Please report software faults here.
11 posts • Page 1 of 1
Hey, first off just wanna say great job on L3DT, it's a fantastic tool. Lately however I've been having trouble when manually editing heightmaps in 3D using sapphire.
I have quite a bit of performance trouble for starters, working on 2048x2048 meshes at a 1m horizontal scale it runs really slow and jerky. It's even pretty jerky if it pull down the triangles rendered to a really low amount. My spec is as follows - should it be proving such a hog on my machine?
Intel Core2 6600 @ 3.0GHz
2Gb PC6400 RAM
ATI Radeon HD 4800 512Gb
Intel 80Gb X-25M SSD (system drive)
Western Digital 7200RPM 500Gb HDD (projects are stored on this)
Windows 7 Home Premium
The real problem however is the 3D cursor not working as intended. More often than not I'll be hovering over a piece of terrain and the 3D tool cursor is nowhere to be seen. By moving the camera around and then moving the mouse cursor about the 3D tool cursor sometimes appears and starts moving where expected, but then a few moments later it'll disappear again, forcing another 30 seconds of faffing about with the camera/mouse cursor to try and get it to appear again. I'm lucky if I get one use of a 3D tool between these disappearing acts, and due to the low framerate and regular hitching I'm getting anyway, the camera moving trick to try and get it to reappear is a real chore to go through.
Sometimes in the event log shows up errors stating "cannot get any pixels for point", but this isn't always the case, and the problem certainly occurs without this message appearing so I don't know if it is related.
Let me know if there's any more information I can provide to help explain it, or if you have any ideas about how to resolve it at my end.
Thanks for the bug report. I've not previously heard of problems with an inaccurate or disappearing 3D cursor, so I'm not immediately able to make any suggestions on how to fix or avoid this problem. Could I ask you to send me the logfile after you run a session with Sapphire? You can get to the log file from the Windows Start menu under 'All programs->Bundysoft->L3DT [version]->L3DT logs'. In that directory, you will find a file called 'log.txt'. Please send it to 'email@example.com'.
From your specs I don't think L3DT should be running too slowly, although that does depend somewhat on your chosen texture resolution. Even so, large textures shouldn't make the cursor unresponsive. Hmmm...
Interesting, sounds like I've gotten somewhat unlucky then!
First of all a small correction to my previous post; it is an ATI Radeon HD 4870 card I have, not the 4800 I seem to have typed before.
I've recorded a video of me using sapphire on my machine. As mentioned before it is 2048x2048 pixels with a distance of 1m, and it is using 2048x2048 texture, normal, and attribute maps. It's pretty jerky as you'll no doubt see! You'll also notice that the 3D cursor is inconsistent in when or where it appears. I've tried to get my mouse visible in the video so you can see the movements I'm making and how that relates to the cursor behaviour.
I'll also mail the log file to you. Hope it helps!
I had a similar issue on an older computer. Turns out, I had old graphics drivers from a different manufacturer left behind.
Try this tool, uninstall all the graphics drivers it shows you, and re-install the newest ATI drivers for your card. Might be worth a shot if you've upgraded your card recently.
Thanks for the suggestion - was worth a shot but sadly it made no difference. I didn't really have any potential mixups there anyway I realised after doing it, as I've had the same graphics card in this machine since my last fresh install of windows, and only a few minor ATI driver updates since then that are unlikely to have caused any conflicts.
Thanks for the log and the video. This looks like it's a problem with raycasting used to determine the cursor position on the terrain.
The log doesn't trace the raycasting calculation, but I'll hazard a guess that the OpenGL 'gluUnProject' function is returning invalid vectors, which results in Sapphire performing raycasts in in the wrong direction. If correct, this problem may be caused by your driver's OpenGL implementation (which explains why you see this problem and other users do not).
Before making any radical changes to Sapphire, I'll first expand the raycasting error detection and logging so that we can see whether that really is the problem. I'll let you know when a developmental build is available.
In the latest developmental build of Sapphire (v18.104.22.168/7) I have modified the raycaster in Sapphire to exit sooner whenever gluUnProject returns invalid vectors, and so this bug shouldn't cause the renderer to slow down.
Please note that this fix will not correct the tool position; that still seems to be caused by faulty drivers, and is not something I can easily work around.
Regarding drivers, we now have three reported cases of this bug, two of which are using ATI Radeon cards, and I am waiting on GPU info for the third. If this is indeed problem with dodgy OpenGL support in new ATI drivers, it would not be the first time (nor the second).
Just wanted to let you know that the latest dev builds of L3DT are running great and I'm not having any troubles at all with either the 3D cursor misbehaving or the horrific slowdown I was suffering. L3DT is a pleasure to use once again
Thanks for the status update. This is pretty much what I expected. The cursor will still go haywire occasionally because OpenGL is still occasionally returning an invalid raycasting vector. The change I made simply aborts the raycast promptly if the vector is invalid. The 'real' fix for this will come when ATI update their drivers.
11 posts • Page 1 of 1
Who is online
Users browsing this forum: No registered users and 0 guests