Fps drops in every game! HELP!

[fbedoya]

I have an Alienware 13 r3 with a 7700HQ and a GTX 1060 6gb, it originally came with 8gb of RAM and 256gb of ssd storage. I upgraded the 8gb to 16gb and bought another 256gb nvme stick. I am running Windows 10 2004, I have had problems with reinstalling windows on this machine and I thought I may have a corrupt installation so I just wiped it and installed a clean image with media creation tool. But the problem still persists. The current installation of Windows only has MS Office, Norton Antivirus, Brave browser and Modern Warfare 2019 which I am using to test this. I used to undervolt the GPU and the CPU to get better thermals (I also thought I was thermal throttling and changed the thermal paste, which did reduce temperatures but that wasn't the problem), but in this installation I'm only using XTU to undervolt the CPU .120 and left the GPU alone. The laptop does get hot, but it doesn't come close to thermal throttling temps, it usually stays bellow 90, and from what I read from HWMonitor (I think) the GPU thermal limit is higher.

 

The problem is that when I'm gaming, I can game for half and hour to an hour with great fps considering the GPU, in normal setting in Modern Warfare getting 70+ fps, and GTA V in ultra getting smooth fps too. But after that first hour, every 3-5 minutes I get an fps drop that lasts about 5-10 seconds and gets as low as 10 fps. After a while, and in a high paced game it gets annoying.

 

I had the impression that it could be my second drive (which I use to install games) since when I was gaming Intel Rapid Storage Technology had a pop up saying that the drive was at risk, but S.M.A.R.T. said everything was okay. In this current installation I have Raid0. I have monitored temps, clock speeds, pretty much everything using HWMonitor, MSI Afterburner and GPUZ and the only thing that really seems weird is that the GPU usage drops from 100% to 20% or lower.

 

Please help!

[FALL1N]

are you playing with battery connected to the power ?

[Hairless Monkey]

Let's get some more data.

 

1. Install HWInfo, (not hwmonitor)
2. Right click the sensors for CPU and GPU temps and clock speeds and select show graph. There should be 10 graphs in total.
3. Play your games for a while with this graphs in the background until the issue occurs.
4. Send a screenshot like this one:

[fbedoya]

2 hours ago, Vova said:

are you playing with battery connected to the power ?

I'm playing with battery connected to the power.

[fbedoya]

2 hours ago, HairlessMonkeyBoy said:

Let's get some more data.

 

1. Install HWInfo, (not hwmonitor)
2. Right click the sensors for CPU and GPU temps and clock speeds and select show graph. There should be 10 graphs in total.
3. Play your games for a while with this graphs in the background until the issue occurs.
4. Send a screenshot like this one:

Thank you so much for your response, I did what you told me and these are the results! Hope they are useful! It seems weird to me that the total CPU usage is highest when everything else drops. Let me know if you know what this is!

Edited June 23, 2020 by fbedoya

[Hairless Monkey]

2 hours ago, fbedoya said:

Thank you so much for your response, I did what you told me and these are the results! Hope they are useful! It seems weird to me that the total CPU usage is highest when everything else drops. Let me know if you know what this is!

It looks to me like a CPU thermal throttle.

 

The CPU core temps slowly ramp up until they reach a threshold temperature, at which point the CPU clocks dive hard and fast to cool off. The relative CPU usage increases because the clock speeds are now lower, and a higher percentage of the available processing power is being utilized. Meanwhile, the GPU clocks and temps dip slightly because they have less work to do because the CPU isn't sending as many frames to render.

 

That's what I see.

[fbedoya]

9 minutes ago, HairlessMonkeyBoy said:

It looks to me like a CPU thermal throttle.

 

The CPU core temps slowly ramp up until they reach a threshold temperature, at which point the CPU clocks dive hard and fast to cool off. The relative CPU usage increases because the clock speeds are now lower, and a higher percentage of the available processing power is being utilized. Meanwhile, the GPU clocks and temps dip slightly because they have less work to do because the CPU isn't sending as many frames to render.

 

That's what I see.

I've already repasted, which was my first time but I felt I did a good job, lifted the backpanel for more airflow, undervolted the CPU, what else can I do?

[fbedoya]

1 hour ago, fbedoya said:

I've already repasted, which was my first time but I felt I did a good job, lifted the backpanel for more airflow, undervolted the CPU, what else can I do?

There is an option in the BIOS to put the fans into performance mode. I activated it and got way better thermals than before, putting me within thermal limit I think, this makes me believe the problem might be something else? I will attach new graphs taken while using this performance mode

 

 

 

[Hairless Monkey]

7 hours ago, fbedoya said:

There is an option in the BIOS to put the fans into performance mode. I activated it and got way better thermals than before, putting me within thermal limit I think, this makes me believe the problem might be something else? I will attach new graphs taken while using this performance mode

In performance mode, your thermals look better and the throttling is reduced in freq. This still points toward a thermal issue. But your CPU should be fine running in the low 80s as it is. So...

 

When you re-pasted, did you move or remove any of the thermal pads? I'm wondering if it is your VRM that is overheating. Does HWinfo show any temperature sensors for the VRM?

[Hairless Monkey]

14 hours ago, fbedoya said:

undervolt the CPU .120

Remove this undervolt, just to eliminate a variable.

[fbedoya]

5 hours ago, HairlessMonkeyBoy said:

In performance mode, your thermals look better and the throttling is reduced in freq. This still points toward a thermal issue. But your CPU should be fine running in the low 80s as it is. So...

 

When you re-pasted, did you move or remove any of the thermal pads? I'm wondering if it is your VRM that is overheating. Does HWinfo show any temperature sensors for the VRM?

I didn't remove or move any thermal pads, and the issue was still there even before I repasted. Also, there isn't a VRM temperature sensor in HWInfo. I did remove the undervolt like you told me to. I've been dealing with this problem for months, please don't give up on me!

[Hairless Monkey]

16 minutes ago, fbedoya said:

please don't give up on me!

Lol.

 

Try underclocking. A steady 3.1Ghz ought to be better than 3.4Ghz with dips to sub 1Ghz. You should be able to use XTU to reduce your boost.

 

If the dips still happen with a healthy underclock, something iffy is going on. Maybe something running in the background. I really don't know.

 

You could also try DDU, but I don't think it will solve this for you.

 

https://www.guru3d.com/files-details/display-driver-uninstaller-download.html

 

If you try it, read and follow all of the recommended instructions.

 

That's all I got. 

[fbedoya]

25 minutes ago, fbedoya said:

I didn't remove or move any thermal pads, and the issue was still there even before I repasted. Also, there isn't a VRM temperature sensor in HWInfo. I did remove the undervolt like you told me to. I've been dealing with this problem for months, please don't give up on me!

 

5 hours ago, HairlessMonkeyBoy said:

Remove this undervolt, just to eliminate a variable.

Since I removed the undervolt the system was running hotter, and I didn't thermal throttle, the indicators on HWInfo says it didn't and the graphs are identical to the ones I've showed before, the drops persisted. Without the performance mode, the GPU and CPU both reached upper 80 but didn't reduce in clock speed, as a matter of fact I got uncomfortable with the high temperatures and corrected the fan curve in HWInfo, and when the system was starting to get cooler, another drop happened. I've tried everything and I'm starting to believe it might not be thermals, I also re flashed the bios.