Options Wont Save & Light Pulse Doesn't Work

[Sandy]

There doesn't seem to be any way to confirm changes made to the options, and once I exit and restart Universe Sandbox, my settings are back to default.

EDIT: And just now, I ran into a problem with light pulse. When I click it, the FPS slows to about 3, and no light pulse is emitted. And it just stays at 3 FPS until I open a new simulation.

[Bla]

Only some of the simulation settings are saved currently.

About the light pulse, have you checked the dust multiplier slider in the simulation settings near the bottom? The higher it is, the more it'll slow down the simulation. 1 should be the default.

[NehNudd]

I have run into the same issue yesterday, and I have decided to test a bit what happens when each of the options is turned on or off.

The case of the simulation slowing down to less than 10 fps happens if you consider that you have a high end computer (like I did) and put the Accuracy mode to RK4.
The number of dust particles doesn't seem to matter.
I have the dust multiplier set to 4.1 and my FPS keeps at around 47 (as long as the Accuracy mode is set to Normal at the moment of sending the light pulse - otherwise it drops to below 10fps even if multiplier is set to 0.5).

What happens than is that light pulse doesn't get shown, and the simulation just get slower and slower by each Light Pulse press. So it is calculating something obviously but it doesn't show it.

If you, however, start the light pulse with the Accuracy mode to Normal, you can switch to the RK4 right after the pulse has been sent out from the body, and it will continue to run normally without it dropping to 4 fps.

I hope this helps

The only problem I see now is that the tick next to Use Multiple Processors in Advanced options doesn't seem to change anything, and it doesn't get saved after the program is shut down.

Does it actually do anything and how can that be tested?

As a reference here are a few details about the HW I have:

CPU: Core i7 860
RAM: 8GB
GPU: ATI 4770 with 512 DDR5

[Bla]

Use multiple processors can be tested by running a simulation and opening the task manager, checking if Universe Sandbox uses multiple CPUs. That's what it's meant to do.

[NehNudd]

Bla, thanks for the tip, but I already did try that..
I was running the sandbox in windowed mode so that I can maybe notice some changes in the Performance CPU display in the Task Manager, but regardless of the setting, it showed the same CPU usage.

When I tried running a more complicated simulation (collision of two galaxies) it unfortunately crashed when I ticked the multiple CPU usage..

Maybe it has something to do with how Windows 7 handles the CPU workload distribution, since the cores are always used regardless of the setting (and they are used when I run just Windows, with no sandbox running - only desktop showing and background processes using the cpu).

[Sandy]

Use multiple processors can be tested by running a simulation and opening the task manager, checking if Universe Sandbox uses multiple CPUs. That's what it's meant to do.

I found that with the option ticked, the simulation was only using 2/4 cores, which I would consider a fairly large problem for such a potentially resource-intensive program. I mean, a physics simulator that can't utilize a quad-core? It's sort of ridiculous.

[Bla]

I found that with the option ticked, the simulation was only using 2/4 cores, which I would consider a fairly large problem for such a potentially resource-intensive program. I mean, a physics simulator that can't utilize a quad-core? It's sort of ridiculous.

It's also a new feature which (I think) is still being developed. So hopefully it'll be improved soon.

[Dan Dixon]

Bla's correct. Proper multi-core support is still in development.

[Joshimitsu91]

Use multiple processors can be tested by running a simulation and opening the task manager, checking if Universe Sandbox uses multiple CPUs. That's what it's meant to do.

I found that with the option ticked, the simulation was only using 2/4 cores, which I would consider a fairly large problem for such a potentially resource-intensive program. I mean, a physics simulator that can't utilize a quad-core? It's sort of ridiculous.

Not really? There are many many posts on these forums about multi-core support, it is only in development at the minute. For most simulations, one core is plenty if you have a decent processor. I've never noticed a bad FPS unless the trail length is not on auto or with this RK4-Light Pulse "bug" (i have a core i5 2500K Sandybridge, thats without my overclock enabled). As with any game for the PC, you can't expect to be able to run it at full graphics (or in this case, full capacity) if you have a basic PC.

[LimitlessAeon]

Nice, since there's a topic about this:

Is there any way we can access a CFG file where these are stored and save the options from notepad?

I ask this because there are many games on the PC where you can access game options from the CFG file, set them to what you want and save them without messing with any in-game menus.

[Dan Dixon]

Are you still having problems Sandy?

[loonybin0]

I can't say if Sandy is, but I know I am.

I have an Intel Core i5 machine with an NVIDIA GTX 670 GPU. What I've discovered is that if the accuracy is set to RK4 Math before you try a light pulse, the fps drops from about 50 to 15 and keeps dropping the more light pulses you make (also, as a previous post mentioned, none of these pulses actually appear). If you switch back to Euler Math after trying a light pulse on RK4, all subsequent light pulses do appear but the fps is still lower than usual. If you make a light pulse in Euler and then switch to RK4, the pulse won't disappear but the fps will drop a little bit (not as much as before).

I hope that helps.