The RX 6000 series Owners' Club

[Felix123BU]

Is it me, or dose the image quality out of the box is better on AMD than on Nvidia ? both in games and in a desktop envoirement. Last time i had an AMD card was in 2014. Since then i only had Nvidia, and after a month or so of using the 6600 XT i switched back to the 1070 Ti to do some tests. And the colors and sharpness are less vibrant and blurrier. I used the same monitor. Is it placebo or ?

And the second thing. I played around with integer scaling. On a 1440p monitor i have to drop down to 720p to make it work but the image quality is poor. And then i tried 1080p + integer on a 4K TV. It looks really good, i am contemplating getting a 4K TV myself and use it as monitor. The only problem is the 60Hz refresh rate. But the bigger size, image sharpness and higher framerate than 1440p make it an interesting option for RTS and RPG type games, maybe even for sims like DCS World or Dirt 2.0
Do you guys have any experience with integer scaling and 4K resolution ?

First, I also had that experience recently, borrowed a friends 2070s to see with my own eyes a comparison between DLSS and FSR, and the first thing I noticed was the slightly washed out colors with the Nvidia card, it was nothing major, but I just felt is was more pale, could have been my imagination. There is a old as time discussion on color compression being different, but I am not really sure if the difference is something clearly quantifiable. And in the end you can adjust colors anyway, so doubt there is any real big differences.

Regarding integer scaling, I can say only one thing, having tried both, FSR is hugely better at upscaling, integer does not come even close in quality. There are apps that can apply FSR on any windowed game, and that is a much better option quality wise from my point of view.

 

[droopyRO]

I forgot about those kind of programs. Someone told me about one, but i found that app doubious since it was in Chinese. And another one that was on Steam for 5$. But at the same time, if they work on Nvidia also, then i should sell this 6600XT and stick to my old 1070Ti that would get a nice boost with FSR. I will make a note and test it out when i'm off work.

 

[Palindrome]

I believe I mentioned it in another thread at some point, but I feel like my Heaven score is somewhat too low..

Spoiler

Any 6600XT owners able to do a run at 1080p max settings so I have something to compare to? My system specs are on my profile, I'm aware that my platform being PCIe 3.0 and the card being 4.0 and only x8 will lower performance somewhat, but I do feel like the card should be scoring higher. Don't get me wrong, I doubled my score coming from an RX580, so I'm very content regardless. Just.... paranoia that something is wrong I suppose

 

[droopyRO]

Shoot me a PM later so i don't forget to run it when i get home.

 

[Palindrome]

Shoot me a PM later so i don't forget to run it when i get home.

I didnt check on this thread till just now, my bad there's no rush though, as long as you can get a benchmark run in with a 6600XT I'll be thankful

 

[droopyRO]

benchmark run in with a 6600XT

Heaven Benchmark i5 8600K@ 5100Mhz, Z370 mobo, 24GB RAM DDR3000, RX 6600 XT Fighter from Powercolor with a slight undervolt.

 

[Palindrome]

Heaven Benchmark i5 8600K@ 5100Mhz, Z370 mobo, 24GB RAM DDR3000, RX 6600 XT Fighter from Powercolor with a slight undervolt.
View attachment 220105

Thanks, that calms my nerves then

 

[OolongGoat]

After having my card for about 6 months now, fixing the Powercolor mounting issue..... and the RGB header issue..... I am very content with my card after some tweaking.

I was still having a 30C delta between my edge temp and my hotspot temp. I took a look at some of the forums around here and looked at undervolting.

Currently I am sitting at a -10% Power Limit and 1100mV while keeping a max clock of around 2500MHz. My hotspot never goes over 80C now. Running 3DMark and other stress tests, my hotspot would get up to 95C.

I don't think a thermal paste swap would help much. I tossed a tube of MX-2 I had that was over 10 years old..... I used the EK Ecotherm this time around on my CPU and GPU. I noticed that the EK paste is a bit more of a thinner viscosity compared to the MX-2 or even the IC Diamond paste I have used in the past.

 

[Felix123BU]

After having my card for about 6 months now, fixing the Powercolor mounting issue..... and the RGB header issue..... I am very content with my card after some tweaking.

I was still having a 30C delta between my edge temp and my hotspot team. I took a look at some of the forums around here and looked at undervolting.

Currently I am sitting at a -10% Power Limit and 1100mV while keeping a max clock of around 2500MHz. My hotspot never goes over 80C now. Running 3DMark and other stress tests, my hotspot would get up to 95C.

I don't think a thermal paste swap would help much. I tossed a tube of MX-2 I had that was over 10 years old..... I used the EK Ecotherm this time around on my CPU and GPU. I noticed that the EK paste is a bit more of a thinner viscosity compared to the MX-2 or even the IC Diamond paste I have used in the past.

