Thomas Failor

Lags on Terrain 2.0 vs. Terrain 1.0 newly upgraded hardware

Recommended Posts

Team, 

I suffer the same FPS degradation as others have reported, but it's just as bad on Terrain 2.0 as it is on Terrain 1.0 saves.  Eventually, after exploring, or putting down tethers,  a save becomes so bogged it just isn't playable anymore below 15 fps and I just delete it.  Terrain 2.0 is no different, and is directly related to the amount of territory I cover.

I just upgraded my GPU to an RTX 2070 (from a GTX 1050) in a ePGU attached via Thunderbolt3 to a Lenovo PC with a 16GB of RAM and a core i5 processor.

The new Exotic Tech demo moon exhibits much more of a hitch / lag while running, walking , driving than previous versions did, even with the new hardware upgrade.  Overall, 0.10.5 on my new GPU is MUCH faster, 55-60fps where I was pulling 40 max on the GTX 1050.  But the Tech Demo Moon runs signifantly worse in terms of lag than the Terrain 1.0 levels do, on the same hardware.

I realize I have a CPU bottleneck on my system, but old saves, new game starts on Terrain 1.0 run buttery smooth.

Share this post


Link to post
Share on other sites

Which i5 processor? A 15 watt dual core i5-7200U performs wildly different than a 95+ watt 6-core i5-9600K.

Is your host's thunderbolt a 4-lane PCIe implementation or a 2-lane?

Edited by Jjandrah

Share this post


Link to post
Share on other sites

It really doesn't matter what hardware you're running. On Xbox, it runs terribly pretty much straight away, where the main game is pretty well optimised at this point. 

The tech demo was meant to be used for less than an hour a time on top-end hardware. It was built back in May/June for e3 so there's probably not much we could say about it that would help the devs, given the development and new content that must've happened in the last 6 months. 

It's not actually a feature for the players to give feedback on, despite how it was announced. They only added it because they consistently promised the community a sample of t2.0 before release. 

Share this post


Link to post
Share on other sites
7 hours ago, Blind Io said:

It really doesn't matter what hardware you're running.

 

It matters, faster hardware gives more fps.

Share this post


Link to post
Share on other sites
4 hours ago, mikk said:

 

It matters, faster hardware gives more fps.

I think the point is it shouldn't matter. Appropriate graphics settings for the hardware should yield a respectable FPS without degradation resulting from prolonged periods of play/many more objects being handled.

A game that's only playable on a top-end PC is a useless game. It targets a very niche market and would absolutely close off console markets.

 

I've experienced degraded performance with a rig that's about 8 years old (due an upgrade) but is still decent and should run a game like Astroneer without a hiccup.

Platform:
Steam Play

Version / Build Number: 
0.10.5

Specifications:

  • OS: Linux (Ubuntu 18.04)
  • CPU: Intel i7-3770
  • GPU: GeForce GTX 970
  • RAM: 2x8GB DDR3 1600 MHz
  • Drive: 2TB SATA Drive 
Edited by cpd

Share this post


Link to post
Share on other sites
4 hours ago, mikk said:

 

It matters, faster hardware gives more fps.

My point was that the tech demo isn't anywhere near optimised enough to be properly usable, so using better hardware will give minimal benefits. Context is important. 

Share this post


Link to post
Share on other sites

They have claimed more than once that the main reason for Terrain 2.0 is/was performance, it is really worrying to see it actually performs much worse than the old Terrain even if it's a very old version of Terrain 2.0. I'm sceptical of the release, they can't even rely on user feedback anymore. The last update is too old.

Share this post


Link to post
Share on other sites

Let's bring this back to things we can measure and compare. OP, can you tell us more about your setup and upload a save game so we can compare our performance to yours?

Share this post


Link to post
Share on other sites
1 hour ago, Jjandrah said:

Let's bring this back to things we can measure and compare. OP, can you tell us more about your setup and upload a save game so we can compare our performance to yours?

Why? The exotic moon is just there to give a glimpse of the new planets. There's no point trying to troubleshoot a build that's over half a year old - there's no point in performance feedback until 1.0 as the game we have is far removed from the game they'll release. 

Share this post


Link to post
Share on other sites
14 minutes ago, Blind Io said:

Why? The exotic moon is just there to give a glimpse of the new planets. There's no point trying to troubleshoot a build that's over half a year old - there's no point in performance feedback until 1.0 as the game we have is far removed from the game they'll release. 

I doubt it has drifted that much already to where our feedback is useless. It's not like they're porting away from Unreal Engine 4.

Maybe there's a particular configuration that bottlenecks this game harder than other games, and we can learn enough to actually help OP out.

Share this post


Link to post
Share on other sites
1 hour ago, Jjandrah said:

I doubt it has drifted that much already to where our feedback is useless. It's not like they're porting away from Unreal Engine 4.

Maybe there's a particular configuration that bottlenecks this game harder than other games, and we can learn enough to actually help OP out.

As seen in the exotic moon, just about everything that's procedurally generated will change due to t2.0. Other than tethers, those items (Terrain, rocks, tress, etc) are what's affecting performance. The whole thing's been rebuilt and will replace the current system. So no point in feedback on the main game. 

We don't have a working model of the current iteration of t2.0, we just have the tech demo that was made for e3. I don't mean to dismiss you, feel free to keep it going, but they'll presumably have the new tech running as good as or better than the current one by release. So not much point in feedback on the moon. 

Edited by Blind Io

Share this post


Link to post
Share on other sites
On 12/10/2018 at 6:14 PM, Jjandrah said:

Let's bring this back to things we can measure and compare. OP, can you tell us more about your setup and upload a save game so we can compare our performance to yours?

Hardware config as follows:

Core i5 7300U (Lenovo Thinkpad 370)

4 lane PCIe - Thunderbolt 3

16 GB DDR4

Akitio Node 

Gigabyte RTX 2070

2560x1440p resolution @59Hz

This config is CPU bottlenecked but terrain 1.0 plays just fine until you explore too far and it gets progressively slower as you spread out, visit other planets.  Particularly bad if you use tethers.  Starting a fresh save runs great.

Terrain 2.0 is significantly worse, game plays at 60fps but has obvious lags when driving, running even though the CPU is at 40%.  It looks beautiful, but the performance is not there even with a worlds top 10 GPU.

Maybe 1.0 will be better?

 

 

Share this post


Link to post
Share on other sites

OK! After the month, I managed to get my hands on something kinda similar. I tried a 15w i7-7500U laptop, a 4-lane TB3 implementation, and a GTX 1060 in an egpu box. This isn't exactly what you're running but it's what I could conjure up.

Starting on Terran, the game behaves beautifully. As shown in task manager, CPU utilization hit 39-41% at 3.48ghz. Even at High/2560 graphics it played smoothly.

Starting on the Terrain Test planet, I found that the two CPU cores were maxed out (task manager showed ~51%, also holding at 3.48ghz). As I built platforms, items, and dug, the CPU utilization bled into hyperthreading (60-70% utilization) and the game would hitch. This CPU load and hitch behavior replicated regardless of graphics settings - High/2560 or on low/1080.

So... ultimately, two conclusions. For OP, fiddle with your settings to try and get as high of a CPU clockspeed as you can push. Worst case, trade up the laptop to one with a quad core CPU, prefurrably one with 45w at it's disposal (an -H instead of -U).

For systemera, your terrain test needs more CPU grunt than typical dual core devices customers are using (laptops). You might wish to consider bumping the minimum system CPU reqs to say 'quad core CPU', as the recommended specs read, before 1.0 launches.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now