HOT RYZEN 9 5950X TOO HOT?

[Hankieroseman]

My Gamer, built for MS Flight Simulator 2020, is buttery smooth running the defaults. Installed 'Core Temp' 1.17.1 and it's reporting a steady 80º C when flying. The CPU is paired with Noctua NH-D15 Chromax with dual fans. It's been suggested to try re-paste the heatsink before going AIO liquid and possibly having to change the case too. Thoughts and experiences?

 

[toilet pepper]

That's normal as that is a CPU heavy game. And no, an AIO is not "better" than the NH-D15. You can try tweaking the curve optimizer.

 

[Hankieroseman]

That's normal as that is a CPU heavy game. And no, an AIO is not "better" than the NH-D15. You can try tweaking the curve optimizer.

I presume the "curve optimizer" in the BIOS? Otherwise known as EZ System Tuning in ASUS UEFI BIOS Set to Normal. CPU @ idle 32º C.

 

[toilet pepper]

Yes, it is in the bios. I don't have an ASUS board but it usually is in the PBO settings in the bios. With the right dial you can overclock it and undervolt it at the same time.

 

[Franzen4Real]

A repaste/reseating of the cooler is never a bad idea just to eliminate that variable. Don't get too wrapped up in the brand of paste or trying to find the absolute (claimed) lowest temp, but make sure you're getting one that spreads easily and evenly. Some times you can have the "best" paste but not the best temps due to it being hard to work with. Beyond that though, we would need a little more info on your setup. Case? number and orientation of fans? gpu?

I have a 5950X with a 280mm AIO, and I'm in the 60's during gaming. The cooler you have should be more than fine, it is the best air cooler on the market and regularly holds its own against 240 aio's. Assuming you do have a good mount/paste, I would first look at your fan speed/curve and make sure you're not just simply running it too slow for the given load. I see you are using an ASUS board, which I do as well. I use the Fan Expert software that comes with it to automatically set my curves, and then adjust if needed for more silence or more cooling. Then I would look at case air flow and determine if you're getting sufficient cool air to the cpu cooler. Also, if you're running a high wattage GPU then look again at the flow though your case and determine if you are exhausting that hot air efficiently or just recycling it through your cpu cooler. There are many things that could be contributing, but that's where I'd start.

 

[Hankieroseman]

A repaste/reseating of the cooler is never a bad idea just to eliminate that variable. Don't get too wrapped up in the brand of paste or trying to find the absolute (claimed) lowest temp, but make sure you're getting one that spreads easily and evenly. Some times you can have the "best" paste but not the best temps due to it being hard to work with. Beyond that though, we would need a little more info on your setup. Case? number and orientation of fans? gpu?

I have a 5950X with a 280mm AIO, and I'm in the 60's during gaming. The cooler you have should be more than fine, it is the best air cooler on the market and regularly holds its own against 240 aio's. Assuming you do have a good mount/paste, I would first look at your fan speed/curve and make sure you're not just simply running it too slow for the given load. I see you are using an ASUS board, which I do as well. I use the Fan Expert software that comes with it to automatically set my curves, and then adjust if needed for more silence or more cooling. Then I would look at case air flow and determine if you're getting sufficient cool air to the cpu cooler. Also, if you're running a high wattage GPU then look again at the flow though your case and determine if you are exhausting that hot air efficiently or just recycling it through your cpu cooler. There are many things that could be contributing, but that's where I'd start.

Well... I changed my Memory speed D.O.C.P. from default @ 2666 MHz to 3200 MHz. and the temperature changed??

Can't figure that out???

Corsair case Obsidian 450D has 2 x 140MM be quiets in front intake and one Noctua 140 MM intake on the bottom. Has 3 x 120MM (2 x be quiets and 1 Noctua) for top exhaust and 1 x be quiet 120MM exhaust fan rear, in-line with the 140MM Noctua HS Fan on the After cooling fins and a 120MM be quiet fan on the Forward cooling fins pushing air aft to the rear exhaust.

[url=https://imgur.com/cqaUATN] [/URL]
[url=https://imgur.com/oJoM38V] [/URL]

[url=https://imgur.com/ppvWv3X] [/URL]

 

[lexluthermiester]

It's been suggested to try re-paste the heatsink before going AIO liquid and possibly having to change the case too. Thoughts and experiences?

Pictures of the case would be helpful. Reapplying your TIM(heatsink compound) couldn't hurt. With a CPU like that, you need a big case with lots of airflow for that heatsink to be effective. Otherwise liquid cooling with at least a 240mm rad is the way to go.

