Does my aio need repasting and are my temperatures abnormal

[HalfMouldyBread]

My Aio is a Corsair h100x and cpu is a ryzen 7 5800x, my idle temperature even around 1 or 2% utilization is around 45 degrees (Celsius). Is this abnormally high?  (im not sure but i feel like it is) when applying the aio originally a little bit of the thermal paste got wiped off by the packaging but it didnt seem like much so i assume its fine, do i need to repaste my aio? I also ran pass mark to check the temps under load and it maxed out around 90 degrees but i think that might have just been because it was throttling 

[An0maly_76]

That doesn't sound too bad. My 5900X is air-cooled and idles at 29C-39C, 47C-57C nominal, hasn't topped 64C even disabling half my fans. I have a Corsair 4000X case though, with six fans and a Scythe Mugen 5 cooler. I also keep ambient temp around 20C as well. The 5800X, BTW, AMD is on record as saying is safe to 90C under full load, though I wouldn't let it get beyond 75C... probably starts throttling somewhere around 70-80C.

 

FWIW, I use an iCUE Commander Core XT, with a custom fan curve and temp-controlled RGB scheme....

 

Fans ramp 70%-100% from 30C to 80C.

RBG temp monitor settings - Yellow 67C-72C, Blue below 67C, red above 72C

Edited March 19, 2022 by An0maly_76
Revised, more info

[MrSmiles]

Can’t hurt to re-paste considering you noticed some wipe off. Also depends where the aio rad is mounted, the temps in your room, where the pc is...

[MrSmiles]

But it’s not bad as @An0maly_76says considering you have a whole load of headroom still

[GuiltySpark_]

8 minutes ago, HalfMouldyBread said:

My Aio is a Corsair h100x and cpu is a ryzen 7 5800x

A 240 AIO is borderline being enough for a 5800x, at least if you want to take advantage of PB and its boosting capabilities. All 8 cores packed into a single CCD means that's a lot of heat in a small space. 5900x has a tendency to run cooler as its a 6 and 6 layout. What you're seeing is normal for that cooler. Tweak voltage with the Curve Optimizer and a beefier cooler will net you better results. Idle temps aren't helpful for much, however. Throw it under a multicore load with Cinebench R23 and check back.

6 minutes ago, An0maly_76 said:

 probably starts throttling somewhere around 70-80C.

Closer to 90c

 

I have done a per-core undervolt with the Curve Optimizer and with AutoOC, maintain 4.7Ghz all-core while looping Cinebench R23. Temp settles in at 82-83c after 10 minutes of looping. 

[An0maly_76]

7 minutes ago, An0maly_76 said:

The 5800X, BTW, AMD is on record as saying is safe to 90C under full load, though I wouldn't let it get beyond 75C... probably starts throttling somewhere around 70-80C.

 

Just now, GuiltySpark_ said:

Closer to 90c

 

Yeah, I'm confused on this one because I've found multiple articles claiming AMD said the 5600X was safe to 95C, the 5800X / 5900X / 5950X were all safe to 90C -- under full load though. So I'm not sure why it wouldn't start throttling back before it hit that point? Not arguing, just asking. Seems awfully close to the danger zone for my liking.

[GuiltySpark_]

5 minutes ago, An0maly_76 said:

 

 

Yeah, I'm confused on this one because I've found multiple articles claiming AMD said the 5600X was safe to 95C, the 5800X / 5900X / 5950X were all safe to 90C -- under full load though. So I'm not sure why it wouldn't start throttling back before it hit that point? Not arguing, just asking. Seems awfully close to the danger zone for my liking.

Nothing to worry about. I'm still boosting under a full Cinebench load at over 80c. 

 

https://old.reddit.com/r/Amd/comments/jsy8bw/robert_hallocks_response_to_all_zen_3_thermal/

 

If anyone with Zen3 CPU wants to take full advantage of PB/PBO and push their chip to its potential, this guide is wonderful. More power, less heat. All using the tools built into these chips and the B550/X570 platform. 

