM, Andres Felipe Molina Villamizar
<af....@un...> wrote:
> I'm not sure about the IRC channel, but activity on the mailing lists has
> close to null recently.
>
> Sorry I can't be of more assistance.
>
> -Andres
>
> -----Original Message-----
> From: Andrew Jackman [mailto:kd...@gm...]
> Sent: Saturday, April 18, 2009 3:16 PM
> To: chr...@li...
> Subject: [Chromium-users] IRC
>
> Is there a chromium IRC channel somewhere?
>
> Andrew Jackman.
>
> ----------------------------------------------------------------------------
> --
> Stay on top of everything new and different, both inside and
> around Java (TM) technology - register by April 22, and save
> $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
> 300 plus technical and hands-on sessions. Register today.
> Use priority code J9JMT32. http://p.sf.net/sfu/p
> _______________________________________________
> Chromium-users mailing list
> Chr...@li...
> https://lists.sourceforge.net/lists/listinfo/chromium-users
>
>

Re: [Chromium-users] IRC

[Andres F. M. V. <af....@un...>]

I'm not sure about the IRC channel, but activity on the mailing lists has
close to null recently.

Sorry I can't be of more assistance.

-Andres

-----Original Message-----
From: Andrew Jackman [mailto:kd...@gm...] 
Sent: Saturday, April 18, 2009 3:16 PM
To: chr...@li...
Subject: [Chromium-users] IRC

Is there a chromium IRC channel somewhere?

Andrew Jackman.

----------------------------------------------------------------------------
--
Stay on top of everything new and different, both inside and 
around Java (TM) technology - register by April 22, and save
$200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
300 plus technical and hands-on sessions. Register today. 
Use priority code J9JMT32. http://p.sf.net/sfu/p
_______________________________________________
Chromium-users mailing list
Chr...@li...
https://lists.sourceforge.net/lists/listinfo/chromium-users

[Chromium-users] IRC

[Andrew J. <kd...@gm...>]

Is there a chromium IRC channel somewhere?

Andrew Jackman.

[Chromium-users] Chromium and X

[Andrew J. <kd...@gm...>]

My name is Andrew Jackman.  I've had my eye on chromium for some time
now and I've finally gotten the opportunity to put it into action.  At
my disposal I have a laptop, a desktop with three nvidia graphics
cards, and another lesser desktop box.  The laptop is configured with
a single screen, the desktop with three screens, and the lesser box
with a single screen as well.  Moreover, I've been using a remote X
server (cygwin/X) to attempt to understand how Chromium interacts with
X.  It is because I don't understand how Chromium uses X that I am
writing this email.

As far as I can tell from the documentation, the app nodes and
mothership nodes do not interact at all with X.  It is the server
nodes (specifically the rendering nodes) that deal with X.  Is that
correct?  This would make sense to me as it seems that it is X's
responsibility to render the final GL stream using its loaded drivers.

To test this, I set up X on the greater box (three heads) with
Xinerama extensions disabled and a light window manager (like
fluxbox).  I used the crdemo.conf and the simplemural.conf examples to
try and get multiple windows going.  Despite various attempts, the
/single/ rendering window only showed up on the screen where the
appfaker was launched.

I did a similar test with the laptop and the remote X server and
consistently the single window appeared on the X server where appfaker
was launched.

I also did the following test on both setups and the result was only
one window rendering the atlantis demo without errors (no blank
window):

1: % cd mothership/configs
2: % python crdemo.conf atlantis &
3: % crserver &
4: % crappfaker

>From the two above experiment scenarios, my original hypothesis (that
crserver is the app that interacts with X) was disproved.

I have concluded that either Chromium is simply not doing what it's
supposed to do (because I only get one window despite any change in
the setup) or I am too ignorant to do what I want to do.

I am either asking for my fundamental ignorance to be cleared up or
some troubleshooting help.  Thank you!

