1 2
1988RedT2
1988RedT2 MegaDork
1/21/21 8:41 a.m.

What's the best system temperature monitoring software utility for Win10?

Built son #2 a system for Christmas.  Ryzen 5, Radeon RX5700XT running dual monitors.  I know, I know.  Not the graphics card I wanted, but availability ahead of Christmas was poor, and the Radeon was what I could find in the performance ballpark in which I wanted to be.

System works great, but it shut down several weeks ago while he was playing COD.  I moved the system into a new case by Thermaltake with a literal wall of fans.  Everything worked fine until yesterday, when it shut down again.

I'd like to know if it's the CPU or the GPU that's causing the shutdown.  I know the graphics card is a furnace, but the case doesn't seem hot with all the air flowing through it.  Stock AMD cooler on the CPU is probably my next upgrade candidate, but I don't know what kind of temps he's running.  I know I can view temps in the BIOS, but I'd like something I can put on his desktop.

What's the best system temperature monitoring utility?  What's the best free one?

Thanks!

szeis4cookie (Forum Supporter)
szeis4cookie (Forum Supporter) Dork
1/21/21 8:45 a.m.

I think AMD's Ryzen Master is your ticket to start here for CPU temps. 

Have you experienced crashes in other loads on the system? What happens if you fire up Cinebench R23 and let it loop for a while?

bigdaddylee82
bigdaddylee82 UberDork
1/21/21 8:51 a.m.

The stock Ryzen Wraith cooler is pretty skookum shouldn't have any issues with stock power/clock levels.

Thermal Grizzly's Kryonaut is the best thermal paste I've used, and I've used most of them, I still have a fond nostalgia for Arctic Silver 5.  I don't recall what if any thermal paste or pad the Ryzen Wraith cooler comes with.  If you didn't use any, get some Kryonaut.

If the Ryzen's stock cooler is installed properly, the Ryzen isn't over clocked, and you used a reasonable amount of thermal paste.

I'd focus on the GPU.

Since you're using an AMD CPU & GPU, AMD's System Monitor should tell you everything you need to know about temps.

Edit: tell us about the power supply, what brand and how many watts?

1988RedT2
1988RedT2 MegaDork
1/21/21 9:00 a.m.
bigdaddylee82 said:ps.

Edit: tell us about the power supply, what brand and how many watts?

The PSU is an "Antec Earthwatts Gold Pro Series EA750G Pro 750W Semi-Modular, 80 PLUS GOLD, 120mm Silent Fan, PhaseWave Design, 7 Year Warranty."  Purchased at the same time as the graphics card.

bigdaddylee82
bigdaddylee82 UberDork
1/21/21 9:12 a.m.

Antec is decent, and 750W should be more than enough.

Looks like that card is known for overheating, you have the latest drivers?
https://www.reddit.com/r/Amd/comments/cb0zph/5700_xt_overheating/

BoxheadTim (Forum Supporter)
BoxheadTim (Forum Supporter) MegaDork
1/21/21 9:14 a.m.

Another vote for using Ryzen Master to monitor temperatures.

1988RedT2
1988RedT2 MegaDork
1/21/21 9:37 a.m.

Installed both AMD system monitor and Ryzen Master.

AMD system monitor appears to be pretty dated.  It does properly identify CPU, GPU, and RAM, but only percent usage, not temps.

Ryzen Master does show realtime CPU temps.  Good at idle, but that is expected.  I will see how it goes.  He's still on his school laptop this morning.

1988RedT2
1988RedT2 MegaDork
1/21/21 9:49 a.m.

Latest driver is June 11 of last year.  I'll have to check later, but it's hard to imagine that he's got an older one installed.  I will verify.  Anybody know if there's a way to monitor GPU temps?

szeis4cookie (Forum Supporter)
szeis4cookie (Forum Supporter) Dork
1/21/21 10:04 a.m.

I'd probably use AMD's Radeon Adrenalin or MSI's Afterburner to look at GPU temps/work with settings.

As I'm reading back through your original post, is it that you've only had the two random system shutdowns so far? Can you load the system up to the point where it does it again?

