Capping framerate

SouthernGorillaSouthernGorilla United States Join Date: 2017-07-26 Member: 232057Members
I see in the console that the max framerate is set at 144 Hz. I'd like to cap it at 60 Hz to see if that helps stop the crashes I'm having. But I can't find any way to adjust it. I have the cap set in the Radeon settings, but I'm curious as to whether there's a conflict going on when Subnautica tries to draw faster than 60 Hz and that's why the graphics crash. Capping the FPS at 60 Hz would be just one more thing I've tried to resolve the issue.

Comments

  • Kouji_SanKouji_San Sr. Hινε Uρкεερεг - EUPT Deputy The Netherlands Join Date: 2003-05-13 Member: 16271Members, NS2 Playtester, Squad Five Blue
    edited August 2017
    I think I read something about the console command vsync by accessing the console, keys [F3] -> [~] and typing:

    vsync 0 (is no limit capped at 144hz/144FPS I guess, with Vsync off)
    vsync 1 (is the using Vsync to limit it to your monitor refresh rate, so for 60Hz you get 60FPS)
    vsync 2 (is 30FPS, or rather half the monitor refresh rate?)
  • SouthernGorillaSouthernGorilla United States Join Date: 2017-07-26 Member: 232057Members
    Yeah, I saw that and tried it. But it didn't make any difference. At least, the target FPS shown in the console didn't change. Also tried a "maxfps 60" command I read somewhere. But, again, nothing changed that I can see.

    I don't monitor my FPS very often. I seriously doubt it's ever getting far above 60, if it even hits 60. I'm just trying everything I can think of to keep the game going. It's not the game crashing, just the video. We can still hear the audio and the controls still work. But the graphics crash, we get a blue screen, then it goes to the desktop and we can't get back into the game. It's very strange. I'm running a new RX460 with the most current driver and stock Radeon settings. The rest of the system is a bit old, but it should be capable of handling a game, especially with 8 GB of RAM. Adjusting the video settings doesn't make a difference either.

    What gets me is that I can run the Unigine Valley benchmark and OCCT system stress tests without any problem. We can go days without a single crash. Then some days we can't get it to run more than 5-10 minutes without crashing. We've even run the game with the internet disconnected to test if it was a network issue, we still got a crash. I've replaced the power supply since the OCCT test showed some suspicious power fluctuations. I swapped in a Coolermaster Hyper-Evo 212 CPU cooler to improve thermals. The only thing left in the system that may be causing the issue is the Q6700 quad-core processor or the Asus P5K motherboard. My next step is to overclock the RAM to see if memory speed will help.
  • Kouji_SanKouji_San Sr. Hινε Uρкεερεг - EUPT Deputy The Netherlands Join Date: 2003-05-13 Member: 16271Members, NS2 Playtester, Squad Five Blue
    edited August 2017
    maxfps is a console command from Natural Selection 2, one of UWE's other game. AFAIK Subnautica doesn't actually have a command similar to that. Hence I thought of Vsync, which is a weird way to cap your FPS as it can introduce inputlag. Not that it would matter all that much for a singleplayer game, I guess :D


    That benchmark isn't comparable to Subnautica as this game is far from optimized and actually running on an engine (unity) that is at this moment in time being pushed beyond what it should be capable of. Even current gen high end gaming rigs can run into FPS dips, which is the code/engine causing issues, not the rig.

    Well, that Kentsfield Q6700 from 2007 is an ancient CPU in gaming terms and missing a lot of newer instruction sets... It seems Intel Haswell (2014) is the minimum requirement (although Sandy 2011 and Ivy 2012 can handle it easily as well). Also you're certainly not going to get max performance out of that RX460 in any case (which is still an entry level gaming card), when running it alongside a CPU that is quite a lot of generations out of date

    The difference in performance between a Quadcore from the second generation of Core2 Duo/Quad (6xxx) and Sandy Bridge is huge, while the difference between Sandy and Haswell isn't all that big



    I used to own a Core2Duo E6420 from the same family and while a superb CPU in it's day, could be overclocked to neat speeds, it was starting to show it's age around 2011 and later with newer more demanding games. I'd advice you to step into the iCore family, come join us :naughty:
  • SouthernGorillaSouthernGorilla United States Join Date: 2017-07-26 Member: 232057Members
    Oh, I know the CPU is a tad dated. Really only bought the rig because it was cheap and had Windows installed so I could start playing another game. I've been messing with Linux on a laptop for a decade and needed a cheap way to get back into a Windows desktop. I plan to upgrade at some point. Sorry to disappoint, but I'll be going Ryzen when I do. B)

    Wouldn't the crashes be pretty much constant if it was a hardware issue? There's no consistent trigger for a crash. Sometimes it crashes while the game is paused, sometimes it crashes when we're just standing in a base, sometimes it even crashes during the load screen. It's also crashed while streaming Hulu and while watching YouTube. Which definitely points to some sort of issue external to the game. But neither Hulu nor YouTube will stress an RX460 or Q6700 to the breaking point. I used that driver-remover software to completely delete the original Nvidia drivers in case there was any interference from them and I've installed, removed, and reinstalled the Radeon drivers.

    Didn't have any issues at all for about the first month I had the rig. I would play around with Star Citizen, which is far more stressful, without crashes. It wasn't until I started back with my Steam games that the problems started. Which is why I tried playing offline, to see if the overlay or network was the problem.
  • Kouji_SanKouji_San Sr. Hινε Uρкεερεг - EUPT Deputy The Netherlands Join Date: 2003-05-13 Member: 16271Members, NS2 Playtester, Squad Five Blue
    edited August 2017
    Are you overclocking anything and is it running stable on acceptable voltages? Cause those random crashes kinda feel like something is not entirely stable in an overclock. Especially because you've replaced the PSU, an old and failing PSU is usually one of the bigger suspects if you get random unexplained crashes

    Heck it could quite a lot of things, ranging from hardware failure (memory corruption) which could cause random crashes as well (MemTest86 time?), HDD on it's last legs, voltage settings, or perhaps the motherboard wants to retire etc...
  • 0x6A72320x6A7232 US Join Date: 2016-10-06 Member: 222906Members
    For the Hard disk, see these tools (I've used CyrstalDiskInfo and HDD Health from that list, which are both free) http://carlcheo.com/best-smart-monitoring-tools. As Kouji_San said, MemTest86 (or MemTest86+) for the RAM, let it run through at least one full pass (not individual test type, I mean a full pass of all tests, you'll see when examining the screen). PSU (power supply) can be the devil to diagnose, as it makes all the other components look like they're failing. Only way I know of is to take a known good one and swap it out temporarily. They make PSU testers, but you'd need one that puts real-world levels of load on it, as at zero load, a bad PSU can test just fine (sort of like a car alternator, man that's a PITA when that happens. "Your alternator's fine! Check the battery! Check the wires! Maybe it's the computer!" *0x6A7232 goes to change the alternator* "Wait, no, I said the alternator's fine! Oh.. maybe it wasn't fine..." Score: 0x6A7232 - 1, Parts store tech - 0
  • SouthernGorillaSouthernGorilla United States Join Date: 2017-07-26 Member: 232057Members
    Not overclocking anything. In fact, I dropped the voltage in Radeon Settings at one point just to see if that stabilized everything. It didn't. I was going to try overclocking the RAM but I'll probably hold off on that for a while.

    I haven't tested the memory yet. Guess I can do that when I go to bed this morning and just let it run. I replaced the PSU because the OCCT stress test showed a lot of huge dips in the power. The power is much more stable now in the test. But it still seems low, the 12v rail isn't reading 12v in the test. But I don't know how accurate the sensor is or if it's just a calculated result. I'm still worried that I should have replaced the 500W PSU with a bigger one instead of sticking with 500W. But that should be plenty to run a fairly simple rig like I have. The GPU doesn't even have an extra power connector, it can't be drawing that much. But that's another thought I had, maybe the motherboard isn't feeding enough power to the PCI-E slot. According to this GPU-Z program I installed the GPU is only running at 8x instead of 16x for some reason. So maybe there's a bandwidth problem.

    Yeah, ten thousand things it could be. Or any combination of any number of things. That's what makes it so aggravating. Pretty much all you can do is keep replacing parts til the problem goes away. The next step would probably be just turn the case upside down, dump everything out, and throw the Newegg catalog at it.
  • SouthernGorillaSouthernGorilla United States Join Date: 2017-07-26 Member: 232057Members
    Started trying something different. They said this motherboard was designed for OC... I don't think they were kidding.
    8zzs7jhipw7d.png
  • 0x6A72320x6A7232 US Join Date: 2016-10-06 Member: 222906Members
    Hmm. You have the latest BIOS available for you board, right? Sometimes there's little quirks with certain types of hardware (or just a bug in the BIOS) that they release updates for which fixes it.
  • SouthernGorillaSouthernGorilla United States Join Date: 2017-07-26 Member: 232057Members
    I did have the last BIOS they released for the board. I had to clear the CMOS yesterday, that doesn't alter the BIOS, does it? It just resets it to factory settings, right?

    Got the CPU from 2.66 GHz up to 3.34 GHz and the 667 MHz RAM up to 883 MHz without touching the voltages. And still got a crash. It's just got to be too old. Either that, or the strange x8 setting on the PCI-E slot is causing problems. I can't find out how to fix that. It's a x16 card obviously and it's in the x16 slot, but it's only running at x8.

    Probably just best to set the rig on fire and start over.
  • 0x6A72320x6A7232 US Join Date: 2016-10-06 Member: 222906Members
    I did have the last BIOS they released for the board. I had to clear the CMOS yesterday, that doesn't alter the BIOS, does it? It just resets it to factory settings, right?

    Got the CPU from 2.66 GHz up to 3.34 GHz and the 667 MHz RAM up to 883 MHz without touching the voltages. And still got a crash. It's just got to be too old. Either that, or the strange x8 setting on the PCI-E slot is causing problems. I can't find out how to fix that. It's a x16 card obviously and it's in the x16 slot, but it's only running at x8.

    Probably just best to set the rig on fire and start over.

    Try re-seating the GPU? Re-seat everything, actually. Pull it all out, blow it off with compressed air, (besides CPU, that should be fine I think) and try that. Resetting the CMOS just clears settings back to factory default.
  • SouthernGorillaSouthernGorilla United States Join Date: 2017-07-26 Member: 232057Members
    Had to pull the GPU to get at the CMOS battery and jumper. Didn't blow it out though.
Sign In or Register to comment.