Jump to content

Pacific Demo performance


Mitovo

Recommended Posts

Hello everyone,


So, been poking at T3D again and wanted to take a look at Pacific demo again, as I remember it featuring a lot of T3D's features to reference.


I noticed something kinda surprising, though.. my frame rate was LOW. Like, really low.


I'm on a 2070 Super, which is getting easily well over 200+ FPS on other games, and even in the other T3D demos. But i'm hitting as low as the 20s in the Pacific demo. This is on DX9. For some reason if I try it in DX11 (by copying the assets into a current T3D setup), GEForce Experience won't show the FPS overlay. So, I don't know if there's any difference in DX11.


So, I"m kinda curious what would cause such a huge difference in performance? Is it an optimization thing? Too much going on? Not good enough occlusion, etc?


Thanks :)

Link to comment
Share on other sites

You need to be more specific for example what demo you tried, since there are multiple versions, there is the original physx one and there is the (badly) ported one or what settings you used and what hardware you got. Then you need an example of a similar game and similar settings and how it runs.

I did not notice the Pacific Demo being super slow, but Torque3D performance overall a little sluggish.

Link to comment
Share on other sites

I first downloaded and tested the Pacific Demo from section 2 of the Downloads page. Performance was as I described, as low as the 20s in FPS with hiccuping while turning the camera.


I then tested Pacific Demo by downloading the Art/Level Package from section 3 of the same downloads page, and extracted it over a new install of 10.3.1.


I run it on max settings @ 2560x1440 (though resolution changes make negligible difference)


Ryzen 2600

NVidia 2070 Super

32 Gigs RAM

Samsung SSD


Not sure what other games are out there that would be a good apples-to-apples comparison with T3D... but I run Witcher 3 on High/Ultra settings and get 50-60 FPS consistently. I dare say Witcher 3 at those settings is far more demanding than the Pacific Demo.


It's worth noting, too, that other demos for T3D run just fine, and I get well over 100, even 200 FPS. This is why my question was "is there something with the way Pacific is designed", and not about the engine itself. I already know the engine can hit up to 200+ fps. Pacific's performance is very much the exception, not the rule. Would like to reference it for "what to avoid" as well as "what to do".

Link to comment
Share on other sites

So I just tested the one you linked and it is the badly ported broken one, but even though the performance was ok for me, I still get my 50-60 fps on max settings on my Geforce 1050, so not much unusual there.

You could try the original Pacific demo from Garagegames, this one is far worse quality wise with bugs in it, it seems someone just ported it over in a haste not considering the things that broke.

Maybe you are encountering some kind of bottleneck that limits the performance that is specific to your setup, hardware or software.


To show the FPS ingame you can enter into console : metrics (fps);

Link to comment
Share on other sites

I did test it by dragging the assets from the direct demo download onto a fresh setup of 3.10.1 and the performance improves by about ~20 FPS in all cases. So that's something, but still well under what I would expect all else considered.


As for it possibly being my hardware as the bottleneck...

- If I was seeing similarly poor results on other T3D demos I could consider that. But that's not the case. I get upwards of 200+ FPS on the other T3D demos.

-If I wasn't getting solid 60+ FPS on games like Witcher 3, Outer Worlds, and any other recent titles, at high or highest settings, I could consider that. But, I get 50-60+ fps in those games, with those settings.


All that considered, my hardware being the bottleneck, specifically and only for the Pacific Demo, is something I'm not even going to entertain here.


So, whatever the version, with all else being equal, it's *something* in the way the Pacific Demo itself is put together. That's what I'm driving at figuring out. What was done differently in the Pacific Demo (which ever version) that makes it perform so much lower than all other T3D demos I can run, as well as newer, more demanding games?


Now, if it comes down to the coding, and the map design (by) itself isn't the culprit, then I can say "okay, so the map/environment itself isn't the problem" and be satisfied with that. However, if there are some "oopsies" in how that map was put together that can so greatly impact performance, then I would like to know what they are, so I can avoid making those mistakes in my own work.


Anyone else familiar with the design or other possible aspects of the Pacific Demo that might be at issue here?

Edited by Mitovo
Link to comment
Share on other sites

Im not more familiar than others , sry , but , if you suspect that it could be an asset you can comment out things in art/datablovks/datablockExec.cs , also , there is even some reference in the art/datablocks/environments.CS concerning a problem causing a crash which at the time appears to not have been resolved . Of course you can also comment out things in the scripts/client/init.cs , but to check the assets Id start with the datablock execs . Maybe you can just comment out everything , I think Id be more concerned with the physics shapes and maybe the postFx but IDK .