1988RedT2
1988RedT2 MegaDork
1/21/21 10:21 a.m.

In reply to szeis4cookie (Forum Supporter) :

Yah, just the two shutdowns that I know of.  One was after a pretty long session on COD with the video settings maxxed.  I suspect we could do it again.  What am I looking for?

red_stapler
red_stapler SuperDork
1/21/21 10:25 a.m.

MSI Afterburner can monitor both CPU and GPU temperatures.  If you're getting shutdowns, I'd have it log the temperatures so that you'll know what they were when the system shut down.

szeis4cookie (Forum Supporter)
szeis4cookie (Forum Supporter) Dork
1/21/21 10:25 a.m.

Essentially, you're looking for a definitive relationship between load and the shutdowns, as opposed to a random event (maybe a very short power interruption to the power supply, or a software crash).

Try a loop of Cinebench for CPU load, or 3dMark Time Spy for GPU load, to see if you can make a load-related crash happen again: 3DMark Download v2.16.7117 + Time Spy (guru3d.com) 

BoxheadTim (Forum Supporter)
BoxheadTim (Forum Supporter) MegaDork
1/21/21 10:32 a.m.

In reply to 1988RedT2 :

June 11 seems to be way out of date for the drivers. I have a Vega 56 in my PC and I'm pretty sure it's Radeon driver had an update something like a week or so ago.

Are you using the manufacturer-provided drivers or the official AMD drivers? Or the ones from Windows Update?

Edit: You probably want to get your drivers here.

szeis4cookie (Forum Supporter)
szeis4cookie (Forum Supporter) Dork
1/21/21 10:35 a.m.
BoxheadTim (Forum Supporter) said:

In reply to 1988RedT2 :

June 11 seems to be way out of date for the drivers. I have a Vega 56 in my PC and I'm pretty sure it's Radeon driver had an update something like a week or so ago.

Are you using the manufacturer-provided drivers or the official AMD drivers? Or the ones from Windows Update?

 

Good point - AMD's been working on stability challenges with these cards for a while - you probably want the newest Radeon drivers from AMD while you're downloading stuff from AMD's website.

1988RedT2
1988RedT2 MegaDork
1/21/21 12:57 p.m.

Re the drivers, I just assumed that the manufacturer's website would be the best place to get it, and the card is made by Gigabyte.  The June 11 date is from their support.  I will have to check the actual driver version when the power user gets out of class.  99% sure it's whatever driver windows installed for the card right after I fired it up.

fromeast2west
fromeast2west Reader
1/21/21 1:43 p.m.

I've seen RAM issues lead to random crashes under load too.  To rule out a windows issue I'd launch diagnostics from a USB drive and then run a memory check.   

WonkoTheSane (FS)
WonkoTheSane (FS) SuperDork
1/21/21 1:56 p.m.

If testing seems to indicate that it's CPU (Which it really, really shouldn't, a modern CPU should just throttle-down until it's essentially a 386 SX, but maybe it could snap overheat?), I'd pop the heat sink off and make sure there's no peel-off plastic shield that was supposed to be removed before installation.  A lot of the pre-thermal-goo'd heatsinks have a little peel off sticker over the goo, and I've seen 2 computers myself that people have left them on accidentally, and at least a few more on /r/pcmasterrace. 

WonkoTheSane (FS)
WonkoTheSane (FS) SuperDork
1/21/21 2:06 p.m.

A "significantly less user-friendly, but contains a lot more data" program that I've used before is HWiNFO64 (https://www.hwinfo.com/download/).

It's freeware, and it can record in addition to just displaying info.  It will tell you peak temps in addition to current.   I don't have any AMD/ATI video cards handy on my machines, so I don't know how good it can deal with those.  But here's a screenshot of the sensor output on the Surface Pro I'm tying this on:


BoxheadTim (Forum Supporter)
BoxheadTim (Forum Supporter) MegaDork
1/21/21 2:09 p.m.
1988RedT2 said:

Re the drivers, I just assumed that the manufacturer's website would be the best place to get it, and the card is made by Gigabyte.  The June 11 date is from their support.  I will have to check the actual driver version when the power user gets out of class.  99% sure it's whatever driver windows installed for the card right after I fired it up.

