Page 2 of 6

Posted: 23.12.2007, 01:14
by chris
cartrite wrote:I tried everything I can think of short of putting my old 5500 card back in and I can't reproduce these blocks. The wire mode shot looks similar to mine in wire mode as far as polygon size.
cartrite


And the only thing I haven't done is put my 7900 back . . . No idea what's going on here now.

--Chris

Posted: 23.12.2007, 01:26
by t00fri
Toti,

how are things with your machine?

Bye Fridger

Posted: 23.12.2007, 01:34
by chris
Here's what I see. Note how important cloud shadows are to avoid the 'painted on' appearance of clouds when the viewer is near a planet.

Cloud shadows on:
Image

Cloud shadows off:
Image

The effect looks really nice as the sun angle changes.

--Chris

Posted: 23.12.2007, 01:35
by t00fri
I have also tried a fresh install without any non-default textures. ==> All the same...

F.

Posted: 23.12.2007, 01:45
by Toti
Fridger,

I can't run Celestia in the OpenGL2.0 renderpath (my new pc has one of these very limited intel built-in cards)

Just in case, have you tried using a low res base texture (and no specular/night/normal maps) with those clouds instead of a VT?

EDIT: You just answered this question in cross posting :)

Posted: 23.12.2007, 01:49
by cartrite
Seems like a card issue to me. I just put in my old 5500 and the blocks are there. I was running driver 169.07 when I took this.
Image

cartrite

Posted: 23.12.2007, 02:41
by Johaen
I'd love to test it on both the 7900GT and the 6200 I have installed on my machine, but I am unable to build from the source. I've been doing some research on how to do so. I've been unable to find a place to download Visual Studio 2005 Express C++.

Posted: 23.12.2007, 03:47
by selden
Johaen,

It's documented in the Celestia WikiBook. See http://en.wikibooks.org/wiki/Celestia/D ... 2_platform

Posted: 23.12.2007, 04:27
by Johaen
selden wrote:Johaen,

It's documented in the Celestia WikiBook. See http://en.wikibooks.org/wiki/Celestia/D ... 2_platform


I am aware of this. ;) I was already there and that's how I got to where I'm at. I already DLed the SDK, but the link to Visual Studio 2005 Express C++ only leads to the 2008 edition, and that page you linked says that Celestia is not compatable with 2008.

I suppose I could DL the disk image and work with that... I'll have to try I guess.

Posted: 23.12.2007, 07:25
by Johaen
OK. I have the SDK installed. I have Visual Studio 2005 Express C++ installed. Following the directions at http://en.wikibooks.org/wiki/Celestia/Development/Win32_platform/Tortoise_CVS I have TortoiseCVS installed. But when I try to checkout the sourcecode, I get the following error: connect to cvs.sourceforge.net:2401 failed: No connection could be made because the target machine actively refused it.


Bummer. :cry:

I'm going to bed. Hopefully someone will have a fix for this when I wake up. :)

Posted: 23.12.2007, 10:09
by cartrite
Johaen,
Before everyone gets upset that this thread was highjacked, I suggest that we continue this somewhere else. Have another look at the wiki page. There was an error in the original. Look at the entry for Server.
cartrite

Posted: 23.12.2007, 10:43
by t00fri
cartrite wrote:Seems like a card issue to me. I just put in my old 5500 and the blocks are there. I was running driver 169.07 when I took this.
Image

cartrite


Excellent cartrite!

while the issue is still open, we have at least now clear evidence from THREE independent (lower end) cards for the blocks:

- FX 5900 Ultra
- FX 5500
- Quadro NVS 110M (G72 chip) ~ FX 7300

while G80 cards ONLY have been shown to work fine with Chris' code so far.

We also used various drivers, in all cases quite recent ones.
There is a small chance that there is a common driver bug in those recent drivers is responsible.

There is also a clear PERSISTENT driver bug in my Quadro NVS 110M driver, since at least TWO driver updates by DELL and also in all relevant drivers from http://www.laptopvideo2go.com/
that require modification of the DELL *.inf file.

I have reported about this bug in detail before.
http://www.celestiaproject.net/forum/viewtopic ... ptop#88449

It consists of a reproducable extra "mirror dialog popup" appearing only in fullscreen mode . As a reminder, here is a lousy screendump I made with my handy. That second popup on the right should be absent!

Image

So it might be worthwhile to try a considerably older driver with these older cards.

Bye Fridger

Posted: 23.12.2007, 10:51
by cartrite
t00fri wrote:
So it might be worthwhile to try a considerably older driver with these older cards.

Bye Fridger

I was thinking that myself. These newer drivers have updates for the 8000 series and may introduce bugs into older cards. I'll check. It will be a while though. I put my 8600 card back in.
cartrite

Posted: 23.12.2007, 11:34
by cartrite
I still see the blocks after running driver 1.0-8776 on the GeForce Fx5500 card.

Posted: 23.12.2007, 12:22
by t00fri
cartrite wrote:I still see the blocks after running driver 1.0-8776 on the GeForce Fx5500 card.


Good. It seems at least that we are "closing in" somehow ;-)

In my view it is VERY important to find out whether Chris code also works with chips < G80. So we need further "block-hunting" volunteers that have e.g. FX7900 cards.

Cheers,
Fridger

Posted: 23.12.2007, 12:34
by Paolo
In order to have some new bug testers on this feature is very easy to publish an updated version of the exe so also who is not able to compile from CVS ar are too lazy, like me.

Posted: 23.12.2007, 13:10
by cartrite
This probably doesn't affect the code (or could it?) but I was trying to use Kompare to see the differences between the new patch and cvs originals and I couldn't because Kompare shows the whole file shaddermanager.h as being different with respect to the original cvs version of the file because of the dreaded ^M character. I seen them in Emacs. Every line ends with ^M. The other 3 files have the ^M character too.
EDIT: Another thing I found is that the 8000 series started to support 128 bit floating point in case that matters.
cartrite

Posted: 23.12.2007, 13:19
by Adirondack
FYI:

6600 / 163.71:
Image

8800GTS / 163.69:
Image

Adirondack

Posted: 23.12.2007, 13:19
by t00fri
cartrite wrote:This probably doesn't affect the code (or could it?) but I was trying to use Kompare to see the differences between the new patch and cvs originals and I couldn't because Kompare shows the whole file shaddermanager.h as being different with respect to the original cvs version of the file because of the dreaded ^M character. I seen them in Emacs. Every line ends with ^M. The other 3 files have the ^M character too.
EDIT: Another thing I found is that the 8000 series started to support 128 bit floating point in case that matters.
cartrite


cartrite,

the ^M is nothing but the Windows line ending displayed in an UNIX editor. This happens since Chris L. works under Windows. Since some devs use Windows some use Linux, that's quite OK and taken care of.

The 128bit floating point might be an issue, of course. We really need to know whether there is a card just below G80 that does NOT show the blocking.

Bye Fridger

PS: When do you sleep actually? ;-)

Posted: 23.12.2007, 13:27
by cartrite
Cat Naps only. I'm used to getting up early because when I actually work, I got an 100 mile commute to the Philadelphia area down the PA autobaun.
cartrite