Question 3600x hitting >100 temps

Julianbuzz

Honorable
Jan 3, 2014
18
0
10,510
0
I recently built a machine using a 3600x with the stock Wraith Spire cooler. I ran some tests to make sure my temps are ok, but I'm getting over 100 degrees in less than 10 minutes in a Prime 95 Blend test. It sits around 82 degrees until about 7 minutes in and then the temperatures skyrocket. I also ran AIDA64 Extreme for 10 minutes, but temps stayed in the 80s.

My question is, are these normal temps for the 3600x with the stock cooler? I understand this cooler has limitations, but I only expected thermal problems with overclocking.

Do I need to reapply thermal paste and reseat the cooler? Or is this normal?

System Specs:
CPU - 3600x stock cooler, no overclock
Motherboard - Asus Prime X570-P
RAM - G.SKILL TridentZ RGB series 16GB (2x8GB) 3200
GPU - Gigabyte GeForce RTX 2090 Super GAMING OC 3X 8G
m.2 SSD - WD Blue NAND 1TB Internal SSD
Case - Thermaltake Level 20MT ARGB
PSU - Cooler Master MWE Gold 750W 80 PLUS Gold
 

eibelbilly

Great
May 6, 2020
138
10
95
1
The stock coolers arent the best but this sounds like you with have a bad fan curve, or a bad fan I would go in bois and see if the fan curve is bad if so reply , make Shure the fan is spinning well,
 
Jan 9, 2020
57
5
45
2
Make sure you have enabled overtemperature protectión in BIOS, then try to read the temps with ryzen master (check the PPT EDC and TFC % these should be below 100%) and hwinfo (check power deviation, high values above 100% could mean that the motherboard power readings are not accurate), if the temps readings are still getting high in both programs, you could try disabling all auto OC settings just to verify, if the problem gets fixed.
 

Julianbuzz

Honorable
Jan 3, 2014
18
0
10,510
0
Make sure you have enabled overtemperature protectión in BIOS, then try to read the temps with ryzen master (check the PPT EDC and TFC % these should be below 100%) and hwinfo (check power deviation, high values above 100% could mean that the motherboard power readings are not accurate), if the temps readings are still getting high in both programs, you could try disabling all auto OC settings just to verify, if the problem gets fixed.
Used both of these programs and turned off all auto overclocking and I am still getting these temps. My PPT, EDC, and TFC are all well below 100%.
 

Julianbuzz

Honorable
Jan 3, 2014
18
0
10,510
0
The stock coolers arent the best but this sounds like you with have a bad fan curve, or a bad fan I would go in bois and see if the fan curve is bad if so reply , make Shure the fan is spinning well,
Checked the fan curve and I'm running at max RPMs when I hit 70 degrees.
 

Zerk2012

Titan
Ambassador
My temperature is in the low 40s for all of these.
Did you do this?
Download OCCT and run it nothing to change just start the test.

If your mainly gaming with the PC your temps should not get higher than OCCT puts it.
i HAVE NO CLUE WHY PEOPLE THINK THE STOCK COOLERS ARE ANY GOOD THEIR TONS OF POST ABOUT HEAT ISSUES. They see a completely unrealistic full load on the processor and freak out.
 
Reactions: Flayed

Julianbuzz

Honorable
Jan 3, 2014
18
0
10,510
0
Did you do this?
Download OCCT and run it nothing to change just start the test.

If your mainly gaming with the PC your temps should not get higher than OCCT puts it.
i HAVE NO CLUE WHY PEOPLE THINK THE STOCK COOLERS ARE ANY GOOD THEIR TONS OF POST ABOUT HEAT ISSUES. They see a completely unrealistic full load on the processor and freak out.
This is exactly why I'm posting. I am not freaking out. I just want to know if this is normal behavior. I just don't know enough about this specific cpu to know if this is normal behavior. I've been running OCCT for about 15 minutes and my temps are sticking around 81 degrees.

Thank you for your help.
 

Zerk2012

Titan
Ambassador
This is exactly why I'm posting. I am not freaking out. I just want to know if this is normal behavior. I just don't know enough about this specific cpu to know if this is normal behavior. I've been running OCCT for about 15 minutes and my temps are sticking around 81 degrees.

Thank you for your help.
81C will limit the CPU boost speed but is not that bad for that program. I would replace the stock cooler on ANY processor I bought their just not really good enough to get the full performance from a AMD processor. OCCT should be a bit more of a load than when gaming.

15 minutes is fine it's not really going to get much higher but you need to account for when gaming your video card is going to be putting off extra heat. I would like to see a AMD processor running in the 70 /72C area to ensure I get the full boost speed.

If money is tight you could run it like that for a while MAKE SURE TO CHECK THE TEMP WHEN GAMING hopefully you will be under 80C more in the 75C area but I would be counting on buying a decent cooler it dont take a lot to be better than stock.
 
