• We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.
Status
Not open for further replies.
Why a cap at all? The only thing you will protect are your ears.
It also can help improve graphical quality and performance overall.

Modern GPUs will throttle from heat and it takes some time to transfer that to a heat spreader and then to a cooling system. The load in a game tends not to be even so it is constantly making adjustments and cooling lags behind these adjustments. You could run with 100% cooling all the time but you will have wind tunnel level noise that will distract from your gaming experience as well. Even a water cooling system will have some fans that are going to be loud at 100% speed.

It is also wasteful of energy which will not only increase your electricity bill but can also shorten the life of some components such as fans that have a shorter MTBF when run at max speed. The extra waste heat will also require more elaborate cooling or otherwise build up in your case which could cause throttling.

Lastly the cap also smoothes out performance at extremely high fps which can make it more visually appealing. It can help avoid rapidly changing fps and related poor frame times.

tldr: A cap can make your system work better overall.
 
I would say nope.
I have a completly Air Cooled system and i haven't seen my 1080 ti throttle down. I haven't seen the cards throttle down at Games Exhibitions were they run on 100% load from the morning until the last guy leaves the booth. And there we have a non healthy environment for GPU's. 40°C (often more) outside of a case are the normal environment on exhibitions. And you want to tell me that in a normal room at home the cards will throttle down. When this happens the cooling system overall is lackluster and should be upgraded.

Have you ever run a mining cluster?
If yes then you should know better. If no then you should do this to see how far away your assumption is. I did run 3 cards in one case (with adequate cooling) with 100% load for weeks without them throttling down. How do i know that they didn't throttle? A stable Hashrate over the whole mining operation.

The only point i say yes. It is wasteful of energy, that's true. But not really, my 1080 ti draws less power then a 780 ti or a GTX 580. I can run my High-End system with a 600W power source. Some of my systems in the past wouldn't even start with only 600W. So yeah you are right, when you forget how wasteful older Gaming systems have been.
 
Just a heads up from personal build experience: my EVGA NVIDIA GTX 1080 reference card would throttle down during stress tests. And no, I don't overclock it. I stress tested it when I first installed it back in 2016. I then took apart its stock cooler, cleaned off the sloppy thermal paste application from the factory, applied Arctic Silver 5 (my long time go-to thermal compound), re-installed cooler, and stress tested it overnight. I then ran it for a couple weeks to ensure some curing of the compound before tracking temps and comparing them with stock.

The result: now my 1080 never throttles down even while at full load and stress testing. This is with the default auto fan curve. I run my own, more aggressive, but still quiet fan curve, and never have issues with temps or performance.

I did the same when the GTX 680 was brand new, except I lapped that heatsink and changed the exhaust bracket. I did a mix of custom heatsinks, coolers, lapping, and more on a lot of cards in the past, including 7600GT cards in SLI, 8800 GTX cardsin SLI, GTX 285 SC, GTX 295 Co-Op, GTX 680, and now my GTX 1080. Lesson learned from all of that: the process of OEM heatsink mounting sucks. And generally speaking the cheaper the OEM, the worse the install of the cooler. Second lesson learned: nothing wrong with reference coolers that exhaust hot air out of the back of your system. Open air 3rd party coolers can do a better job of removing heat from the GPU and other components, but you need additional case fans (or better fans) to remove that heat dumped inside the case. I've used a mix of both over the years, from reference coolers to OEM designed coolers and 3rd party coolers. The only reference cooler I didn't particularly like on modern video cards was the one for the 295 Co-Op, as it was anemic for what that card could do. If you go back further to the ATI (now AMD) glory days, NVIDIA had horrible reference coolers while AMD had some nice ones.

So lesson for other users who don't want to go through the hassle of removing stock coolers to re-do the thermal paste: run your fan curves higher or just add more/better case fans. Stock installs of coolers generally sucks.
 
I am not so much talking about severe throttling from severe overheating. They will also throttle slightly during normal operation because of localized heat within the die that can't get removed quickly enough from that point. This has become more of an issue as die process has shrunk and density has increased. The complicated automatic heat management designed in deals with this without any user involvement. The reported heat is likely to even be fine because these localized hot spots are normal. But the management system will be reducing the maximum possible sustained performance. It may be a significant hit or a negligible one but it is an avoidable degradation by just capping FPS somewhere above the maximum refresh rate of your display and thus reducing the source of extra heat and power.
 
They will also throttle slightly during normal operation because of localized heat within the die that can't get removed quickly enough from that point.
I call BS on that sentiment. If any card throttles when it's new, it's due to poor system cooling, poor heatsink mount from the OEM, or it's defective. GPUs are not supposed to throttle under regular conditions.

As for the statements about framerate beyond a display's refresh rate, I agree. There's no need for that, at least not in excess. Typically a game is never able to produce a perfectly stable framerate as one frame possibly takes longer to render than another, so to get rid of those dips people run framerates higher than refresh or they run adaptive vsync, gsync, etc.
 
So I suppose it is entirely coincidental that this am it was fine bit the moment I play this one, a game others have reported the same exact issue, it was damaged. I suppose it it coincidental as well that each instance I tried to lauch the game the card became less and less stable.

If I spend 40 years eating too many hamburgers and not enough healthy food, and then I go run a marathon and have a heart attack, is it the marathons fault or my own ? This is essentially whats going on with your GPU. It was probably faulty or insufficient to begin with, and when put under load, it crumbled. My several year old graphics card was able to handle the strain just fine though.
 