Andrew Roy Jackman.

jackman@xe3 ~/cr-1.9/mothership/configs $ X -version

X.Org X Server 1.5.3
Release Date: 5 November 2008
X Protocol Version 11, Revision 0
Build Operating System: Linux 2.6.27-gentoo-r8 i686

[Chromium-users] Adding OpenGL2.0 support

[Andres F. M. V. <af....@un...>]

Hello all,

After working on a high resolution tiled display using Chromium and DMX, I
have come across several issues (documented in the users mailing list
archives) which lead me to believe I require OpenGL 2.x support.

As part of my graduation project, I have decided to take on the task of
implementing the necessary extensions over the course of the next few
months.

 

As far as I know, there is still no OpenGL 2.0 support to date, with little
work done on it so far.

 

I’ve started going over the documentation as well as mailing list archives
for pertinent information, but I would greatly appreciate any suggestions or
pointers on the matter.

 

Many thanks,

-Andres

 

--------------------------------------------

Andrés Molina

Systems and Computer Engineering

Universidad de Los Andes

IMAGINE Group

 

Re: [Chromium-users] Chromium on OSX?

[Benjamin K. <be...@co...>]

hi Alan

thanks for your reply
my guess is that then Chromium doesn't make use of the PBuffer at all...

just a n00b question then, how does Chromium work under OSX?
does PBuffers ever gets called?

i faked a copy of all the calls in opengl_stub/cgl.c but returning 0  
just to get crappfaker to launch correctly and not terminating on  
symbol errors.
and crappfaker manages to launch by with a 0,0 dimension app that just  
hangs.

are my edits causing this error?
if so... why? i thought that PBuffers where never used.

thanks lots
----
bENJAMIN kHOO
qIU wENDA
http://www.copperblue.per.sg
http://himiko.dnsalias.net:8000
gpg --search-keys be...@co...

perl -e "while(length(\$s='1079797275786966')-(++\$i*2)>=0) 
{push(@_,substr(\$s,length(\$s)-\$i*2,2))}print pack('c'.scalar(@_),@_)"





On 21/01/2009, at 6:35 PM, Alan Hourihane wrote:

> On Wed, 2009-01-21 at 02:11 -0800, Benjamin Khoo wrote:
>> hi
>>
>> i have tried building Chromium on OSX and has gotten as far as being
>> able to complete the build.
>>
>> however, when i run crappfaker, i get an error that CGLCreatePBuffer
>> symbol isn't found in the OpenGL library, which is supposed to be  
>> part
>> of libcrfaker.dylib
>>
>> after some probing around, i have found that these symbols are
>> available in the OpenGL.framework, but not replicated for libcrfaker
>> which pretends to be part of the OpenGL framework at runtime.
>>
>> CreatePBuffer was just one of the symbols. apparently, there is a  
>> host
>> of other symbols that doesn't replicate itself from OSX to  
>> libcrfaker.
>> does anyone know how to get past this problem or have a proposed
>> solution?
>
> PBuffers have never been supported in Chromium, but could be added  
> with
> a little work.
>
> Alan.
>
>

[Chromium-users] Chromium on OSX?

[Benjamin K. <be...@co...>]

hi

i have tried building Chromium on OSX and has gotten as far as being  
able to complete the build.

however, when i run crappfaker, i get an error that CGLCreatePBuffer  
symbol isn't found in the OpenGL library, which is supposed to be part  
of libcrfaker.dylib

after some probing around, i have found that these symbols are  
available in the OpenGL.framework, but not replicated for libcrfaker  
which pretends to be part of the OpenGL framework at runtime.

CreatePBuffer was just one of the symbols. apparently, there is a host  
of other symbols that doesn't replicate itself from OSX to libcrfaker.
does anyone know how to get past this problem or have a proposed  
solution?

thanks lots

----
bENJAMIN kHOO

[Chromium-users] Parallel Rendering

[Pierre <pi...@co...>]

Dear all,

I am currently testing chromium into different configuration, but I am 
having troubles to understand how to run it for parallel rendering:

I managed to run the atlantis demo on a single computer (so in 
sequential) like it is described in the documentation (even through 
VNC). Now, I would like to do the same but in parallel for the following 
2 configuration (but NOT through VNC to start):

I have 3 computers (comp1, comp2, comp3) that can communicate with each 
other, each with a graphic card and a screen.

