Safe temperatures for RX 580?

[zPanic]

Hi!

 

I recently purchased a Sapphire RX 580 Pulse, and installed it into my computer.

Upgrading from a Sapphire R9 280X Dual-X, it's quite an improvement.

However, I've noticed my new RX 580 goes above and beyond my 280X's max temperature.

The highest I've seen with my old 280X was about 72 degrees Celcius.

The highest I've seen now is 75 degrees Celcius.

 

That being said, I'm not that knowledgable on finetuning my hardware, not even through AMD's WattMan, sadly.

My GPU's max temp in there is 84 degrees Celcius, and target is 75.

Now, this occured when I played Playerunknown's Battlegrounds on ultra with very low shadows, so that might've been it.

I never played PUBG on my old 280X on Ultra before, only on very low.

 

Idle temperature seems to hang around 45-52 degrees Celcius, while on Tom's Hardware, for example, they stress tested a Sapphire Nitro+ RX 580 maxing out at 78 degrees Celcius. Here's another.

Could it be a faulty PSU or bad airflow in my case? My PSU is an Energon 750W CM, and my case is an Aerocool Strike X Advance Black Edition.

 

Would love to hear your replies.

Thanks in advance.

[c0d0ps]

24 minutes ago, zPanic said:

*SNIP*

Try removing the side panel and check your temps again.

[-rascal-]

Some reviews tends to use an open test bench, so their temperatures will be lower compared to a enclosed system (e.g. in a case).

To rule that out, as @c0d0ps mentioned, take the side panel off and monitor the temperatures again -- if it drop, then the problem lies on your case airflow.

 

Also, the RX-580 fans don't spin up during idle / load loads. It only starts to kick once it reaches a certain temperature limit, or when the GPU is being worked (e.g. game).

 

From what I recall from my own Dual-X and Vapor-X HD 7970's (AKA 280X), both of the fans do not COMPLETELY stop.

At least one will still be active even during idle / low loads.

 

Another thing you can look at is making your own custom fan curve / profile...using Sapphire TriXX (or MSi Afterburner, etc).

Something like this (keep in mind, this is my own custom fan curve for my Sapphire NITRO R9-Fury...little different from a RX-580).

Spoiler

 

[zPanic]

14 hours ago, c0d0ps said:

Try removing the side panel and check your temps again.

 

11 hours ago, -rascal- said:

Some reviews tends to use an open test bench, so their temperatures will be lower compared to a enclosed system (e.g. in a case).

To rule that out, as @c0d0ps mentioned, take the side panel off and monitor the temperatures again -- if it drop, then the problem lies on your case airflow.

 

Also, the RX-580 fans don't spin up during idle / load loads. It only starts to kick once it reaches a certain temperature limit, or when the GPU is being worked (e.g. game).

 

From what I recall from my own Dual-X and Vapor-X HD 7970's (AKA 280X), both of the fans do not COMPLETELY stop.

At least one will still be active even during idle / low loads.

 

Another thing you can look at is making your own custom fan curve / profile...using Sapphire TriXX (or MSi Afterburner, etc).

Something like this (keep in mind, this is my own custom fan curve for my Sapphire NITRO R9-Fury...little different from a RX-580).

  Reveal hidden contents

 

 

Thank you both for your comments.

I have tested, and it makes near to no difference to take the side panel off.

Airflow seems to be good in that case.

I wasn't aware most reviews used open bench tests for their reviews, so I suppose these temperatures are fairly normal.

I haven't been able to properly test the card's performance (I mostly play mid-tier games like Playerunknown's battlegrounds, h1z1 kotk, CS:GO, LoL), but I'll give it a shot using The Division.

So far it doesn't seem like much of an improvement over my old R9 280X, but it might just be that I'm not testing it properly.

 

[c0d0ps]

34 minutes ago, zPanic said:

Thank you both for your comments.

I have tested, and it makes near to no difference to take the side panel off.

Airflow seems to be good in that case.

I wasn't aware most reviews used open bench tests for their reviews, so I suppose these temperatures are fairly normal.

I haven't been able to properly test the card's performance (I mostly play mid-tier games like Playerunknown's battlegrounds, h1z1 kotk, CS:GO, LoL), but I'll give it a shot using The Division.

So far it doesn't seem like much of an improvement over my old R9 280X, but it might just be that I'm not testing it properly.

 

Test lots of different games and keep looking at gpu usage. No frame limiter (until you have around 99% gpu usage).

[zPanic]

Just now, c0d0ps said:

Test lots of different games and keep looking at gpu usage. No frame limiter (until you have around 99% gpu usage).

I'll give that a shot.

I don't usually look at GPU usage, mostly CPU.

I always assumed that when the GPU heats up to around it's full load temp (75 degrees for the RX 580 since that's the target in AMD WattMan) I assumed it was working at full load.

Which, to be honest, I find questionable with games such as PUBG, H1Z1, etc.

Just to clarify for myself, as a hardware newbie; GPU temp =/= GPU load?

[c0d0ps]

4 hours ago, zPanic said:

I'll give that a shot.

I don't usually look at GPU usage, mostly CPU.

I always assumed that when the GPU heats up to around it's full load temp (75 degrees for the RX 580 since that's the target in AMD WattMan) I assumed it was working at full load.

Which, to be honest, I find questionable with games such as PUBG, H1Z1, etc.

Just to clarify for myself, as a hardware newbie; GPU temp =/= GPU load?

Higher gpu load will lead to a higher gpu temperature (compared to idle where there is little to no load) but it is not the same.

 

To check if you are using full gpu power aka if the game is bottlenecking by being unoptimised or if the cpu is bottlenecking you are looking at gpu usage.

 

Look up techquickie on youtube I am sure they have something about this (Linus and other people explaining tech stuff under 5-6 minutes).

[chicofrito]

Quote

My GPU's max temp in there is 84 degrees Celcius, and target is 75.

While on Tom's Hardware, for example, they stress tested a Sapphire Nitro+ RX 580 maxing out at 78 degrees Celcius

There is no comparison between Nitro+ and Pulse. The Heatsink from the Nitro+ is much bigger.

 

Heatsink comparison: Sapphire RX 580 Pulse (top) vs. Nitro+ (bottom).