 

[Hankieroseman]

Pictures of the case would be helpful. Reapplying your TIM(heatsink compound) couldn't hurt. With a CPU like that, you need a big case with lots of airflow for that heatsink to be effective. Otherwise liquid cooling with at least a 240mm rad is the way to go.

Pics posted above.

 

[Franzen4Real]

To me it doesn't seem like the CPU would be getting choked for air just judging by the number of fans and the case you have. Again I guess I would be looking into fan speeds/curves of your front intake fans and the cpu cooler fans and just making sure you're putting enough flow through the case. Also I see you're running a 3090 which definitely puts out some serious heat. Have you tried playing with your case side panel off? You have 3 fans exhausting on top so I'm guessing that could possibly be encouraging airflow out of the gpu heatsink upwards and around your cpu cooler. Perhaps if you try with the panel off and let that beast exhaust all of the hot air outside the case instead of circulating it through the case you may see a difference. This again to just start eliminating the easy to test possibilities. It may just be the combo of a cpu intensive game, and a whole lot of heat inside the case from the gpu, requiring a lot more work/higher rpms from your case fans and cpu cooler.

I jumped on GamePass to grab FS and try it myself to give you a more apples to apples comparison on my temps, but at 152GB download it's ganna be tomorrow evening before I can try it out... my 50mb/s connection is crying at the moment. I just got the AirLink setup on my Oculus up and running and have been looking for some new things to try out, and I'm betting this sim looks incredible in VR.

 

[damric]

80C isn't bad for your setup for FS2020. That CPU will basically turbo its clocks up hard until about 90C or a power limit is reached, and that is by design, and that game is a real CPU eater. I bet it performs very well though compared to most modern rigs which really struggle to run FS2020.

 

[Space Lynx]

80C isn't bad for your setup for FS2020. That CPU will basically turbo its clocks up hard until about 90C or a power limit is reached, and that is by design, and that game is a real CPU eater. I bet it performs very well though compared to most modern rigs which really struggle to run FS2020.

yep, if it were my rig, i would just adjust the noctua fan settings to be a bit more aggressive in the BIOS. that alone will prob bring it down to 75. (if you don't mind the extra noise when at full load)

 

[Morbius2021]

I would also request some zoomed out pictures, it might just be me but it doesn't seem very flow optimized in there. What size is your case?

 

[Space Lynx]

woops, I forgot, not just the fans on the Noctua, but also adjust the fan curve on the exhaust fans. i have mine set very aggressive... and my temps stay very cold. you would be surprised how much a stronger fan curve can accomplish.

 

[Hankieroseman]

I would also request some zoomed out pictures, it might just be me but it doesn't seem very flow optimized in there. What size is your case?

 

[lexluthermiester]

That looks like good airflow. So you're likely just pushing the CPU to it's max and it's hitting it's upper temp range. If it starts thermal throttling, look into some better cooling. Otherwise you're alright, IMHO.

 

[tabascosauz]

My Gamer, built for MS Flight Simulator 2020, is buttery smooth running the defaults. Installed 'Core Temp' 1.17.1 and it's reporting a steady 80º C when flying. The CPU is paired with Noctua NH-D15 Chromax with dual fans. It's been suggested to try re-paste the heatsink before going AIO liquid and possibly having to change the case too. Thoughts and experiences?

Use either HWInfo or Ryzen Master for temps on Ryzen. CoreTemp and HWMonitor are pretty pointless for Ryzen as they can only read the fan control-bound Tctl/Tdie which doesn't tell the whole story of what's going on between your 2 CCDs.

As for airflow, those Shadow Wings fans don't really push much air or have much rpm range either. You might get slightly better performance by replacing the Be Quiet fan on the cooler with another NF-A15, but it's more a recommendation than an actual problem.

But honestly it's normal, the 5900X and 5950X make aggressive use of their 90C throttle limit. Usually get pretty hot in 1- or 2-core loads - so, demanding games basically, not uncommon to see a game pull 15-16W each out of your best 2 cores and reach 80-85C on air despite generally low power draw. Multi-core temps (so, benchmarks) on these two chips is actually pretty tame as long as you have PBO disabled, I'm guessing you probably sit around just 70C in Cinebench.

Ryzen 5000 can reach wildly different temps in "gaming" depending on what you're playing, so take opinions on "gaming temps" with a grain of salt. Something less demanding sees like 65-70C average on my 5900X, while demanding games can reach 85C on warmer days. Conversely, I can play older games from say 2005 and some of them also keep the CPU pegged at 80C, because they're almost entirely CPU-bound from being so old.