Configuration 1: I run atlantis on comp1 and the rendering is done on 
comp2 and comp3, with each displaying its part of the application on its 
own screen (tile display).

Configuration 2: I run atlantis on comp1 and the rendering is done on 
comp2 and comp3 and the result sent back to comp1 which display the 
image on its screen.


Could any one explain me (or re-direct me to a documentation page) how 
to achieve this?

Thank you ever so much in advance,


       -Pierre.



This e-mail message may contain confidential and/or privileged information. If you are not an addressee or otherwise authorized to receive this message, you should not use, copy, disclose or take any action based on this e-mail or any information contained in the message.
If you have received this material in error, please advise the sender immediately by reply e-mail and delete this message. Thank you.
Allinea Software and Streamline Computing are trading divisions of Concurrent Thinking Limited: Registered in England and Wales No: 03913912
Registered Address: The Innovation Centre, Warwick Technology Park, Gallows Hill, Warwick, CV34 6UW, United Kingdom

[Chromium-users] interactive apps: input devices?

[Christopher C. <cro...@ym...>]

Hello,
I am quite new to Chromium. Somewhere I read that simple mouse interaction had been implemented in a development version, but I cannot find this link again, and it doesn't work for me.

Is it true, that interactive control of the Chromium applications is possible? If not, could you outline which are the hardest problems to expect when implementing mouse events forwarding from the display servers to the client application? Are there similar open source projects that have this functionality implemented?

cheers
Christopher



      

Re: [Chromium-users] Modern Chromium

[Byungil J. <myp...@gm...>]

Hi

It is interesting that SAGE is discussed on this mailing list.

As the main developer of SAGE, I have a little concern about that
SAGE is regarded as developed for image/video applications.
Though it is true that image/video is the most easily deployable
SAGE application, SAGE is mainly developed to push extremely
high-res scientific visualizations onto scalable displays.

But one big huddle is application transparency.
The SAGE OpenGL wrapper solves the problem
for a single desktop application but it has resolution limitation
something like 1280x1024.
Since SAGE streams rendered image, this may yield image artifacts
when it is scaled up to a huge tiled display as Chang pointed out.

To avoid that, users have to do off-screen rendering very high-res
images on a powerful graphics server or use parallel rendering apps.
But this requires users to write or modify their apps using SAGE interface,
though the API is a few lines of code.

If you want to show multiple apps on your tiled display at a time,
SAGE should work for you.
For 4x4 tiled display, I don't expect serious artifacts.
If you are happy with running single app on entire tiled display,
use Chromium or other parallel viz apps for tiled display.