 

 

[HalfMouldyBread]

9 minutes ago, GuiltySpark_ said:

. Throw it under a multicore load with Cinebench R23 and check back.

 

Just finished running Cinebench a few times through and whilst it did sit at 90 degrees for the whole time i dont think it was throttling at all as it said 100% utilization throughout 

[An0maly_76]

10 minutes ago, HalfMouldyBread said:

Just finished running Cinebench a few times through and whilst it did sit at 90 degrees for the whole time i dont think it was throttling at all as it said 100% utilization throughout 

Good to know, but makes me feel that much better about my temps...

(Link to video clip, has coarse language)

https://youtu.be/ZAmwnD-Sv3c?t=14

 

Edited March 19, 2022 by An0maly_76
Revised, more info

[GuiltySpark_]

11 minutes ago, HalfMouldyBread said:

Just finished running Cinebench a few times through and whilst it did sit at 90 degrees for the whole time i dont think it was throttling at all as it said 100% utilization throughout 

Utilization at 100% is expected when running a multicore load, that's by design. If you were sitting at 90c you were absolutely being limited but temp without clockspeed doesn't really tell you anything. You're going to want to watch your clockspeed AND temp at the same time. I use both HWInfo and Ryzen Master.

 

What I CAN tell you is as I suspected, that cooler is not up to the task of running at chip at 100% load. You don't want it sitting at 90c as its likely pulling back clockspeed to keep it there and from going higher. You might get lucky with an undervolt with the Curve Optimizer but I wouldn't be comfortable with that cooler at all.

 

What was the final score for the Multicore test in Cinebench?

[An0maly_76]

2 minutes ago, GuiltySpark_ said:

What I CAN tell you is as I suspected, that cooler is not up to the task of running at chip at 100% load. You don't want it sitting at 90c as its likely pulling back clockspeed to keep it there and from going higher. You might get lucky with an undervolt with the Curve Optimizer but I wouldn't be comfortable with that cooler at all.

*slow clap*

I have the Scythe Mugen 5, and OP might want to think about the Noctua NH-D15 if overclocking.

[GuiltySpark_]

1 minute ago, An0maly_76 said:

*slow clap*

I have the Scythe Mugen 5, and OP might want to think about the Noctua NH-D15 if overclocking.

Its not even really about overclocking as he's having this issue at stock. In reality, these chips "overclock" by default out of the box with Precision Boost and his (likely) can't even do that if its hitting 90c right off the bat.

[GuiltySpark_]

End of the day, 5800x is a good 280/360 AIO or NH-D15/FUMA 2 chip if you really want to take advantage of it and let it breath.

[An0maly_76]

  

5 minutes ago, GuiltySpark_ said:

Its not even really about overclocking as he's having this issue at stock. In reality, these chips "overclock" by default out of the box with Precision Boost and his (likely) can't even do that if its hitting 90c right off the bat.

 

*shrugs* All I know is that I must have an angel on my shoulder. Everyone else is posting higher temps than I am, even at idle. And even shutting down three of six fans and turning off my air-conditioning for about 20 minutes (dangerous to me as I have breathing problems), I have yet to break 70% of the temp threshold for throttling. Big shout-out @GeorgeMKane, @Somerandomtechyboiand @--SID--for the cooling recommendations that were made in my planning thread.

Edited March 19, 2022 by An0maly_76

[GuiltySpark_]

4 minutes ago, An0maly_76 said:

  

 

*shrugs* All I know is that I must have an angel on my shoulder. Everyone else is posting higher temps than I am, even at idle. And even shutting down three of six fans and turning off my air-conditioning for about 20 minutes (dangerous to me as I have breathing problems), I have yet to break 70% of the temp threshold for throttling. Big shout-out @GeorgeMKane, @Somerandomtechyboiand @--SID--for the cooling recommendations that were made in my planning thread.

The answers are in the numbers, though i'm not sure what the OP's situation really has to do with yours. 5900x due to having its cores split up between two CCD's (6 each) vs 5800x's all 8 in one CCD, means its a known hotter chip. Either way, a screenshot of the top section of Ryzen Master while under a Cinebench R23 load will explain why you're seeing the temps you're seeing. 