 

[erocker]

Well... I changed my Memory speed D.O.C.P. from default @ 2666 MHz to 3200 MHz. and the temperature changed??

Was it for the better or worse? If the temps went up, it's most likely due to increased load/voltage on the memory controller.

 

[Hankieroseman]

Was it for the better or worse? If the temps went up, it's most likely due to increased load/voltage on the memory controller.

CPU temp went down. Now I'm seeing low 50's and collaborated by other software. ??

My new ETEKCITY Infrared Thermometer reports 78º F Intake and 106º F Exhaust air temps. The 3DMark time Spy tests never showed over 62º C in the CPU during tests. So original temps were erroneous or the change in memory speed helped. The exhaust fan turns full blast @1800 RPM. The Core Temp now reports 51-53º C last night while flying the Turboprop French Daher 930 TBM. Looks better now.

Use either HWInfo or Ryzen Master for temps on Ryzen. CoreTemp and HWMonitor are pretty pointless for Ryzen as they can only read the fan control-bound Tctl/Tdie which doesn't tell the whole story of what's going on between your 2 CCDs.

As for airflow, those Shadow Wings fans don't really push much air or have much rpm range either. You might get slightly better performance by replacing the Be Quiet fan on the cooler with another NF-A15, but it's more a recommendation than an actual problem.

But honestly it's normal, the 5900X and 5950X make aggressive use of their 90C throttle limit. Usually get pretty hot in 1- or 2-core loads - so, demanding games basically, not uncommon to see a game pull 15-16W each out of your best 2 cores and reach 80-85C on air despite generally low power draw. Multi-core temps (so, benchmarks) on these two chips is actually pretty tame as long as you have PBO disabled, I'm guessing you probably sit around just 70C in Cinebench.

Ryzen 5000 can reach wildly different temps in "gaming" depending on what you're playing, so take opinions on "gaming temps" with a grain of salt. Something less demanding sees like 65-70C average on my 5900X, while demanding games can reach 85C on warmer days. Conversely, I can play older games from say 2005 and some of them also keep the CPU pegged at 80C, because they're almost entirely CPU-bound from being so old.

[url=https://imgur.com/1Umqwrw] [/URL]

Ryzen Master

 

[Morbius2021]

OK, well , WOW that CPU cooler is just massive !, I honestly thought this might be an MATX or ITX configuration with the initial pictures.

You seem to have adequate intake but I wonder if the backside exhaust doesn't get a bit overwhelmed due to the limited clearance.

Even with that you shouldn't be running so hot.

I just got one of the Ice Giant coolers myself, it would likely be about the same size as your current solution but according to LTT even the Engineering sample of that particular unit
might be what you are looking for. I haven't deployed it yet but according to the testing of the engineering sample by Linus the unit actually "BEAT" a 360MM Liquid cooling system.

You might want to consider a bigger case if you go that direction.

the costs are not terrible for the Ice Giant solution either. About $169.00, they had some issues with getting it out the door, but seem to have deliveries in under 30 days shipping now.

 

[jesdals]

Well temps seems fine to me - but why the fan grill on the intake on the cpu cooler and the out take behind?

 

[Max(IT)]

That's normal as that is a CPU heavy game. And no, an AIO is not "better" than the NH-D15. You can try tweaking the curve optimizer.

Normal ? Using FS 2020 with my 5900X I'm not going over 70/71°
I don't think the 5950X is very different, temperature wise.

Ryzen 5000 can reach wildly different temps in "gaming" depending on what you're playing, so take opinions on "gaming temps" with a grain of salt. Something less demanding sees like 65-70C average on my 5900X, while demanding games can reach 85C on warmer days. Conversely, I can play older games from say 2005 and some of them also keep the CPU pegged at 80C, because they're almost entirely CPU-bound from being so old.

do you remember the game you were playing when read 80° ?
Because I never saw anything like that, playing quite a few different games.

 

[Morbius2021]

TDP wise they are rated the same, 105 Watts, however in testing its been found that the 5950x can draw about 145 watts.

 

[Max(IT)]

TDP wise they are rated the same, 105 Watts, however in testing its been found that the 5950x can draw about 145 watts.

there is little difference between 5900X and 5950X. Surely not 10° if properly cooled.

 

[Morbius2021]

I would agree with that, the only difference is one more chiplet, 4x more cores, I wish I had more performance data regarding the cooler he is using though.

 

[Max(IT)]

I would agree with that, the only difference is one more chiplet, 4x more cores, I wish I had more performance data regarding the cooler he is using though.

yep.
I'm using a Noctua DH15