Regards,
Byungil


On Thu, Dec 11, 2008 at 10:38 AM, Chang Yuan <cha...@gm...> wrote:
> Hi all,
>
> Here are my two cents.
>
> SAGE is great for image/video applications, but may not be the best choice
> for distributed graphics rendering. It handles OpenGL applications by
> intercepting glSwapBuffer call and then streaming the buffer content
> (instead of OpenGL calls) to the display tiles. The images are rendered with
> inevitable upscaling artifacts which are quite visible on the display tiles.
> It does allow running unmodified OpenGL applications, though.
>
> For my own application, I am using the OGRE rendering engine
> (http://www.ogre3d.org), which uses some advanced shader programs (possibly
>>OpenGL 1.5). I tried to connect it with Chromium, but the scene is rendered
> with missing content (due to the lost shaders I guess) with a very low frame
> rate. Has everybody have successful experience with running OGRE through
> Chromium? Right now I'm sending the pixel buffers through SAGE as a backup
> solution.
>
> I plan to try another distributed rendering engine, Equalizer
> (http://www.equalizergraphics.com/), that claims to support all versions of
> OpenGL calls. But it requires modifying the program, which could be a
> turn-off for some people.
>
> Best,
> Chang
>
> On Thu, Dec 11, 2008 at 7:11 AM, Brock Palen <br...@um...> wrote:
>>
>> Looks like cglx will not support unmodified GL applications,
>> From:
>> http://vis.ucsd.edu/~cglx/docs.html
>> Do I need to recompile my OpenGL application against CGLX ?
>>
>> YES. CGLX needs to intercept a few OpenGL calls to provide a
>> distributed OpenGL context on your visualization system.
>>
>> It does look like SAGE will work though:
>> http://www.evl.uic.edu/cavern/sage/applications.php#wrapper
>>
>> Thanks,
>>
>> Brock Palen
>> www.umich.edu/~brockp
>> Center for Advanced Computing
>> br...@um...
>> (734)936-1985
>>
>>
>>
>> On Dec 10, 2008, at 3:58 PM, Andrew Shewmaker wrote:
>>
>> > This might be a good time for me to ask if anyone has compared cglx
>> > to chromium?
>> >
>> > http://vis.ucsd.edu/~cglx/
>> >
>> > Sorry, I can't answer your Chromium questions.  The last powerwall I
>> > ran it on was RHEL4.
>> >
>> > --
>> > Andrew Shewmaker
>> >
>> >
>>
>>
>>
>> ------------------------------------------------------------------------------
>> SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas,
>> Nevada.
>> The future of the web can't happen without you.  Join us at MIX09 to help
>> pave the way to the Next Web now. Learn more and register at
>>
>> http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/
>> _______________________________________________
>> Chromium-users mailing list
>> Chr...@li...
>> https://lists.sourceforge.net/lists/listinfo/chromium-users
>
>
> ------------------------------------------------------------------------------
> SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada.
> The future of the web can't happen without you.  Join us at MIX09 to help
> pave the way to the Next Web now. Learn more and register at
> http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/
> _______________________________________________
> Chromium-users mailing list
> Chr...@li...
> https://lists.sourceforge.net/lists/listinfo/chromium-users
>
>



-- 
Byungil (Brent) Jeong
Software Engineer
Advanced Visualization Group
Sun Microsystems
(650) 786-7687 / FAX: (650) 786-2514

14 Network Cir
Menlo Park, CA 94025

Re: [Chromium-users] Modern Chromium

[Andres F. M. V. <af....@un...>]

What kind of setup (hardware, network, mothership configuration) are you
using when running Ogre on chromium?

 

I know someone I work with had a go at Equalizer, but we ended up choosing
Chromium for precisely the reason you mention, Chromium does not require you
to modify the applications.

 

 

From: Chang Yuan [mailto:cha...@gm...] 
Sent: Thursday, December 11, 2008 1:39 PM
To: Brock Palen
Cc: Andrew Shewmaker; chr...@li...
Subject: Re: [Chromium-users] Modern Chromium

 

Hi all,

Here are my two cents.

SAGE is great for image/video applications, but may not be the best choice
for distributed graphics rendering. It handles OpenGL applications by
intercepting glSwapBuffer call and then streaming the buffer content
(instead of OpenGL calls) to the display tiles. The images are rendered with
inevitable upscaling artifacts which are quite visible on the display tiles.
It does allow running unmodified OpenGL applications, though. 

For my own application, I am using the OGRE rendering engine
(http://www.ogre3d.org), which uses some advanced shader programs (possibly
>OpenGL 1.5). I tried to connect it with Chromium, but the scene is rendered
with missing content (due to the lost shaders I guess) with a very low frame
rate. Has everybody have successful experience with running OGRE through
Chromium? Right now I'm sending the pixel buffers through SAGE as a backup
solution. 

I plan to try another distributed rendering engine, Equalizer
(http://www.equalizergraphics.com/), that claims to support all versions of
OpenGL calls. But it requires modifying the program, which could be a
turn-off for some people.

Best,
Chang


On Thu, Dec 11, 2008 at 7:11 AM, Brock Palen <br...@um...> wrote:

Looks like cglx will not support unmodified GL applications,
From:
http://vis.ucsd.edu/~cglx/docs.html <http://vis.ucsd.edu/%7Ecglx/docs.html> 
Do I need to recompile my OpenGL application against CGLX ?

YES. CGLX needs to intercept a few OpenGL calls to provide a
distributed OpenGL context on your visualization system.