 

[An0maly_76]

That's sort of my point. I have a known hotter chip and still seeing lower temps than I've seen anyone else post as of yet with a Ryzen 5000 series. And OP's idle temps are what I see under light load -- with, as you say, a known hotter chip. Just makes me think I must be living right. Glad I took the advice given by the aforementioned members in my last post, might be a different story otherwise.

[GuiltySpark_]

9 minutes ago, An0maly_76 said:

That's sort of my point. I have a known hotter chip and still seeing lower temps than I've seen anyone else post as of yet with a Ryzen 5000 series. And OP's idle temps are what I see under light load -- with, as you say, a known hotter chip. Just makes me think I must be living right. Glad I took the advice given by the aforementioned members in my last post, might be a different story otherwise.

Do you not have a 5900x? Your profile seems to indicate that you do (5800x is the hotter chip). Again though, just being "OK" with the temp doesn't mean much if its performing lower to achieve that and that's only tested with benchmarking and monitoring the aforementioned statistics. Temp alone without clockspeed and power does not tell the whole story.

 

Forgive me i'm not meaning to push but it seems like you're just thinking you've achieved something with low temps without investigating why they're low. This stuff isn't magic and its easily brought to light with testing. 

[HalfMouldyBread]

34 minutes ago, GuiltySpark_ said:

Utilization at 100% is expected when running a multicore load, that's by design. If you were sitting at 90c you were absolutely being limited but temp without clockspeed doesn't really tell you anything. You're going to want to watch your clockspeed AND temp at the same time. I use both HWInfo and Ryzen Master.

 

What was the final score for the Multicore test in Cinebench?

Just ran it again to check the clock speeds and it was running around 4.4 to 4.45, which is defiantly a bottle neck as my motherboard automatically over clocked it and i have seen it go a lot higher, the final Cinebench score was 14332 

[An0maly_76]

10 minutes ago, GuiltySpark_ said:

Do you not have a 5900x? Your profile seems to indicate that you do (5800x is the hotter chip). Again though, just being "OK" with the temp doesn't mean much if its achieving less performance to achieve that and that's only tested with benchmarking and monitoring the aforementioned statistics. 

 

Forgive me i'm not meaning to push but it seems like you're just thinking you've achieved something with low temps without investigating why they're low. This stuff isn't magic and its easily brought to light with testing. 

Yes, I have the 5900X. I guess I misunderstood you, I thought you said the 5900X was a hotter chip than the 5800X, and that the OP seeing higher temps without overclock was an issue. So I'm not so much thinking I achieved something, as being surprised they're lower than what others are posting, that's all. As for investigating why temps are low, I pretty much already know. I didn't play around with my cooling (six 120mm fans plus the cooler with a high-air-flow case) and listened to advice from people with more experience.

 

Check out my planning thread. I was going to use an Asus B550M-K, which did not have sufficient (if at all) VMR cooling, and I was also thinking the Hyper 212 that served my 1700 okay would be fine for the 5900X. Wrong again... LOL. I'm just glad others chimed in and kept me from making those mistakes. They also recommended a case with better airflow, and I happened to find the Corsair 4000X and liked it.

 

Once I was made aware cooling would be an issue with this chip, I pulled out all the stops -- with much better results than expected, I might add. But to be fair, I have a respiratory condition and severe allergies that require I have steady air-conditioning year round -- around 70-71F (31C?) I start coughing, and it gets worse as temp rises, so my ambient intake air temp is 20C, which is of course, a factor in all this.

 

FWIW, I do think OP should consider the Scythe Mugen 5 or Noctua DH-15.

Edited March 19, 2022 by An0maly_76
Revised, more info

[GuiltySpark_]

Just now, HalfMouldyBread said:

Just ran it again to check the clock speeds and it was running around 4.4 to 4.45, which is defiantly a bottle neck as my motherboard automatically over clocked it and i have seen it go a lot higher, the final Cinebench score was 14332 

4.4Ghz isn't great. Cinebench R23 scores should be closer to 15600. 