Last edited:

Zerk2012

Titan
Ambassador
I will definitely consider that. Is there an air cooler around $50 or less you would recommend? I'm mostly familiar with the Hyper 212.
Fairly good cooler and should be able to still see your memory. Should be fine for a 3600X.

https://pcpartpicker.com/product/8GBrxr/scythe-mugen-5-rev-b-5117-cfm-cpu-cooler-scmg-5100

Your board with that cooler if you look at the picks you can see all 4 memory slots.

https://pcpartpicker.com/b/dsk6Mp

A lot of people are using the esports dual but it will block some memory slots hate to loose your RGB.
 
Last edited:
Reactions: FoxVoxDK

Julianbuzz

Honorable
Jan 3, 2014
18
0
10,510
0
Fairly good cooler and should be able to still see your memory. Should be at least good for 10C difference.

https://pcpartpicker.com/product/8GBrxr/scythe-mugen-5-rev-b-5117-cfm-cpu-cooler-scmg-5100

Your board with that cooler if you look at the picks you can see all 4 memory slots.

https://pcpartpicker.com/b/dsk6Mp

A lot of people are using the esports dual but it will block some memory slots hate to loose your RGB.
Thank you! I really appreciate the suggestion.
 

Karadjgne

Titan
Ambassador
I'm not a fan of occt. Not for temps anyways. Apart from the cut off temps settings being useless additions, it uses linpack which is an extreme stress tester that includes a decent amount of ram. Because of the way it works during conversions, the load changes every so often so it's not a constant load. Which is bad for baseline temps. Linpack was created to replace p95 for Stress testing, not temp testing and is brutal enough to drive a cpu upto an artificial 130%±. And then drop to 80%. Occt, IBT, LinX also use outdated binaries from 2012, the newer Linpack Extreme is considerably better for Stress tests.

Prime95 small fft with AVX, AVX2 (256) disabled (or version 26.6 which has no AVX) gives a continuous 100% load using the same instruction sets games use. That's going to give you a single target that most games won't breach. There are some games that use AVX in small amounts with PhysX for particals during explosive scenes etc, but it's not enough to alter temps to any major degree.

Not sure why Blend decided to skyrocket temps after 7 minutes of test, even Blend is only larger unit with more ram usage, but still a stable output. That's not a normal occurance. So if anything, that'd be one test I'd repeat until the cause is found, because it's possible that since p95 uses the same instructions as games, then a game could possibly see the same 7 minute result. If it's bad sensor readings, change in bios settings, power plans, I'd want to know.
 

Zerk2012

Titan
Ambassador
Thank you! I really appreciate the suggestion.
If you get that cooler you have to clean off all the thermal paste off the CPU, don't use the normal recommended pea size amount in the middle of the processor like on Intel chips. Either make a X or use the pea size but spread it out AMD processors needs more of a full coverage because of the way their made.
 

Zerk2012

Titan
Ambassador
I'm not a fan of occt. Not for temps anyways. Apart from the cut off temps settings being useless additions, it uses linpack which is an extreme stress tester that includes a decent amount of ram. Because of the way it works during conversions, the load changes every so often so it's not a constant load. Which is bad for baseline temps. Linpack was created to replace p95 for Stress testing, not temp testing and is brutal enough to drive a cpu upto an artificial 130%±. And then drop to 80%. Occt, IBT, LinX also use outdated binaries from 2012, the newer Linpack Extreme is considerably better for Stress tests.

Prime95 small fft with AVX, AVX2 (256) disabled (or version 26.6 which has no AVX) gives a continuous 100% load using the same instruction sets games use. That's going to give you a single target that most games won't breach. There are some games that use AVX in small amounts with PhysX for particals during explosive scenes etc, but it's not enough to alter temps to any major degree.

Not sure why Blend decided to skyrocket temps after 7 minutes of test, even Blend is only larger unit with more ram usage, but still a stable output. That's not a normal occurance. So if anything, that'd be one test I'd repeat until the cause is found, because it's possible that since p95 uses the same instructions as games, then a game could possibly see the same 7 minute result. If it's bad sensor readings, change in bios settings, power plans, I'd want to know.
I had much rather tell somebody that does not really know what to run to run OCCT than just hearing they ran Prime 95. Nothing specific to do just run the test if your temps are in a OK range then gaming should be really no big deal.

Your average user knows nothing about downloading specific versions of prim 95 or what settings to set. To me when people are having problems and don't specify what their actually running it just makes it easier to me as I have depended on it for years.

EDIT no disrespect meant or taken from any comments .

In general it takes the guess work out for me when talking about load temps. Everbody can have their own opinion but when you see people talking about high temps under stress I will recommend OCCT every time because I know actually what their looking at then. (I have had a mod PM me on this before and I said the same thing to them they could not really disagree)

To sum it up I take out the unknowns and insert what I know. Is this the 100% right way well yes and no all I know if you just download it and run it with no changes I can go from their without any more information because it's what I use. All I need is the highest temp they seen running it.

Other potential problems, motherboard defaulting to to much voltage, cooler not installed properly. Nothing I can do about that I just use what I know and using the stock cooler is not really a good thing unless you limit the boost clock they actually sux and people for some strange reason think their fine, any cooler that will limit the full boost speed of the processor is just junk in my opinion.
 
Last edited:

ASK THE COMMUNITY