Probably the cold plate is not perfectly even, in those cases a thermal pad can be better than paste, but only in limited scenarios. But clocks of 2.5ghz and only 80c on the hotspot is excellent for an air cooler

 

[Dristun]

Ladies, fellas, good day, sorry for barging in. Is there any actual difference in cooling between Sapphire's Nitro+ 6600XT and Pulse 6600XT? Any owners or people lucky to compare here maybe? Judging by the photos on their website the heatsink looks exactly the same - it's just that one of the fans is for some reason reversed on the Nitro+. These two are somehow the cheapest 6600XTs available here now and might just buy one cause 1660 just doesn't cut it for smooth 1440p in 2042 and Vanguard anymore. Thanks in advance.

 

[Valantar]

Ladies, fellas, good day, sorry for barging in. Is there any actual difference in cooling between Sapphire's Nitro+ 6600XT and Pulse 6600XT? Any owners or people lucky to compare here maybe? Judging by the photos on their website the heatsink looks exactly the same - it's just that one of the fans is for some reason reversed on the Nitro+. These two are somehow the cheapest 6600XTs available here now and might just buy one cause 1660 just doesn't cut it for smooth 1440p in 2042 and Vanguard anymore. Thanks in advance.

They do indeed look essentially identical, and I wouldn't expect two coolers that similar to perform noticeably differently. If I were to guess, I'd assume the Nitro has a better binned GPU and possibly more VRM phases populated (i.e. the Pulse might leave one or two out). Sapphire tends to make excellent coolers and GPUs, so I would guess you're getting a good card either way. Just go with whichever suits you best, whether that's down to price, looks, or something else.

 

[HD64G]

They do indeed look essentially identical, and I wouldn't expect two coolers that similar to perform noticeably differently. If I were to guess, I'd assume the Nitro has a better binned GPU and possibly more VRM phases populated (i.e. the Pulse might leave one or two out). Sapphire tends to make excellent coolers and GPUs, so I would guess you're getting a good card either way. Just go with whichever suits you best, whether that's down to price, looks, or something else.

Agreed! The differences for such a low power GPU aren't enough to spend even $30 more on Nitro vs Pulse.

 

[lunatik]

Hi everyone, happened to find this forum looking for answer about mpt.
Anyone know what could be tweaked on 6700xt? Have not touched mpt yet..
Could you only change max power or smth else too? Downloaded 3dmark timespy and with 5600x/6700xt i got 1st place atleast with 1 gpu overall (13459).

 

[Butanding1987]

Hi everyone, happened to find this forum looking for answer about mpt.
Anyone know what could be tweaked on 6700xt? Have not touched mpt yet..
Could you only change max power or smth else too? Downloaded 3dmark timespy and with 5600x/6700xt i got 1st place atleast with 1 gpu overall (13459).

Basically you just increase the GPU power limit (watts). I suggest you leave GFX TDC limit (amp) and SOC TDC limit (amp) alone.

 

[lunatik]

Basically you just increase the GPU power limit (watts). I suggest you leave GFX TDC limit (amp) and SOC TDC limit (amp) alone.

Seems that power limit is not a problem for me since i could not improve my score. I went from 211 to 235 +15% (270W total). Most likely will still try and change smth else when i get the time to learn about it more.

 

[mechtech]

Was hoping to get an RX6600 this Christmas, but it's not looking good lol oh well.

 

[Venom_Chris]

Hello all new here just posting the Information I have to help someone who might google my setup in fufture.

Short background; My last rig consisted of an Asus Ares II gpu and a i7 3960x I've been away from the tech game for that long (about 6 years)

Anyway decided to get back into it with a laptop. Alienware Area 51m (desktop i9-9900k, RTX 2080) It runs HOT with both the GPU and CPU. I'm taking the 9900k hits 100c daily multiple times 101c sometimes.

Decided to try an Egpu with an Asus ROG Strix LC OC 6800xt .

Egpu Is Alienware Graphics amp running at 3.0 4x PCIE.

Card works flawlessly and Is a beast even in an egpu at 3.0 4x PCIE.

My firestrike graphics score Is hitting about 56k+ with only a 2.5ghz overclock, nothing on memory.

Annoyingly though my card can do max clocks just not In firestrike. It will sit at 2755mhz all day In games and stress tests like Superposition 8k, Furmark, Metro Exodus enhanced extreme with ray tracing, OCCT, Heaven, multiple games, etc.

Any Idea why firestrike crashes as soon as It's loaded unless It's down to about 2.5Ghz? It's super annoying. I'm certain this card would do over 2.8 If It wasn't locked off.

Oh, also, I tried It In my laptop with a tired old I7 7700HQ (so a mobile cpu) to see how It scaled - It scales amazingly. This GPU does not discriminate. I'll upload a pic of gpu-z later when on that rig.

Thanks for reading. Hope it's useful to someone In future who's looking for such Information.

Edited* I'm actually running at PCIE 3.0 4x lanes not 8x. Figured I'd correct that. I also hit 2.8Ghz briefly this morning.

Second edit* Thinking about trying MPT but wondering If that will achieve anything as I'm already at max? Assume not and I've no clue what I'm doing with it.

 

[Butanding1987]

Fire Strike needs A LOT of voltage to run at 2,800 MHz on the 6800 XT — definitely more than ~1.25v.

 

[sepheronx]

Was hoping to get an RX6600 this Christmas, but it's not looking good lol oh well.

