Random Frame Drops with GTX1080

[ksteib]

Hello,

My MSI AERO GTX 1080 is dropping frames lately, I have been using it for awhile and it was getting to like 87C so I put the NZXT Kraken G12 cooler on it with a Corsair H55, and the temps are now always below 60 - 65C, but since I have noticed my frames don't tend to stay locked with Vsync at 72 (same as my Hz on monitor..) and have lower drops when Vsync is off. (I did have like 3 days of really solid performance once I put the new cooler on it). Mainly tested with games like Rocket League and GTA V, so it is making it very jumpy to play. Any ideas? I did switch my fans around and my CPU is getting closer to 58C now but that shouldn't be awful.. I have the i5-6600k

Thanks.

 

[FauxisFox]

Your CPU temps, and probably load seem fine, so we can probably rule that out. What's the power supply that you're using? It could be supplying dirty voltage or unstable power levels. It could also be that some thermal paste that you applied has leaked onto the GPU board. Not all pastes are dielectric (nonconductive). What was your application method? For GPUs and water blocks, I recommend using the drip and spread evenly with an old credit card method. The heatsink doesn't apply enough pressure to use the rice-grain method. How much did you put? Also, are you using the latest drivers? Do you have the games running in Windowed or Fullscreen(Windowed) modes (they can negatively impact performance)? Are you using a mechanical hard drive? Many, many factors...

 

[ksteib]

Your CPU temps, and probably load seem fine, so we can probably rule that out. What's the power supply that you're using? It could be supplying dirty voltage or unstable power levels. It could also be that some thermal paste that you applied has leaked onto the GPU board. Not all pastes are dielectric (nonconductive). What was your application method? For GPUs and water blocks, I recommend using the drip and spread evenly with an old credit card method. The heatsink doesn't apply enough pressure to use the rice-grain method. How much did you put? Also, are you using the latest drivers? Do you have the games running in Windowed or Fullscreen(Windowed) modes (they can negatively impact performance)? Are you using a mechanical hard drive? Many, many factors...

I have a EVGA 750 Watt gold PSU, I am using the preapplied thermal paste from my cooler because it is new so that may be it I suppose. I always play full screen, updated drivers, and I ave games playing on a SSD as well as a 7200 rpm drive. same thing. GTA is using 100% cores I noticed but it never dropped before recently so I am curious as to why it is a problem now, obviously a game like Rocket League wouldn;t cause that much pain.

 

[FauxisFox]

Your CPU temps, and probably load seem fine, so we can probably rule that out. What's the power supply that you're using? It could be supplying dirty voltage or unstable power levels. It could also be that some thermal paste that you applied has leaked onto the GPU board. Not all pastes are dielectric (nonconductive). What was your application method? For GPUs and water blocks, I recommend using the drip and spread evenly with an old credit card method. The heatsink doesn't apply enough pressure to use the rice-grain method. How much did you put? Also, are you using the latest drivers? Do you have the games running in Windowed or Fullscreen(Windowed) modes (they can negatively impact performance)? Are you using a mechanical hard drive? Many, many factors...

I have a EVGA 750 Watt gold PSU, I am using the preapplied thermal paste from my cooler because it is new so that may be it I suppose. I always play full screen, updated drivers, and I ave games playing on a SSD as well as a 7200 rpm drive. same thing. GTA is using 100% cores I noticed but it never dropped before recently so I am curious as to why it is a problem now, obviously a game like Rocket League wouldn;t cause that much pain.

Your PSU is actually the same as mine, so it can't be the problem. I have yet to have a problem with power quality on EVGA and XFX PSUs. Try running a game with Rivatuner running in the background. Have it monitor your GPU clock, temps, and memory allocation while playing a more demanding game like GTA V. Then look at the results. If there are clock and memory spikes, it's probably because MSI put a custom BIOS that doesn't respond well to low temps. When it sees low temps, it tries to kick in GPU Boost 3.0 and go as high as it can, but it fails when it hits a certain clock speed (the 1080 that you have isn't stable at the target clock speed and freezes) and drops back down to normal speed. Then it will try again after a certain number of cycles. MSI has done this before and made the argument that you shouldn't use an aftermarket cooling block because they already sell one preinstalled (removing the fan shroud and provided heatsink void warranty), which I think is utter bollocks. What you can do to combat this is by using MSI afterburner and locking the clock speeds to stock speeds and test it's stability then. If nothing changes and it still stutters, then it has to be something else, but I suspect that aggressive GPU Boost is the culprit.