Unfortunately that's often not the case. Most manufacturers tend to package the manufacturer's drivers anyway and potentially just add some "extras" that you usually don't want.

You might also want to check that the system has the latest AMD chipset drivers installed as those may also have an impact.

1988RedT2
1988RedT2 MegaDork
1/21/21 3:12 p.m.

Okay, so the actual driver in use by Win10 is dated a few months later, near the end of 2020.  "Update driver" and letting windows search yields "driver is up to date" message.

Should I trust Windows (laugh) or go ahead and install a newer driver from the AMD website if such exists?

I will check on the chipset drivers as well.

WonkoTheSane (FS)
WonkoTheSane (FS) SuperDork
1/21/21 3:15 p.m.
1988RedT2 said:

Okay, so the actual driver in use by Win10 is dated a few months later, near the end of 2020.  "Update driver" and letting windows search yields "driver is up to date" message.

Should I trust Windows (laugh) or go ahead and install a newer driver from the AMD website if such exists?

I will check on the chipset drivers as well.

Whenever the opportunity presents itself, I'd go with the the core technology providers' drivers.  AMD in your case, or Intel, or nVidia, etc.

BoxheadTim (Forum Supporter)
BoxheadTim (Forum Supporter) MegaDork
1/21/21 3:30 p.m.

What WonkoTheSane said.

Hardware manufacturers don't generally get every driver version vetted and approved by MS (and thus they're not on the driver update site). Especially graphics card drivers move quickly as they tend to get tweaked for games very regularly.

1988RedT2
1988RedT2 MegaDork
3/1/21 11:53 a.m.

Update:

We are still having random crashes, but it's happening rather infrequently.  I believe the screen momentarily goes green, and the system crashes and reboots.  Seems to be a driver issue, and I'll get around to wiping and reinstalling someday soon.

But that isn't the reason I wanted to post here today.

In chasing this problem down, we discovered that the CPU is running pretty hot.  Running certain games, it would see temps north of 80 C.  Well that can't be good.  So, open Mr. Wallet and get a Noctua NH-D15 chromax black cooler.

https://noctua.at/en/nh-d15-chromax-black

Yes, 100 clams is pricey for a CPU cooler, but this thing might just be worth it.  Nicely engineered, clear directions, easy install, and it cut CPU temps in half.  You don't often hear me raving about how good a product is.  This is a nice damn cooler.  Just make sure it fits under the hood, because it ain't small!

bigdaddylee82
bigdaddylee82 UberDork
3/1/21 12:16 p.m.

You obviously never lived through the Smithfield Pentium D days if you're worried about 80° C. cheeky

Max temp on a Ryzen 5 3600 is 95° C, if it's seeing peaks of 80° under heavy load, I wouldn't worry too much.

Pull the Wraith off, put some good thermal paste on it, and carry on.

 

 

I had a PD 820 up to 3.4 Ghz (stock was 2.8) ran it at 3.2 most of its life.  That thing was a toaster oven, used a mash up of a dual 120mm and single 80mm radiators in a water cooling loop just for it.

1988RedT2
1988RedT2 MegaDork
3/1/21 12:21 p.m.
bigdaddylee82 said:

You obviously never lived through the Smithfield Pentium D days if you're worried about 80° C. cheeky

Max temp on a Ryzen 5 3600 is 95° C, if it's seeing peaks of 80° under heavy load, I wouldn't worry too much.

Pull the Wraith off, put some good thermal paste on it, and carry on.

 

 

I had a PD 820 up to 3.4 Ghz (stock was 2.8) ran it at 3.2 most of its life.  That thing was a toaster oven, used a mash up of a dual 120mm and single 80mm radiators in a water cooling loop just for it.

Yeah, but I'm working from second-hand info.  When I say "north of 80," that likely means "above 90."

Cooler is bound to be better over the life of the system.  The Noctua is on there and working fine.

1 2
Our Preferred Partners
cAWO8UD6aO4Lvn5fVqbEXgUtfP7BMQqRpDAeq5uzfzrVn2FzQPEtvTXUMX72IIZj