Zelenia

Okay I got the new memory...I did notice that I got Mortar Max version so it's different to your after all. It's running 3600Mhz but I got only 1 slot out of 2 showing... I hope I didn't break anything. 😞 I try to re-seat the memory.

What a coincidence... I did just order 3600mhz CL18 RAM as well and I thought it would be supported on this same motherboard you got. I'll let you know what happens with mine when I get them.  🤣 
 
Well I do need more RAM anyway. It would've been a worse gamble to buy 2 random 8GB 3200Mhz sticks (I couldn't find exact same ram sticks anywhere new or 2nd hand) and hope they work together with the rest of the build. 

It's the core hotspot temperature so the hottest spot at one point in time on the core.
 
On my 3060 it was 10 C to 12 C higher than the average ...

Hello again, just little update to this topic. 🙂
 
Those Thermalright coolers I saw earlier were actually for older CPUs and didn't fit to AM4 processors... Too bad.
 
I found a second hand Noctua NH-U12S for 40€ (brand new AM4 installation kit included). I did pick it up so I didn't need to pay for mailing. 
 
It's like day and night difference in temperatures and noise... Like full load temps dropped by 10C. I guess the previous cooler just was that much weaker? I don't think the previous paste had dried out since this machine is only 1 year old. 

Hey,
 
I'm planning to upgrade my low profile CPU cooler with Scythe Mugen 5 Rev.B and it's around 50€ in my country.  Seems to be quite decent at cooling and it doesn't get too loud. I don't feel like spending 100€+ to cool down R5 3600.  I mainly will need that extra cooling if I decide to start video editing again or start to do some 3d modeling and other processor heavy tasks as my hobby. When rendering the CPU gets quite hot (84C-95C) as it is and I can't run it at 4,2Ghz.
 
My other options that are available are Noctua NH-U12S (chrome black) and be Quiet! Dark Rock Pro 4 which are both around 75€.
 
Is it worth it to extend my budget or stick with Scythe? My main criteria here is to get "pleasant" humming under load and not a jet engine. 😛 So decibels are fine but it's about the high-pitches... I've been quite happy with Noctua but I feel like they are bit too expensive for what they give.
 
 

I'd personally take a macho over the mugen 5. more thermal mass, better fan. it really a kick ass cooler at a solid price.

I can help you through it. However I don't think it will make much difference if you pull down both voltage and clock speed of the GPU. But yes it's possible. I did something similar first because I was so scared of breaking the card. 
 
Anyway, you can contact me with DM if u wish and I'll walk you through the process. 🙂 I did spend half year playing around with the voltages so I got it memorized pretty much. 
 

Been there, done that.  I've went through exactly the same vicious cycle as you did with that card. Sadly the cooler isn't any good so there isn't much you can do about it.
I was able to tolerate the card for 7 months and then I did end up selling it (the buyer did know it's noisy and still wanted it because of GPU shortage).
 
So no, I don't have any profiles left because I'm using a different card. All I got is that old topic and those values I've listed already. 
 
It's possible to force the card to lower voltages without overclocking but that's basically downclocking. It's almost same thing you've been doing by adjusting "core clock" slider in Afterburner. You will lose some performance by doing that.
 
I got no idea about GPU tweak but you can always try and see if it works for you. 
 
On positive side note, I found myself a new hobby with PC building thanks to this card. 

You don't have to overclock when undervolting. But it's usually done to get more performance from your card AND decreasing the amount of heat generated. So you get a performance boost and the card is running cooler than it would otherwise. And also TUF 2060 got a terrible stock curve, at least the OC one did. I'm not sure if undervolting did affect the lifespan of my card but with the stock settings the fans would've died sooner. 
 