[HalfMouldyBread]

1 minute ago, GuiltySpark_ said:

4.4Ghz isn't great. Cinebench R23 scores should be closer to 15600. 

yeah i might just have to get a better cooler i suppose 

[GuiltySpark_]

5 minutes ago, An0maly_76 said:

Yes, I have the 5900X. I guess I misunderstood you, I thought you said the 5900X was a hotter chip than the 5800X, and that the OP seeing higher temps without overclock was an issue. So I'm not so much thinking I achieved something, as being surprised they're lower than what others are posting, that's all. As for investigating why temps are low, I pretty much already know. I didn't play around with my cooling (six 120mm fans plus the cooler with a high-air-flow case) and listened to advice from people with more experience.

 

Check out my planning thread. I was going to use an Asus B550M-K, which did not have sufficient (if at all) VMR cooling, and I was also thinking the Hyper 212 that served my 1700 okay would be fine for the 5900X. Wrong again... LOL. I'm just glad others chimed in and kept me from making those mistakes. They also recommended a case with better airflow, and I happened to find the Corsair 4000X and liked it.

 

Once I was made aware cooling would be an issue with this chip, I pulled out all the stops -- with much better results than expected, I might add. But to be fair, I have a respiratory condition and severe allergies that require I have steady air-conditioning year round -- around 70-71F (31C?) I start coughing, and it gets worse as temp rises, so my ambient intake air temp is 20C, which is of course, a factor in all this.

I looked into your planning thread and subsequent threads you made regarding cooling. It doesn't look like you've truly loaded the CPU (Cinebench R23 multicore test, games are not a load test) and monitored it with proven and trusted tools (HWInfo or Ryzen Master). Armory Crate and Corsairs utilites are known to be quite poor and untrustworthy and definitely not something you'd use for proper testing. Sheesh, especially Armory Crate, that thing is basically a virus. 

 

Either way, i'd be curious what you see in HWInfo while looping Cinebench as most if not all the testing you're reading others doing are using that. 

 

I can run a CPU heavy game and never see over 60c, yet Cinebench? Right up to the 80's.

[GuiltySpark_]

3 minutes ago, HalfMouldyBread said:

yeah i might just have to get a better cooler i suppose 

A more appropriate cooler and some tweaking and you're good to go.

[An0maly_76]

16 minutes ago, GuiltySpark_ said:

I looked into your planning thread and subsequent threads you made regarding cooling. It doesn't look like you've truly loaded the CPU (Cinebench R23 multicore test, games are not a load test) and monitored it with proven and trusted tools (HWInfo or Ryzen Master). Armory Crate and Corsairs utilites are known to be quite poor and untrustworthy and definitely not something you'd use for proper testing. Sheesh, especially Armory Crate, that thing is basically a virus. 

 

Either way, i'd be curious what you see in HWInfo while looping Cinebench as most if not all the testing you're reading others doing are using that. 

 

I can run a CPU heavy game and never see over 60c, yet Cinebench? Right up to the 80's.

Actually have run Cinebench and scored 3-5% above typical. Running it again on 30 minute stability test and monitoring... 90-100% usage, most cores at 4174 Mhz, hasn't topped 62C yet. I'll keep you guys posted.

 

Edit: Just re-read this....

 

16 minutes ago, GuiltySpark_ said:

Armory Crate and Corsairs utilites are known to be quite poor and untrustworthy and definitely not something you'd use

for proper testing. Sheesh, especially Armory Crate, that thing is basically a virus.

The plot thickens...

 

 

So is HWinfo part of Win10? Or is that a download I have to get somewhere?

Edited March 19, 2022 by An0maly_76
Revised, more info

[GuiltySpark_]

8 minutes ago, An0maly_76 said:

 most cores at 4174 Mhz, hasn't topped 62C yet.

Preliminarily, that's definitely explaining things a bit. I would expect to see closer to 4.4-4.5Ghz all core. Perusing the 5900x Cinebench R23 score threads on reddit, for stock, you're looking at ~19000 points or so. PBO enabled and some curve optimizer tweaking it should be in the 22000 range.