Datsville


RE: Datsville rev528
#51
(2018-01-15, 18:28)Roland Melkert Wrote:
(2018-01-15, 18:23)Merlijn Wissink Wrote: I've got about 6fps on my GTX1070, but according to Task Manger (ok, probably the best monitoring tool Tongue) gpu usage is 'only' about 50%. Any reason for that?

Probably because it's hitting a bottleneck due to the insane amount of triangles.

But what would be the bottleneck then? cpu usage sits at about 10%, ram usage is about 700MB (16GB available) and vram usage is about 200MB (of the 8GB available).
Reply
RE: Datsville
#52
I created a higher resolution render and uploaded it here:

https://mjhorvath.github.io/Datsville/do...ender.html

https://mjhorvath.github.io/Datsville/do...n-map.html

The image is 16384x8192px split into 128 smaller images.

[ed]

The render is currently hard to use in a phone, so I recommend a tablet or desktop.
Reply
RE: Datsville
#53
I have a faster computer now, and was navigating through Datsville using LDView's "Walk" mode, and have to say the experience is really, really cool! You should try it!

Smile

I get about 2.5 FPS currently. I should be upgrading my video card sometime this year and will hopefully at least double that amount.

[ed]

In the meantime, if someone who has a powerful video card right now could record themselves exploring the town at "eye level", and upload the subsequent recording to YouTube, I would greatly appreciate it! Thanks!
Reply
RE: Datsville
#54
I made a render of datsville_rev006.124_inlined_y_boxed_n.ldr in 4K in Blender Cycles using a WIP importer. Blender memory usage in task manager was around 4-5 GB. Importing into Blender took about 12s. LDView and LeoCAD take a lot longer than that to load the model. I'm not sure how much longer since I didn't let it finish loading after about 30s. This is the largest LDraw model I've found so far, so it's helpful for benchmarking.

   
Reply
RE: Datsville
#55
(2020-02-02, 18:09)Michael Horvath Wrote: I created a higher resolution render and uploaded it here:

https://mjhorvath.github.io/Datsville/do...ender.html

https://mjhorvath.github.io/Datsville/do...n-map.html

The image is 16384x8192px split into 128 smaller images.

[ed]

The render is currently hard to use in a phone, so I recommend a tablet or desktop.

Correct URL to page: https://mjhorvath.github.io/Datsville/re...n-map.html
Reply
« Next Oldest | Next Newest »



Forum Jump:


Users browsing this thread: 5 Guest(s)