Here is my old post:
 
In this picture I'm "overclocking" the card from 862mv to 950mv. The card was boosting to stabile 1860Mhz. From 950mv to 1050mv (max boost voltage) the card is "underclocking" and staying at 1860Mhz. You can see it's underclocking because MSI Afterburner is showing that light grey line that is suggesting me to drag the curve higher. 🙂 
 
Anyway, chips are different even in same card model so you should try to make your own curve. You can do this with either with the game running in a benchmark mode or use a software that simulates gameplay. You can also run the games with voltage + mhz overlay turned on to see what kind of range is safe for your card.

Don't get that Asus TUF 1660... That cooler is terrible. It can't handle GPU boost without throttling and you'll need to undervolt it. If u check my old posts you see I've posted like 10 topics just to solve issues with this particular model (with RTX 2060 tho). People told me it's just bad airflow in my case or not enough fans etc but the moment I swapped the card out of my build all issues were gone. 
 
It was quite loud at the moment I did purchase it but the noise got louder every month. It also did develop a coil whine over the time which was insult to injury.
 
Ofc there is a chance that I got a bad one but both store and Asus refused to take it back because "it was working as intended".
 
 

I'm bit hesitated to give any advice (I've been trying to write this message for 2 hours now)... Just make sure you got good quality PSU before doing this stuff. 🙂 
 
Anyway, for me it was worth it. I got 2-3°C lower temperatures while gaming on my TUF RTX 2060 so it was great. There was also less stutter because the GPU didn't need to downclock itself. 
 
The downside was that the GPU had sometimes rather odd behavior... like it sometimes would be stuck in some voltage and it refused to come down. I'm not sure if it was Afterburner issue or what. My card also was never able to have stabile 1930Mhz while downclocked. I did try even some manual overvolting/overclocking to see if I can get stabile 1930Mhz but the cooler wasn't able to handle it. 
 
For the numbers...  My 2060 TUF card was working on 0.730Mv-1.060Mv range (idle v.s boost). I was able to get stabile 1860-1870Mhz when undervolted.  You should play around with MSI Afterburner to see what works for your particular card.
 
I got one old post that has the voltage curve I did use but it got a slight overclocking going on. So if you don't want to overclock your card, don't use it. 😛

Yes basically.
How it works is that the gpu boosting goes in intervals based on temperature, meaning if the temperature is lets say between 55 and 60C (random numbers) the gpu will boost to a certain maximum clock speed as long as it has enough power headroom. The maximum clock speed decreases a little bit if the temperature reaches the next level (60-65 in this example). This is independent from the temp limit. So the gpu will clock itself down bit by bit the higher the temperature reaches, regardless of what you do, until it reaches the limit either by power consumption or by temperature. It will then throttle down enough to stay under that limit.
So when you decrease the temp limit, you just set it to throttle down earlier, it can still theoretically boost to 1950MHz or so when the temp is low.
 
 

Yeah, the wires are drawn poorly here. I should've taken a photo instead. ? They are there just mainly to "demonstrate" that they are clumped together here. Fans are spinning properly so I assume they are connected all right. 
 
Hmm well I'm not really 100% sure. No idea how I'd test that. This Ryzen 5 3600 processor was indeed quite pain in the butt at start because it was accelerating all the time. Even when I did move a mouse it had to make some noise so I know it's working hard.  I did change power setting to "balanced" so it did stop doing that. 

Yes, it's the GPU making the noise. I guess it's just that type of GPU that has "high pitch" that is unpleasant. It gets fairly hot too compared to other RTX 2060 around.  I should've checked this before purchase and just get silent GTX 1660ti instead.  Well you can't get good things for cheap.
 
//edit. I'm not actually sure if fans are connected to mobo or not. But they should be running 100% all the time if they weren't, no? 
 
//edit 2. Tbh I don't know anymore. It's probably both GPU and CPU that add to the noise and the case isn't helping. I'll need to look into it later when I get more money. 

I finally did figure this out in case someone else has same issue with this particular GPU. ?
 
