Awesome RX 590 overclock

[rx590overclocker]

I have a gigabyte rx 590, and after hours of messing with settings i have the perfect overclock found.  the temps never run above 80 in my case, fans aren't super noisy.  GPU clock runs at about 1700 mhz, and VRAM at 2600x4, so 10,400 mhz.  It does suck as much as 135 watts though so keep that in mind.  I am gonna put a link to the file so you can load it as a profile in your radeon settings.   I do not know if this will work on other brands of rx 590s, but I would love to find out.  I have run it for well over 30 minutes in FurMark with no issues, even while doing other stuff such as browsing.  In fact it was running while i typed this post.  Here is the link to the profile, it is also added to this post as a file.  

https://drive.google.com/file/d/19ESgSqR5Cap-EsMJbLFIx-s8hztHy0_1/view?usp=sharing

I have also posted a screenshot of furmark running at the 32 minute mark.

 

1700GPU2600VRAM.xml

[tikker]

Overclocking is subject to the silicon lottery, so there's a good chance your settings won't work for others. Nontheless that's a pretty good overclock from what I can see floating around!

 

Also the obligatory be careful with FurMark. Have you ran an actual benchmark to see how it performs? Something like Heaven or 3D Mark that gives you a score to compare with others.

[rx590overclocker]

4 hours ago, tikker said:

Overclocking is subject to the silicon lottery, so there's a good chance your settings won't work for others. Nontheless that's a pretty good overclock from what I can see floating around!

 

Also the obligatory be careful with FurMark. Have you ran an actual benchmark to see how it performs? Something like Heaven or 3D Mark that gives you a score to compare with others.

yea, furmark ran well, but i tried a different benchmark and it froze up. so i dont think its super stable.  I am still playing with it.  And i will definitely try those benchmarks thx.

[tikker]

19 hours ago, rx590overclocker said:

yea, furmark ran well, but i tried a different benchmark and it froze up. so i dont think its super stable.  I am still playing with it.  And i will definitely try those benchmarks thx.

Those benchmarks might push your card harder than Furmark in some regards. The reason I say be careful with Furmark is because it's not a really a benchmark (unless it's a reasonable scenario for you running your 590 with such a load ). It's a stress/torture test with an extreme unrealistic load. I don't know how AMD cards respond to it, but Nvidia cards with their current protections power throttle when something like Furmark is ran. In the past they even put explicit measures against it in their drivers, because it has destroyed cards This throttling also makes it harder nowadays to judge stability with it.

 

Although in that screenshot you are running at your wanted clocks and not extremely hot either given that it's Furmark + OC.

[Dr0y]

What are your stock clock on that card ?

 

The best thing to do would be to revert to stock and run your benchmark suit (maybe benchmark one or two game you play too) to make a good baseline. 

OC with one of your benchmark to test and push it the most you can.

When you've reached the limit run all the benchmark you did in the first place to see the boost in score and more importantly to see if it's stable in all of them.  One can be stable but not some other one.  Also try some gaming as it can be stable in benchmark but not in one or two particular game. 

 

I had a really fun time oc my rx470 ... I did gain 8% to 10% depending of the test and game I ran ... all that without even touching the voltage.  I found out that I had more gain from memory oc than from core oc.

 

Now I kinda want to try it out with my rx590 too ... I have the sapphire nitro + SE with samsung memory, stock core clock 1560 stock memory clock 2100 ... so it should be interesting to see if it's a good chip or not. 

 

Keep us posted about your OC.

[rx590overclocker]

On 8/22/2020 at 7:36 PM, Dr0y said:

What are your stock clock on that card ?

 

The best thing to do would be to revert to stock and run your benchmark suit (maybe benchmark one or two game you play too) to make a good baseline. 

OC with one of your benchmark to test and push it the most you can.

When you've reached the limit run all the benchmark you did in the first place to see the boost in score and more importantly to see if it's stable in all of them.  One can be stable but not some other one.  Also try some gaming as it can be stable in benchmark but not in one or two particular game. 

 

I had a really fun time oc my rx470 ... I did gain 8% to 10% depending of the test and game I ran ... all that without even touching the voltage.  I found out that I had more gain from memory oc than from core oc.

 

Now I kinda want to try it out with my rx590 too ... I have the sapphire nitro + SE with samsung memory, stock core clock 1560 stock memory clock 2100 ... so it should be interesting to see if it's a good chip or not. 

 

Keep us posted about your OC.

OK, so my stock is 1545 from the factory.  I plan to do some more work on my OC later today, but school is a cram so maybe not.  I will keep working on it tho.  I do plan to switch away from furmark and maybe switch to something else.  Any suggestions?

[Dr0y]

Unigine heaven and superposition

Firestrike

Timespy

 

Those are the ones I used when I OC my rx470