It does look like SAGE will work though:
http://www.evl.uic.edu/cavern/sage/applications.php#wrapper

Thanks,


Brock Palen
www.umich.edu/~brockp <http://www.umich.edu/%7Ebrockp> 
Center for Advanced Computing
br...@um...
(734)936-1985




On Dec 10, 2008, at 3:58 PM, Andrew Shewmaker wrote:

> This might be a good time for me to ask if anyone has compared cglx
> to chromium?
>
> http://vis.ucsd.edu/~cglx/ <http://vis.ucsd.edu/%7Ecglx/> 
>
> Sorry, I can't answer your Chromium questions.  The last powerwall I
> ran it on was RHEL4.
>
> --
> Andrew Shewmaker
>
>


----------------------------------------------------------------------------
--
SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada.
The future of the web can't happen without you.  Join us at MIX09 to help
pave the way to the Next Web now. Learn more and register at
http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/
_______________________________________________
Chromium-users mailing list
Chr...@li...
https://lists.sourceforge.net/lists/listinfo/chromium-users

 

Re: [Chromium-users] Modern Chromium

[Chang Y. <cha...@gm...>]

Hi all,

Here are my two cents.

SAGE is great for image/video applications, but may not be the best choice
for distributed graphics rendering. It handles OpenGL applications by
intercepting glSwapBuffer call and then streaming the buffer content
(instead of OpenGL calls) to the display tiles. The images are rendered with
inevitable upscaling artifacts which are quite visible on the display tiles.
It does allow running unmodified OpenGL applications, though.

For my own application, I am using the OGRE rendering engine (http://www.*
ogre*3d.org), which uses some advanced shader programs (possibly >OpenGL
1.5). I tried to connect it with Chromium, but the scene is rendered with
missing content (due to the lost shaders I guess) with a very low frame
rate. Has everybody have successful experience with running OGRE through
Chromium? Right now I'm sending the pixel buffers through SAGE as a backup
solution.

I plan to try another distributed rendering engine, Equalizer (
http://www.equalizergraphics.com/), that claims to support all versions of
OpenGL calls. But it requires modifying the program, which could be a
turn-off for some people.

Best,
Chang

On Thu, Dec 11, 2008 at 7:11 AM, Brock Palen <br...@um...> wrote:

> Looks like cglx will not support unmodified GL applications,
> From:
> http://vis.ucsd.edu/~cglx/docs.html<http://vis.ucsd.edu/%7Ecglx/docs.html>
> Do I need to recompile my OpenGL application against CGLX ?
>
> YES. CGLX needs to intercept a few OpenGL calls to provide a
> distributed OpenGL context on your visualization system.
>
> It does look like SAGE will work though:
> http://www.evl.uic.edu/cavern/sage/applications.php#wrapper
>
> Thanks,
>
> Brock Palen
> www.umich.edu/~brockp <http://www.umich.edu/%7Ebrockp>
> Center for Advanced Computing
> br...@um...
> (734)936-1985
>
>
>
> On Dec 10, 2008, at 3:58 PM, Andrew Shewmaker wrote:
>
> > This might be a good time for me to ask if anyone has compared cglx
> > to chromium?
> >
> > http://vis.ucsd.edu/~cglx/ <http://vis.ucsd.edu/%7Ecglx/>
> >
> > Sorry, I can't answer your Chromium questions.  The last powerwall I
> > ran it on was RHEL4.
> >
> > --
> > Andrew Shewmaker
> >
> >
>
>
>
> ------------------------------------------------------------------------------
> SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada.
> The future of the web can't happen without you.  Join us at MIX09 to help
> pave the way to the Next Web now. Learn more and register at
>
> http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/
> _______________________________________________
> Chromium-users mailing list
> Chr...@li...
> https://lists.sourceforge.net/lists/listinfo/chromium-users
>

Re: [Chromium-users] Modern Chromium

[Brock P. <br...@um...>]

Looks like cglx will not support unmodified GL applications,
From:
http://vis.ucsd.edu/~cglx/docs.html
Do I need to recompile my OpenGL application against CGLX ?