I did try to install drivers again, went through BIOS maybe 100 times, replaced case fans with expensive ones and flipped them around, moved cables around for better airflow etc. Yeah the case has better cooling now but it didn't help this GPU much.
 
This did the trick however.
 

 

 
 
Someone might ask, why though? I'm not sure why but GPU was overclocking itself to 1935Mhz (GPU boost 3.0?) when playing any game. This was making GPU to run quite hot and fans spinning like 70-80%. That together with terrible case fans didn't mix well. 
 
Now the GPU clocks around 1400Mhz and stays around 70-75°C under load, the fan speed never goes above 50%. I still can play my games (relatively old ones) in 60-100FPS at 1080p and it's silent. ?
 
And yeah, I'll let the GPU release it's fury when it needs to.

I've actually switched back to the default fan profile also known as auto (disabled the fan curve) and just left the core clock at -502 mhz & and the rest of the settings to their default values as well and now my temps are even better ( -2 / -3 degrees celsius ) + I don't know why but for some reason the default fan profile isn't loud anymore...hmm. And by looking at your fan curve at around 65 degrees it stays at 40% while mine is 45% at 65 degrees as well so it's a small difference anyway. As long as it's under 50% it should be really quiet. ( I do also care about temps since it could increase the lifespan 🙂 )

Welcome to the LTT forums! I'm glad more people are still finding this old topic useful. 🙂 As long as your temperatures are fine and noise levels are under control you don't really need to touch power limit in MSI Afterburner. I wouldn't do that in FPS games anyway since you definitely don't want any performance loss (stutter) v.s other players. I did end up undervolting my card to match the "advertised" clock speed to keep it even more cooler and more silent.  
 
Irony here is that Asus did add "silent mode" button in their next line of TUF cards (and better heatsink). I guess lots of people complained about the noise. 

Your only real option is Afterburner which really doesn't use much RAM like you seem to be worried about. Between it and RivaTuner I'm reporting about 5MB.

@Zelenia I signed-up just to say: thank you so much for these settings! I got the same GPU ( the non-OC version which does the same exact thing XD, and it stays at 1890 MHz ) but with these settings now everything is more quiet. 1 question, should I also modify the power limit to 78%? I only play Rainbow Six Siege and it works pretty good for me as for now, but I haven't changed that option yet.

UPDATE: I went for it, and I'm so glad I did. This is gonna hold me over much better until I can get my hands on a current gen GPU. Might even wait until the next gen. BOCW runs at 1080p max settings well above my monitor's refresh rate, and it can even handle Cyberpunk at medium settings with a very playable 45-60 FPS.

I missed the edit above where you said you've tried older drivers. I think if we find what's causing the CPU usage the problem will be fixed. 

As matter of fact it's very useful,I bought a similar card from Gigabyte.
These cards are very cool,and overclock very well,and the temps are low even after a moderate overclock,which leaves room for further overclocking.
Also lower temps increase the lifespan since the silicon is less likely to fail from thermal stress - which kills a lot of GPUs.
Cooler cards are more power efficient,and as a result the card can boost higher.
 
In short headroom for overclocking and higher lifespan of the silicon,higher boost clocks and better efficiency.

Honestly it makes me mad that retailers and manufacturers just let miners snatch all the cards with bots/scalpers. They really aren't doing anything to stop it for real because "free market". 😡 I don't like BC in general because of how that currency is used in big scale in dark webs (for illegal stuff). Companies selling these cards to miners should at least check backgrounds of the buyers. 

The difference between 120 and 140mm AIO is very marginal. The difference increases when comparing 2x120 and 2x140 or 3x120 and 3x140 as surface area difference ends up being much higher.
 
Tho you might get even better temps with a decent air cooler, not just a massive one. 120 and 140mm AIOs are a bit weird IMO and probably suit the SFF builds best, where the space is limited. In pretty much every other case, a decent air cooler or a bigger AIO is a much better approach.

note: also depends on quality, Some 120mm have better cooling than cheaper 140mm