Link to comment
Share on other sites

Did you not read what I wrote? I said the Pacific Demo you downloaded is not the real Pacific Demo, someone just took the assets and threw it into the new engine, I instantly see it by how it is done, not much effort was put in to make the visuals, audio and mechanics work like the original demo.


With saying the hardware and software as bottleneck does not mean that your hardware is not good enough, but some kind of setup of hardware and software can create a bottleneck even if the hardware and software normally can run it properly, it just takes a few settings done wrong and your performance goes down the drain.


But even if it was that bad, I cannot confirm it, I get 50-60 FPS all the time in the bad Pacific Demo, so nothing special there. I also know the assets inside out and yes there are some design flaws in the assets as well as how it was put together, but they should not be that big. You are dealing with a non issue here, because the Pacific Demo is not the real Pacific Demo as I said many times, try the original one and see if the problem is still there.

Link to comment
Share on other sites

Did you not read what I wrote? I said the Pacific Demo you downloaded is not the real Pacific Demo, someone just took the assets and threw it into the new engine, I instantly see it by how it is done, not much effort was put in to make the visuals, audio and mechanics work like the original demo.

 

Yes. I read what you wrote. Every single word of every single post. And I've responded.


I'll summarize: I said I'd downloaded and tested it multiple ways, with multiple versions. I tested the original GG version after your first mention, but didn't specifically include that detail, because the results didn't change. The only improvement I saw was ~20fps when I copied the full Pacific Demo's assets over T3D 3.10.1, as I stated in another post.


To clarify, so you can move past that argument: Yes. I tested the original GG version. No, the results were not notably different. FPS pikes up to ~60 at points, but nowhere near the 200+ it hits in other demos. More importantly, it drops down into the low 30s in a number of areas.

 

With saying the hardware and software as bottleneck does not mean that your hardware is not good enough, but some kind of setup of hardware and software can create a bottleneck even if the hardware and software normally can run it properly, it just takes a few settings done wrong and your performance goes down the drain.

 

Okay, you keep putting this theory forward in vague terms, but never get specific.

What hardware/software/settings could prevent a system that runs every other Torque Demo at 200+ FPS, and literally any other modern game at a solid 50-60+ FPS, from running all tested versions of the Pacific Demo, including the original GG version, so much lower?


Specifically what "hardware/software/settings" issues could possibly account for a ~150 FPS difference, only for one specific demo, with every version tested?


You're pushing this theory pretty hard, so I expect you can elaborate and give some specific insight/details. If not, stop putting forward arguments you can't support.

 

But even if it was that bad, I cannot confirm it, I get 50-60 FPS all the time in the bad Pacific Demo, so nothing special there. I also know the assets inside out and yes there are some design flaws in the assets as well as how it was put together, but they should not be that big. You are dealing with a non issue here, because the Pacific Demo is not the real Pacific Demo as I said many times, try the original one and see if the problem is still there.

 

What do you mean "even if it was that bad, I cannot confirm it".... Literally wow. The way such egocentric arrogance just rolls off your fingers is amazing.


Sorry to disappoint, Duion, but my experiences are not subject to your "confirmation". You aren't the all-knowing arbiter of all things T3D.


Now, you've replied several times to this thread, and haven't provided anything beyond "it's fine for me", vague references to "software/hardware/settings" issues without giving any kind of specifics, and "you're testing the wrong version".


Ironically... you claim to know of the issues in the Pacific Demo's design, and dismiss them as "not that big". Yet, you never bother to describe any of those issues, even though that's precisely what I've been asking about since my first post. Par for the course, I suppose.


Now, since I know you're going to reply again, can it be:

1) Specific examples of environment design issues you allude to.

2) Specific examples of the hardware/software/settings issues you claim could be causing problems, which I can actually test and verify one way or the other.


If you could do that, at least one of your posts in this thread would be contributing to the original topic/question.

Link to comment
Share on other sites

Did you switch off every graphic setting there is to narrow the issue down? How are we supposed to figure out what your issue is? I can just make guesses. You are actually the arrogant one trying to blame Torque3D or others for an issue that only exists for you, if you have an issue, you have to figure it out, that is the way it goes usually.


So at least we can say it is not the bad port of the demo, you also said the bad performance also exists with the assets copied over to a normal Torque3D, which would lead me to guess it is the assets causing the problem, so go and remove the assets one by one until the performance issue goes away, kind of an obvious procedure or not?

