Strange things hardware related

[GaspardSavoureuxx]

Hello all.

5 Days ago, i changed my CPU from a r5 2600 to a r5 5600x. Everything went fine.

For the first 2 day after the upgrade, my GPU was like everytime (more than 800hours) at 45-50°C with 65% usage+-.

3 days from now, i had a strange bug where my GPU always was at 65-70°C wherever i was in game. In Limsa with a lot of people or alone in a whole map didn't change a thing. What is really weird is that my temp goes up but my usage goes way down (from 65% to 40%).

I have no fps drop, everything is smooth but i wonder why i gain 20°C on my GPU for nothing...

Everything is up to date and clean. Every of my fan works.

My rig :
R5 5600x - RTX 2060Super - 16GB ram 3200mhz - 1tb nvme - 1080p - MB b450 A-PRO - Bios updated - Win 10

CPU idle temp : +-30°C
GPU idle temp : +-25°C

Did someone have an idea? It's not the end of the world but i liked when the game didn't make my PC running like a jet

Thanks

EDIT : So, the problem seems solved. After disabling the fTPM in the BIOS, everything runs like before now. The fTPM seemed to work great with my r5 2600, but after the upgrade to the 5600x problem happened.

Thanks all <3

[MiniTofu]

I assume you clean your PC interior dust as well.

Reason why because the original R5 2600 was holding back the GPU performance (known as Bottleneck) and when you put a bigger straw (R5 5600X) you remove that limit. That GPU can actually scale up the performance.

[GaspardSavoureuxx]

Hey,

Thanks, but i don't think so.

If my CPU was holding back my GPU, my GPU usage would be higher now and not smaller.

Like i wrote on the OP, the first 2 day after the upgrade was perfect, the same temp on the GPU as before. It's only after 2 days that my gpu went from 50 to 70°C.

It even happend once after a PC restart that the game run at 50°C for all afternoon. After disconnection and reconnection on the evening, the GPU was back to 70°C.

It's software related (ff14/windows/bad app) i don't know. I have windows 10 only for gaming and don't have much of app installed.

And yes, everything is clean, i clean dust once a month, the inside of my PC is like new

I just wonder if other people like me had this problem and if they found a possible fix.

[MiniTofu]

Not sure, could be the GPU is doing some work for the CPU since it was bottlenecking. FFXIV is CPU side, I only hit around 80fps in Limsa with my 3080FE+10600k.

Screaming at 74'c. I did set a power limit on my 3080FE to keep the fan from making noise over 60%.

[GaspardSavoureuxx]

I understand what you say.

The only illogical thing is that my GPU usage is 25% less now and have 20°C more. That's weird.

[Kirika-]

Likely what has happened:

• Before CPU upgrade, the CPU was the system bottle neck, meaning the GPU was waiting (not working) for the CPU to catch up, which results in lower GPU utilization and lower GPU temperatures over all.
• After CPU upgrade, the system OS has to do a lot of internal reworking, which results in the system not actually reaching peak performance for a few days.... the same thing can happen after a major OS update, as in it take sometimes a few days for the OS to "settle in". So, during that time, your overall performance wasn't reaching its optimal rate, so everything GPU utilization and temperature wise looked about the same as before CPU upgrade.
• Then, once Windows finally finished tweaking itself for the new CPU, the GPU was now allowed to run at a much higher performance, but due to the higher performance, it's temperatures went up. Depending on the specific graphics card's BIOS settings for TDP, fan curves, and voltage/frequency values, the higher temperatures can result in the card "throttling down". The result is that you see higher temperatures and lower utilization.

Try using some GPU tweaking software for your specific card, or use EVGA's Precision X1 or MSI Afterburner to "get into" the card's BIOS settings to see where they are and make any tweaks that can help.

It can also be an issue that now the graphics card can run at a higher performance, you are hitting a limit of the cooling system for it and/or the PC case. As you have said the fans are working, I would then double check the case ventilation and maybe disassemble the graphics card to clean and re-paste the heat sink to GPU die contact and check/replace any thermopads for the VRMs and memory chips. Check YouTube for "Video Card Re-pasting" for How-To's on the process.
You can also try running the PC with the case open to see if the GPU temperatures and performance changes.

[GaspardSavoureuxx]

Likely what has happened:

• Before CPU upgrade, the CPU was the system bottle neck, meaning the GPU was waiting (not working) for the CPU to catch up, which results in lower GPU utilization and lower GPU temperatures over all.
• After CPU upgrade, the system OS has to do a lot of internal reworking, which results in the system not actually reaching peak performance for a few days.... the same thing can happen after a major OS update, as in it take sometimes a few days for the OS to "settle in". So, during that time, you overall performance wasn't reaching its optimal rate, so everything GPU utilization and temperature wise looked about the same as before CPU upgrade.
• Then, once Windows finally finished tweaking itself for the new CPU, the GPU was now allowed to run at a much higher performance, but due to the higher performance, it's temperatures went up. Depending on the specific graphics card's BIOS settings for TDP, fan curves, and voltage/frequency values, the higher temperatures can result in the card "throttling down". The result is that you see higher temperatures and lower utilization.

It can also be an issue that now the graphics card can run at a higher performance, you are hitting a limit of the cooling system for it and the PC case. As you have said the fans are working, I would then double check the case ventilation and maybe disassemble the graphics card to clean and re-paste the heat sink to GPU dia contact and check/replace any thermopads for the VRMs and memory chips.
You can also try running the PC with the case open to see if the GPU temperatures and performance changes.

Damn, thank you for that great explanation.

I have 4 fans in my case and the airflow is pretty good.

The 65-70°C is not a problem though, the max temp on afterburner is 83°C. I was just worried about why a big change like that for no appearing reason.

And thus, you gave it to me

My GPU is still in waranty, so i'm not gonna open it, i don't think the thermal paste is at fault now. If it was at 80-85°C then i would worry.

Thanks again for the good explanation, i just need to learn to live with that

<3

EDIT : I use msi afterburner always with rtss, that's how i monitor everything. After looking, my card will throttle starting at 88°C so i'm good

[GaspardSavoureuxx]

Little update...

Now the game runs at 50°C GPU and there is tons of people on my map (30min in).

I don't get it...

The only thing i did is disable fTPM because it have a bug on ryzen. Hope it's gonna stay that way...

[Larirawiel]

MMORPGs are almost always CPU bound. If you upgrade your CPU then the GPU has also more to compute and it will be hotter then before. Because you framerate is also higher. But do not worry, as long as the GPU does not exceed 100° C.



Cheers

[Thukoolo]

Good morning and thank you for posting in the Final Fantasy XIV Technical Support Forums.

This thread will be moved to the 'Closed Concerns' sub forum as the issue has been resolved.

Thank you for visiting the Final Fantasy XIV Technical Support Forums!