Jump to content

Torque 3D 3.7 released


Recommended Posts

  • Replies 57
  • Created
  • Last Reply

Top Posters In This Topic

Slow compared to what? Did you make any comparisons with measurements?

The big mainstream games are often heavily optimized and a lot of tricks are used to make it looks good and run fast, the engine is not the only factor.

For example Torque uses realtime lights for everything by default, if you bake the lighting instead you can increase FPS by around three times, at least this is what I got when I tested it.

Link to comment
Share on other sites

Thanks Duion. I think it's mainly the cpu then. Granted, my pc is by far not the latest and greatest, but still it's disapointing how slow T3D is.

 

To be sure, there are parts of the engine that are slow, but part of my intent with getting the T3D demos is to give us a very concrete baseline that everyone can use to test.


So now that we have a baseline, we can start getting consistent performance figures like what you've been providing and it gives us a clearer picture of what can be optimized.


One thing I'd like you to try, is to go into scripts/client/prefs.cs


And change

 

$pref::TS::maxInstancingVerts = "200";

 

To...say,

$pref::TS::maxInstancingVerts = 3000";

 

By default, T3D has a rather low vert count for instancing(it's largely meant for groundclutter with this setup). I've seen some improvement if you up the instancing amount so it kicks in for trees, so if you could test that and let me know how that impacts performance?

Link to comment
Share on other sites

I have some stutter problems, but I am getting around 145-170 FPS on the default Desert level. This is with a Galaxy GTX 660 GC and a FX 8350 at 1920x1080 on max settings. RAM is 8GB (2 x 4GB) @ 1333mhz. OS is Windows 7 Ultimate x64.


I haven't finished downloading the Pacific Demo to be able to try it yet.

Link to comment
Share on other sites

Thanks buckmaster, I was able to follow that tutorial earlier to get all the files in the same folder as my executable. Please see my post on page 3 about the error I'm getting when I try to start the project. Please forgive me if my terminology is completely wrong. I'm new at this stuff, but I like to learn.

Link to comment
Share on other sites

@NekoDemon117 would you mind doing some profiling and posting the results? It'd be very interesting to see what your CPU is being spent on.


@triggerfish ah sorry I missed that. Might be good to post the console.log somewhere like pastebin for us to see the whole thing. Sounds like something serious!

Link to comment
Share on other sites

Thanks Duion. I think it's mainly the cpu then. Granted, my pc is by far not the latest and greatest, but still it's disapointing how slow T3D is.

 

To be sure, there are parts of the engine that are slow, but part of my intent with getting the T3D demos is to give us a very concrete baseline that everyone can use to test.


So now that we have a baseline, we can start getting consistent performance figures like what you've been providing and it gives us a clearer picture of what can be optimized.


One thing I'd like you to try, is to go into scripts/client/prefs.cs


And change

 

$pref::TS::maxInstancingVerts = "200";

 

To...say,

$pref::TS::maxInstancingVerts = 3000";

 

By default, T3D has a rather low vert count for instancing(it's largely meant for groundclutter with this setup). I've seen some improvement if you up the instancing amount so it kicks in for trees, so if you could test that and let me know how that impacts performance?

 

JeffR, this helps a bit. i get around 5 fps more this way.

I know shipped games are heavily optimized, I was just thinking that compared to for example unity, it seems T3D performs worse on the same Hardware. Regarding baked lighting, how would I do this....do I Need to buy purelight from the GG Store?

Link to comment
Share on other sites

I just tried the old original T3D 1.1 Pacific demo...and guess what, it runs around 10 fps faster than using the latest T3D 3.7 MIT pacific demo....did you notice the same or might this be something related to my machine or Setup only?

Link to comment
Share on other sites

Also, in alignment with the OP's statement about us focusing on stability for 3.8, I've started drafting a "hard" list of things to test prefacing a release.


It's intended as a checklist to make sure we hit as many parts of the engine to ensure it's functioning without surprise gotchas we didn't catch in general testing, which is how we've had the instability slowly creep up on us until now.


As such, I've started a new wiki page that I'll be filling out with an as-exhaustive-list-as-possible listing that we can walk down and check off to make sure a release is as stable as we're going to be able to make it before getting it to the masses.


That page is here.


If there's anything you can think of that needs to be tested, feel free to modify that page or pass it along here or to me via PM. The entire problem we've been having is it's almost impossible for one person to think of EVERYTHING that needs testing, so even if I detail an exhaustive list, there's no doubt some things I'll not think of or forget that you people have noted or tested yourselves.

Link to comment
Share on other sites


×
×
  • Create New...