And you say that based on what exactly? You have obviously never heard of Furmark. At any rate, stressing your GPU hardware to the absolute limit for no apparent reason (I mean, look at the game! My GTX1080 should be idling most of the time instead of getting the workout of its life!) may very well reduce its lifespan, in addition to wasting a lot of energy on a game with graphics that should run on a toaster.

My point is that a good computer would have been build to handle running at the limit. When someone builds a PC, especially a gaming PC they should run a benchmark that pushes the system to the extreme for a couple days in order to find out if there are any problems with the hardware or cooling setup.
 
Something strange is going on. I can run the game on the crap integrated graphics that my macbook pro has with totally normal temps. I wonder if there's some sort of issue with newer nvidia cards.

Possibly only newer nvidia cards. It runs fine (albeit on medium graphics presets) on my cheap ass GT model that I paid like $80 bucks for several years ago. I was actually thinking of stopping on the way home tonight to get a newer card so I could crank up the graphics settings, but based on what im seeing from people complaining about overheating cards I think I'll just stick with what I have for now.
 
I think the problem with nvidia is the new 397.31 version. I already played Battletech before the said update without problems at all. Then after I updated I experienced too much fps drop with my other games and with Battletech heating up my laptop. I reverted back to previous version and FPS problem is gone as well as sudden temp rise.
 
Even a water cooling system will have some fans that are going to be loud at 100% speed.
Depends on the way the liquid cooled system is designed. If you are using 200mm 900rpm fans in a push/pull setup, even at max speed (900rpm) the fans are nothing more than a low hum in a case. Now there is a small variance in manufacturer but the fan noise is mostly generated by the air movement and the blades, not the bearings or the motor. Six - 200mm, 900rpm fans running at max speed generate about 40-45dB of sound from about 3 feet away from the case at a frequency of an average household oscillating fan set on the lowest setting. It is not distracting in my opinion by any stretch of the imagination.
 
I call BS on that sentiment. If any card throttles when it's new, it's due to poor system cooling, poor heatsink mount from the OEM, or it's defective. GPUs are not supposed to throttle under regular conditions.
The manufacturer will likely call it some type of dynamic boost. It will run at design limit but conditions can allow pushing beyond a limit under certain conditions, often just for certain parts of the chip. Each part can run at a slightly different rate and this can happen on multiple levels of scale. This is used for thermal management, power management, and boosting parts under load while throttling others in order to keep a total load under a specified limit. Throttling as most people are familiar with is when this performance limit is reduced dramatically to manage heat on a large scale. But inside the chips there is often a balancing act going on constantly within certain constraints although this is largely invisible to the user.
 
Like I said earlier, in the meantime, anyone who has an NVIDIA card can disable VSYNC in-game and force either VSYNC or Fast SYNC through the NVIDIA Control Panel, it really isn't that difficult to do.
 
I think this may have just happened to me, too. I've been having some unexplained issues with BattleTech, weird crashes for no reason, and I don't mean little ones either where you can resolve the issue with task manager. I mean full blown crashes where all control over the PC locks up, the sound loops and artifacts appear on the display.

This has been going on for the last couple of days, maybe once per day, but only when I am playing BattleTech. Things like Tom Clancy's Ghost Recon Wildlands or City Skylines or any other game I was playing in that time was perfectly fine, and I've never, ever had an issue with this card before - it's barely even a year old.

Then when I tried to play BT around...thirty minutes ago, I loaded the game up, went to get some water and came back just in time to see the menu freeze and have the exact same crash, and my PC hasn't been able to boot since. I'm. Having to type this on an iPad, but I'm going to try and do some more looking and get in touch with EVGA before I blame BT for it.
 
I think this may have just happened to me, too. I've been having some unexplained issues with BattleTech, weird crashes for no reason, and I don't mean little ones either where you can resolve the issue with task manager. I mean full blown crashes where all control over the PC locks up, the sound loops and artifacts appear on the display.

This has been going on for the last couple of days, maybe once per day, but only when I am playing BattleTech. Things like Tom Clancy's Ghost Recon Wildlands or City Skylines or any other game I was playing in that time was perfectly fine, and I've never, ever had an issue with this card before - it's barely even a year old.

Then when I tried to play BT around...thirty minutes ago, I loaded the game up, went to get some water and came back just in time to see the menu freeze and have the exact same crash, and my PC hasn't been able to boot since. I'm. Having to type this on an iPad, but I'm going to try and do some more looking and get in touch with EVGA before I blame BT for it.

Trade the card in and get a Zotac, I have had no problems with mine.
 
My GTX 1060 appears fine, but I am having a very weird problem with this forum.

Whenever I click to go from Windows to the Paradox forum my game goes to the Bluescreen of Death and restarts. Never seen it before and it does it about every 1 out of 3 times and only when I click to come here.

Never seen this issue before.
 
Whenever I click to go from Windows to the Paradox forum my game goes to the Bluescreen of Death and restarts. Never seen it before and it does it about every 1 out of 3 times and only when I click to come here.
That's unrelated to the game then. So if you're opening your browser and you visit the Paradox forum you get a BSoD? If so, then I'd recommend troubleshooting why. Could be a driver issue or faulty hardware. I'd do a dxdiag run, get a report, and go from there. Make sure your chipset and video drivers are up to date.
 
That's unrelated to the game then. So if you're opening your browser and you visit the Paradox forum you get a BSoD? If so, then I'd recommend troubleshooting why. Could be a driver issue or faulty hardware. I'd do a dxdiag run, get a report, and go from there. Make sure your chipset and video drivers are up to date.

Ok, I`ll look into it.
 
Status
Not open for further replies.