New Build (303)

[Roy Hirshkowitz]

• Work continues on Engine 4.  I made a bunch of "improvements" in this round.  I like the direction this is heading.
• The Preparing Rendering phase of the pre-process should go faster for the Path Tracer and the new engines.  Packet Tracer is unaffected by this change.

[George Ioannidis]

It seems to be OK here. Loads and renders without problems. Currently trying Roy's alternative with caustics. Result looks really promising. Some really great job done, Roy, thanks.

[Roy Hirshkowitz]

Try another model and see what happens.  Send this one if you want me to take a look at it.  

Mario Schipflinger said:

something is wrong here, it started calculating 258 pixel/second over 36 minutes until it reached 3rd pass on a simply 300k resolution...

[George Ioannidis]

Trying to test new build I've got Acad crash two times. First happened after only 30-40 passes, the other after long time (1.5 hour - around 1300 passes), both with engine 4 algorithm. Unfortunately both on the same model. Maybe I'll manage to do some more tests today with different model.

[George Ioannidis]

No, my crashes were different. AutoCAD crash dialogue pops up but in background, nXt window was still active (but not progressing) and I'd successfully saved the rendering BEFORE I hit Close button. After that of course AutoCAD and everything in it died (window was closed automatically).

[kingjin]

To mark the object of alpha channel, engine (E2, 3, 4) don't support. The new engine E4 states don't support the background picture, light seemed to be very good, but the speed and not soon, particle and E1 want to run a long time to improve. This is I encounter.

[Roy Hirshkowitz]

Can I get the model?

George Ioannidis said:

Trying to test new build I've got Acad crash two times. First happened after only 30-40 passes, the other after long time (1.5 hour - around 1300 passes), both with engine 4 algorithm. Unfortunately both on the same model. Maybe I'll manage to do some more tests today with different model.

[George Ioannidis]

Roy, it is the same teapot model.

[Roy Hirshkowitz]

Interesting-- runs fine here.  Wonder what's up.

[George Ioannidis]

Running same file test rendering again yesterday (on my computer this time), no crash appeared (till I manually stoped the render after ~2 hours). I suppose it was due to virus incident two days ago (the brand new SMART FORTRESS 2012 infection from trusted, but hacked!!! site). Up to now, I have found two other applications which were damaged partially by the virus as well. Installing them again over the original installation resolved those problems. Maybe some parts of the nXt or of AutoCAD were damaged too resulting that unusual behaviour. My wife's machine is currently in quarantine running multiple antivirus, antimalware, antispyware and all the others anti-something programs trying to clean it up, hoping that way I'll avoid recovery from the scratch.

[Roy Hirshkowitz]

Memory usage shouldn't be anywhere near 17GB.  I'd love to take a look at this one as well.

Straw shadow is fascinating...

[Roy Hirshkowitz]

George Ioannidis said:

Running same file test rendering again yesterday (on my computer this time), no crash appeared (till I manually stoped the render after ~2 hours). I suppose it was due to virus incident two days ago (the brand new SMART FORTRESS 2012 infection from trusted, but hacked!!! site) ...

I doubt it George-- this is likely an nXt problem-- we just need to get it reproduced here.

[Roy Hirshkowitz]

With a rubble bumpmap on the glass: