FPS sometimes are awfully low when starting game... have to exit and start again

AntaresC64

Distinguished
Jul 7, 2014
7
0
18,510
Hi folks,

I've been having this issue for some time now, though it seems to have gotten worse lately. Sometimes (and it seems pretty random to me), when I launch a game, it runs horrible - getting maybe 14-15 FPS. If I exit the game and start it again, things usually improve... but not much. Maybe getting 35-40 FPS.

Sometimes, if I switch from borderless to fullscreen, things seem to speed up a bit. On other times, I boot the game and achieve solid 60 FPS right from the beginning.

I have experienced it on many games, but it seems especially horrible on GTA V. This afternoon I fired up my PC and played for a good 2 hours at 60 FPS (I have capped framerate with nVidia Inspector) right from the start. Then left to have dinner, went back, turned on my PC again, launched the game... and got 14 FPS. :-(

Any clues? As I said, it seems mostly random, although it is usually worse when I start playing right after booting up my PC. If I do other things previous to a gaming session (like surfing the net, watching videos...) games tend to run smoother. But there's no guarantee either.

Here are my specs:
CPU: Intel i7 4770 @ 3.5 GHz
GPU: Geforce GTX 780ti (factory overclocked)
RAM: 16 GB
OS: Windows 10, 64 bit home edition.
[As for the PSU, I cannot remember the exact model right now -I'm not at home at this moment- but it's a pretty beefy one. 700 W if I recall correctly]
Games are installed on an SSD.

Thanks a lot in advance!

EDIT: last night I launched the game again only to get choppy performance... then I performed a warm reboot, started the game again and... buttery-smooth 60 FPS.

Now that I think of it, last time I encountered that problem was after installing latest nVidia drivers (as of january 2016). I was playing Dragon Age: Inquisition at the time. The game was unplayable right after installing the drivers (note I used DDU to cleanly remove previous drivers), but after a warm reboot it was smooth again.

A couple of google searches have got me wondering about a possible mobo problem... which scares me as hell...
 
it seems like your gpu driver were out of date and now it is up to date so your fps should go sooth with games. if it is still low then try to run the system index which will show you that what is legging in the performance. and its batter to have paging set to custom size rather than system managed size. for batter performance. i am able to run far cry 3 smoothly on my laptop having 4Gb Ram and i5 processor and Gforce 410M GPU (512Mb). and it runs very smoothly.
 


Thanks for your reply :)

Unfortunately, I don't think that is the case. As I said, I have been experiencing this problem for more than 1 year, but never has been this bad. Over this time I have updated my drivers many times. And, as I stated in my post, I can play lengthy sessions of demanding games flawlessly... then turn off my PC, turn it back on a couple of hours later and get 15 FPS. Same drivers, same everything...

Over the last year or so I have been getting some stuttering when turning V-Sync on, but, funnily enough, when I switch from fullscreen to borderless window, and then back to fullscreen, framerate seems to improve enough for the stuttering to go away and for games to run smoothly.

However, as of the last month, I have been running into awfully low FPS... not just some stuttering... I'm talking about totally unplayable games, running (should I say rather "crawling") at 15 FPS. It's not very frequent, mind you... maybe one time in ten. Exiting the game and then launching it again usually solves it - sometimes I hit the 60 FPS mark, other times it just improves a bit over 35 FPS. But, when the game refuses to run smoothly, it seems that a warm reboot usually fixes the problem.

I'm fearing it could be a motherboard-related problem. And that would mean I have been especially unlucky, since my GPU has proven to have an unstable factory OC which lead to freezes in many games.
 


Hi again,

Sorry it took so long, but I have been performing several checks (as well as playing a couple more games, to see if the issue would appear again - hint: it didn't).
My initial conclusions aren't exactly good news for me.

First of all, temps are fine. Actually my GPU has never been plagued by temp issues. Even in graphic intensive games such as The Witcher 3, it never reaches more than 75-80ºC at its peak, and is 70ºC on average. Idle temps are around 30ºC.

CPU temps are good as well.

But the scare came when today I decided to perform a Passmark BurnItTest. My PC *rebooted* only 10 seconds into the test.
I tried a second time and, again, it rebooted at what seemed to be the exact same spot. Not sooner not later. The same spot.

I started to launch individual, separate test. CPU and RAM tests ran fine. Video tests caused a reboot.

Funnily enough, if I launch the tests individually (ie - 2D Graphics, 3D Graphics, GPGPU and Video Playback) they seem to run flawlessly (though I admit I've only ran them for no more than 5-10 minutes each... anyway the reboots always happen only a few seconds into the tests, so I thought that 10 minutes without errors were reassuring).

Launching video tests in pairs (ie - 2D Graphics + GPGPU, or 3D Graphics + Video Playback) caused no trouble at all. A combination of 2D + 3D + GPGPU also ran fine. But as soon as I threw Video Playback into the mix, or launched Video Playback with two other video tests, the reboot happened again (only a bit later this time - maybe 15-20 seconds into the tests).

For the record - I watch a lot of videos on my PC, including lengthy Youtube videos at 1080p and 60FPS. And never-ever had the slightest glitch.

I clearly have a problem with my GPU... or do I? I mean, can I 100% rule out other pieces of hardware (Mobo, RAM, CPU)? Since all crashes and freezes I get in games are always related to Display Drivers / Display Hardware / etc. (according to Windows Event Viewer), and since BIT reboots my PC every time I stress test my GPU, I think it's clear, isn't it? I love my 780 ti, but maybe it's time to start thinking about those shiny upcoming Pascal GPUs 😛

BTW I've already tried to reseat GPU and all power cables. Anyway, everything seems to be perfectly inserted in their slots. Nothing to report there.

I bought my rig from some online PC-builders (actually quite popular and with a good rep in my country) who always ship PCs along with BIT, MEMTEST+ and Furmark results reports. And all the reports they sent me were flawless. They even ran all BIT tests with no problems for more than half an hour... *so how on Earth do I get reboots every 10 seconds*?

Either they have been cheating me (I'd like to think thats unlikely, given this folks have a good reputation) or my PC suffered some damage on its way home (but on a visual inspection everything seems to be perfectly ok).

BTW - my first hard lockup playing a game was about 1 hour after I got my PC. So it's been happening right from the beginning. Maybe I should have RMA it back then, but at the time I thought it was a drivers / software / OS problem, since I had to install everything (including OS) from scratch on an unformated SSD drive, and things were a bit messy.

Erm... sorry for the wall of text 🙂

Anyway I appreciate your help. At this point I'm just thinking if I can be 100% sure that I have a faulty GPU, so I can purchase a new one as soon as I can.
 


Hi again,

Well, my GPU is past its warranty, so I guess I'd have to purchase a new one. The thing is I'd like to be 100% sure that it's a GPU-related problem and not, say, related to the PCI-slot. I have read that, sometimes, signs of a bad PCI slot can mimic a bad GPU. And I don't like that, because I only have 1 PCI slot, and if it's faulty that would mean to get a new Mobo, and perhaps a new CPU. Which, in turn, would mean to get a new Windows, since mine is OEM.

Several weeks ago, when playing GTA V I was getting crashes every 5 - 10 hours. Not bad, but still annoying.

A fortnight ago I was getting one crash per hour or even more. And that was playing an old game (Enslaved: Odyssey to the West - not a especially demanding one). This weekend, I have been playing the same game for 10+ hours. Not a single crash. And everything running over 100 FPS. The difference? I reseated my video card and installed latest nVidia drivers after uninstalling the previous ones using DDU.

But, funnily enough, the bad FPS thing has just appeared again. And worse than ever. At least three times this weekend I had to perform a warm reboot after having AWFULLY low FPS, not only on games, but on Windows as well. Everything was laggy as hell, not only in videogames, but also in Windows.

Just out of curiosity, I launched The Witcher 3. Even on the first screens (which contain only skippable text messages) the game was achieving... maybe 1 - 2 FPS!

Warm reboot, and everything is smooth again. Witcher 3 running at 60 FPS with everything on a mix of Ultra and Very High.

The last time I had that horrible lag, I checked nVidia Inspector and it seemed that the GPU wasn't throttling to full speed, and the core frequency was on 320-ish MHz. But why should that make things like opening up folders on Windows seem so laggy? I mean, we're not talking about demanding games here and the GPU was on its "normal idle state"...

I really don't know what is going on with my PC, but, if all, I hope it's "just" the GPU. Yes, it's an expensive piece of hardware, but one I feel comfortable replacing myself.

I think I have an old Geforce 460 SE lying around. I'll try it on my PC and will let you know.

Again. thank you!