Link to comment
Share on other sites

Did you switch off every graphic setting there is to narrow the issue down? How are we supposed to figure out what your issue is? I can just make guesses. You are actually the arrogant one trying to blame Torque3D or others for an issue that only exists for you, if you have an issue, you have to figure it out, that is the way it goes usually.


So at least we can say it is not the bad port of the demo, you also said the bad performance also exists with the assets copied over to a normal Torque3D, which would lead me to guess it is the assets causing the problem, so go and remove the assets one by one until the performance issue goes away, kind of an obvious procedure or not?

 

Backpedaling and trying to flip the script now, are we?


How are "we" supposed to figure it out, Duion? First of all, who's 'we?' You're the only one in this thread asserting, repeatedly, that it's something with my hardware/software because it runs fine for you and you "can't confirm" the frame drops I'm seeing.


I ask you to give specific details as to what hardware/software it could be, and you resort to "How are we supposed to figure it out"? And "I can just make guesses".


I'm the arrogant one? I'm "blaming Torque3D and others"?

1. I didn't blame Torque3D. I confirmed that, in all other cases, T3D is running just fine for me.

2. You introduced the idea that the P3D demo was screwed up when converted over. The "blaming" was done by you, friend Not me. I was asking what could possibly be the issues unique to that demo causing such a drastic performance drop, whatever they be. You are the one who came in here pretending to have a clue.


You also claim to know exactly what the issues are with the Pacific Demo's design (aka 'blaming others for the performance hits'). Yet, when asked to give examples, you refuse to do that as well, and continue to backpedal instead.


So let's just state the obvious, Duion... You don't know what you're talking about. You didn't when you entered the thread. You don't now. When called to, you can't actually back up anything that you've said - emphatically and repeatedly - in this thread.


If you were a reasonable, well-meaning person, you'd have made your final "suggestions" to begin with. But, because you're Duion, of course you can't do that.


As you do in every other thread, you just came into this thread to crap all over it with your nonsense, without ever addressing or answering the key question - even after you claim to have "exact" knowledge of it.


Pat yourself on the back, Duion.. You've derailed and ruined yet another thread with your BS. I'm sure you'll chalk that up as an accomplishment.

Link to comment
Share on other sites

How can I derail a thread when there is nothing going on? The only person that could be derailed is you right now. I just told you what my thoughts are what the issue could be, if you are more happy to talk to yourself, then go ahead, then I will not respond anymore.

Link to comment
Share on other sites

. . . then I will not respond anymore.

That's probably best.


@Mitovo The problem you're reporting is certainly an interesting one, and whether it's an issue with hardware or art assets it's worth looking into. Torque3D has a nice profiler setup that might help point in the direction of the problem. Perhaps using profilerDump() on the Pacific demo and another demo would allow a side by side comparison. You can see someone else's dump here as an example. Also try playing around with the other metrics (specifically drawcalls) and see what that is showing. Posting the results here or linking to an upload might help some of the resident eggheads spot any issues.

Link to comment
Share on other sites

@Mitovo The problem you're reporting is certainly an interesting one, and whether it's an issue with hardware or art assets it's worth looking into. Torque3D has a nice profiler setup that might help point in the direction of the problem. Perhaps using profilerDump() on the Pacific demo and another demo would allow a side by side comparison. You can see someone else's dump here as an example. Also try playing around with the other metrics (specifically drawcalls) and see what that is showing. Posting the results here or linking to an upload might help some of the resident eggheads spot any issues.

 

Thank you, Caleb. I'll consider those options.


Honestly, if the answer to my question is "We don't know for sure", then that's a perfectly acceptable response to give from the jump. It's quite possible that who ever created the demo, and/or altered it later is long gone and no longer active in this community, or just hasn't been around lately.


Anyway, I'll look into those tools you mentioned, Caleb.


Thanks again.

Link to comment
Share on other sites

Honestly, if the answer to my question is "We don't know for sure", then that's a perfectly acceptable response to give from the jump.

No, that is more like the exact opposite of a "perfectly acceptable" response, it is a useless response.

 

It's quite possible that who ever created the demo, and/or altered it later is long gone and no longer active in this community, or just hasn't been around lately.

 

It is not only quite possible, it is totally obviously the case. The art was created by Garagegames Studios meaning it was commissioned art that was created by professionals, meaning it is very old and they used 3ds max, which means they may have problems importing into Blender or may not fit the workflow in Blender.


So why don't you just make a performance test on the individual assets like I suggested to you? Should take you less than an hour to test them all and you will know if they are the issue, it is not that hard and by the way I imported, improved and converted the Pacific Demo assets myself, that is how I know the issues with them, but as I said they are not that major.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...