YES. CGLX needs to intercept a few OpenGL calls to provide a  
distributed OpenGL context on your visualization system.

It does look like SAGE will work though:
http://www.evl.uic.edu/cavern/sage/applications.php#wrapper

Thanks,

Brock Palen
www.umich.edu/~brockp
Center for Advanced Computing
br...@um...
(734)936-1985



On Dec 10, 2008, at 3:58 PM, Andrew Shewmaker wrote:

> This might be a good time for me to ask if anyone has compared cglx  
> to chromium?
>
> http://vis.ucsd.edu/~cglx/
>
> Sorry, I can't answer your Chromium questions.  The last powerwall I
> ran it on was RHEL4.
>
> -- 
> Andrew Shewmaker
>
>

Re: [Chromium-users] Modern Chromium

[Andrew S. <ag...@gm...>]

This might be a good time for me to ask if anyone has compared cglx to chromium?

http://vis.ucsd.edu/~cglx/

Sorry, I can't answer your Chromium questions.  The last powerwall I
ran it on was RHEL4.

-- 
Andrew Shewmaker

Re: [Chromium-users] Modern Chromium

[Brock P. <br...@um...>]

Good to know that SAGE might have the same functionality as Chromium  
with unmodified Apps, I will check it out more.

Our need to support ISV packages is very important, thus modification  
of apps (other than replacing libGL.so or similar) is not an option.

Brock Palen
www.umich.edu/~brockp
Center for Advanced Computing
br...@um...
(734)936-1985



On Dec 9, 2008, at 9:46 PM, Andres Felipe Molina Villamizar wrote:

> Can't say much on Chromium, but if I'm mistaken it supports up to  
> OpenGL
> 1.5. I've run into some trouble running apps that use OpenGL 1.8/2.0
> features myself.
>
> As for SAGE, I was actually reading about it today, and they offer  
> an OpenGL
> wrapper that's supposed to be able to run OpenGL apps without any
> modification. Additionally, they also have some information on how  
> to modify
> (ever so slightly) applications in order for them to work on SAGE.
>
> Hope this helps.
>
> -Andres
>
> -----Original Message-----
> From: Brock Palen [mailto:br...@um...]
> Sent: Tuesday, December 09, 2008 9:39 PM
> To: chr...@li...
> Subject: [Chromium-users] Modern Chromium
>
> I have inherited a 4x4 tile wall that was built (hardware) and has
> linux loaded but no software.
>
> Now I know DMX is part of Xorg now, but what is the state of
> chromium? The last update on SF is 2006. Does it even work on
> RHEL5 ?  This is my first tiled display, is there a successor to
> Chromium I should be looking at?
>
> I see the Rocks Viz roll provides Chromium and SAGE. I like Chromium
> because of ability to run unmodified applications. Our apps would be
> matlab, visit, paraview (maybe), and another handful of engineering
> apps,  (Fluent, AbaqusCAE, Comsol).  Most of these were built without
> a tiled display in mind (paraview? is the only one aware?) thus I
> liked this feature of Chromium.
>
> Does Chromium support the updates to OpenGL?  Latest data I found was
> OpenGL 1.5.  Is this a problem with newer apps?
>
> Please direct me for what research I should do to understand the
> options.
> is it true SAGE requires modified apps?
>
> Oh and is the TG Threaded Mothership still available? I can't find it
> on their website. Is it even needed?
>
> Brock Palen
> www.umich.edu/~brockp
> Center for Advanced Computing
> br...@um...
> (734)936-1985
>
>
>
>
> ---------------------------------------------------------------------- 
> ------
> --
> SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas,  
> Nevada.
> The future of the web can't happen without you.  Join us at MIX09  
> to help
> pave the way to the Next Web now. Learn more and register at
> http://ad.doubleclick.net/clk;208669438;13503038;i?http:// 
> 2009.visitmix.com/
> _______________________________________________
> Chromium-users mailing list
> Chr...@li...
> https://lists.sourceforge.net/lists/listinfo/chromium-users
>
>
>