I just picked one up.

It isn't a good card by any measure. It isn't worth the price and I actually regret the purchase.

 

[Felix123BU]

Hello all new here just posting the Information I have to help someone who might google my setup in fufture.

Short background; My last rig consisted of an Asus Ares II gpu and a i7 3960x I've been away from the tech game for that long (about 6 years)

Anyway decided to get back into it with a laptop. Alienware Area 51m (desktop i9-9900k, RTX 2080) It runs HOT with both the GPU and CPU. I'm taking the 9900k hits 100c daily multiple times 101c sometimes.

Decided to try an Egpu with an Asus ROG Strix LC OC 6800xt .

Egpu Is Alienware Graphics amp running at 3.0 4x PCIE.

Card works flawlessly and Is a beast even in an egpu at 3.0 4x PCIE.

My firestrike graphics score Is hitting about 56k+ with only a 2.5ghz overclock, nothing on memory.

Annoyingly though my card can do max clocks just not In firestrike. It will sit at 2755mhz all day In games and stress tests like Superposition 8k, Furmark, Metro Exodus enhanced extreme with ray tracing, OCCT, Heaven, multiple games, etc.

Any Idea why firestrike crashes as soon as It's loaded unless It's down to about 2.5Ghz? It's super annoying. I'm certain this card would do over 2.8 If It wasn't locked off.

Oh, also, I tried It In my laptop with a tired old I7 7700HQ (so a mobile cpu) to see how It scaled - It scales amazingly. This GPU does not discriminate. I'll upload a pic of gpu-z later when on that rig.

Thanks for reading. Hope it's useful to someone In future who's looking for such Information.

Edited* I'm actually running at PCIE 3.0 4x lanes not 8x. Figured I'd correct that. I also hit 2.8Ghz briefly this morning.

Second edit* Thinking about trying MPT but wondering If that will achieve anything as I'm already at max? Assume not and I've no clue what I'm doing with it.

Your main problem seems to be that the CPU is overheating and throttling down, which can affect gaming performance and benchmarks. I looked at my Firestrike Ultra and my physics (CPU - 28636)) score is 2.7x better than yours, and even though I run a 5800x, it is by no means 2.7 times faster, which probably means your CPU is constantly overheating and performing badly.

As for the 2800mhz, most Navi 2 cards have a hard limit at ~2.7ghz, as @Butanding1987 said, a lot more voltage is needed to get passed that, and its usually not worth it performance wise. The biggest gain from my personal experience is raising the power limit over the maximum allowed with MPT. Best results in Firestrike I got at around 380w TGP used, ~275w being stock max power usage, the performance increase over stock was around 25% with a combo of power and frequency.

 

[Venom_Chris]

Your main problem seems to be that the CPU is overheating and throttling down, which can affect gaming performance and benchmarks. I looked at my Firestrike Ultra and my physics (CPU - 28636)) score is 2.7x better than yours, and even though I run a 5800x, it is by no means 2.7 times faster, which probably means your CPU is constantly overheating and performing badly.

As for the 2800mhz, most Navi 2 cards have a hard limit at ~2.7ghz, as @Butanding1987 said, a lot more voltage is needed to get passed that, and its usually not worth it performance wise. The biggest gain from my personal experience is raising the power limit over the maximum allowed with MPT. Best results in Firestrike I got at around 380w TGP used, ~275w being stock max power usage, the performance increase over stock was around 25% with a combo of power and frequency.

The firestrike ultra screenshot was with a 7700hq mobile cpu to show you can get half decent performance with lesser cpu's . My 9900k runs at 5ghz all core <80c now the RTX 2080 is out the equation.

This is a normal firestrike run with 9900k (hyperthreading disabled keep forgetting to enable it) which i don't think Is too bad considering egpu at 3.0 4x lanes.

And yeah think I'm gonna be happy with what I got and not tinker anymore. It's good performance. Really tempted to sell this laptop and make a 12900k desktop. Give this card the CPU and Pcie lanes it deserves! Love it.

 

[mechtech]

I just picked one up.

It isn't a good card by any measure. It isn't worth the price and I actually regret the purchase.

No cards are good value right now. Also I would only purchase one if I could get it for $475 Cad $ or less. From the charts it’s quite a bit of a jump over my rx480, but my rx480 exceeds my needs so if I have it for another 5 years. Fine by me. More money in my wallet.

 

[Valantar]

I just picked one up.

It isn't a good card by any measure. It isn't worth the price and I actually regret the purchase.

That sounds overly harsh - the reviews I've seen makes it look like a decent competitor. Is there anything beyond the ridiculous state of GPU pricing making you say that?

 

[Space Lynx]

I miss my rx 6800. Thing was such a powerhouse for $579. These markets really suck.

 

[sepheronx]

That sounds overly harsh - the reviews I've seen makes it look like a decent competitor. Is there anything beyond the ridiculous state of GPU pricing making you say that?

Well, the card is a 5700 in performance while costing a small fortune ($660 CAD after taxes at a retailer). I paid same price for a RTX 3060.

I guess if it was about $400 CAD then it would be a better buy for sure.