Radeon RX 6700, 6700 XT & 6750 XT users club

[Deleted member 185158]

They seem crazy to me as well but for a completely different reason lol
I bought mine at the peak of crypto BS in May '21.
Threw 900 euro at the thing, but I had just sold my Vega 64 for 700 euro (after getting it brand new for 500 in 2018), so yeah...

Looks like the prices are still settling. I buy a card and the next week it goes on sale. But all the 6000 series cards I've purchased are new and unbeaten. So there's that.
Just don't want to get a mining card used.
Even that RX 580 I have for sale was brand new and I used it once. It's been boxed since. 100 bucks and it doesn't sell? I paid like 225$ for the darn thing...... like it's literally 30 days old

 

[amd64skater]

Lol. Small world!! I stomped around there in my youth a bit. I'm in the NW suburbs these days.

Here's Time Spy.
It'll do 14k if I push to the very edge frequency of 2779mhz. But I don't want to keep dealing with crashes, so backed it down a couple mhz.

View attachment 287677

ok I'm downloading and going to bench. Lets see.

 

[Deleted member 185158]

ok I'm downloading and going to bench. Lets see.

And I'm making a sandwich to get ready for work.
Have run a good handful of benches, so I can pull the screen shots when mobile. Most done with the 6700 XT just over the weekend, so they are pretty fresh.

 

[Chrispy_]

Had the same issue btw. Fixed it by booting in the safe mode, removing OOBE drivers with DDU and installing 23.2.1 downloaded from the AMD website.

Try your luck doing that.

This makes more obvious what I paid for.

The frequency with which Microsoft screw up completely good drivers with Windows update is getting out of control.

It was rarely a problem in W7 days, They started screwing up both drivers and WSUS servers in latter W8 days and it's been dire since about 2019. I've had numerous AMD cards, Atheros NICs, laptop trackpads, and various Bluetooth drivers completely buttf***ed by Windows Update over the last few years and I'm getting real sick of this untested, repeat offending BS from Microsoft.

In an ideal world you'd have granularity in Device Manager to say "See these devices here? DON'T YOU F***ING DARE TOUCH THESE DRIVERS EVER!" and let Windows Update automatically install drivers for everything else. Sadly, automatic driver updates are a global off/on and 90% of the time it's a useful feature that you actually kinda need.

 

[amd64skater]

Lol. Small world!! I stomped around there in my youth a bit. I'm in the NW suburbs these days.

Here's Time Spy.
It'll do 14k if I push to the very edge frequency of 2779mhz. But I don't want to keep dealing with crashes, so backed it down a couple mhz.

View attachment 287677

Barely squeaked by. My GPU has a better score. But your CPU has a better score. Overall We have the same system in other words.

Barely squeaked by. My GPU has a better score. But your CPU has a better score. Overall We have the same system in other words.

Barely squeaked by. My GPU has a better score. But your CPU has a better score. Overall We have the same system in other words.

So my GPU and your CPU should have a baby LMFAO. But really with your CPU and my GPU I swear it should break past 15k+

 

[Deleted member 185158]

View attachment 287704


Barely squeaked by. My GPU has a better score. But your CPU has a better score. Overall We have the same system in other words.




So my GPU and your CPU should have a baby LMFAO. But really with your CPU and my GPU I swear it should break past 15k+

Lol, just wish I could OC further than 103 bus clocks. I updated ME firmware and lost that deal. You can read about it here.

Best motherboard for overclocking Non “K” Intel...

www.overclock.net

On topic, the cpu score isn't a huge impact for some 3D mark benchmarks. Some it is a huge impact.

 

[amd64skater]

View attachment 287704


Barely squeaked by. My GPU has a better score. But your CPU has a better score. Overall We have the same system in other words.




So my GPU and your CPU should have a baby LMFAO. But really with your CPU and my GPU I swear it should break past 15k+

But still I should say 6750XT over 6700XT it's has it's few advantages one from another. The reality is AMD made the 6750XT the way the mid card the way it should have been done. NO 6700 vs 6750 crap. Should have been a XTX variant of some kind. 6700XTX 6750XTX Hmmm still not sure. Also should have put 256 bit vs 192bit.

 

[Deleted member 185158]

But still I should say 6750XT over 6700XT it's has it's few advantages one from another. The reality is AMD made the 6750XT the way the mid card the way it should have been done. NO 6700 vs 6750 crap. Should have been a XTX variant of some kind. 6700XTX 6750XTX Hmmm still not sure. Also should have put 256 bit vs 192bit.

Ok Send your 6750 XT over so it can make babies with my 13700K and we'll give them ghetto names like Boniqua or Lil Aleray!

I've been eyeballing a card for sale at 525$ 6900XT bad mofo. But my 1Kw is trashed and my 850 probably isn't enough to run it

 

[amd64skater]

Ok Send your 6750 XT over so it can make babies with my 13700K and we'll give them ghetto names like Boniqua or Lil Aleray!

I've been eyeballing a card for sale at 525$ 6900XT bad mofo. But my 1Kw is trashed and my 850 probably isn't enough to run it

LOL yeah if I had a 5900x or better maybe I would have a better score. Yeah the 6900 series is bad ass. But really get the 6950 variant.

 

[Kodehawa]

You can still see really funny stuff here sometimes.

 

[siluro818]

You can still see really funny stuff here sometimes.
View attachment 287757

MSFS is like the most unreliable GPU benchmark though. I would use it for CPUs exclusively.

 

[Kodehawa]

MSFS is like the most unreliable GPU benchmark though. I would use it for CPUs exclusively.

Yeah. Still only about 10% better than the 6600XT in average.

I wonder if ROP starvation is worse on the 6700 than on the 6700XT -- can the latter just brute force it better with the higher TMU count? lmao

 

[Kyan]

Nice question. And I already had the answer but I didn't think it was valuable to most. The results below.

MaxClocks (MHz)	1994	2294	2590	2670
Voltage@AMD driver (V)	1	1,15	1,175	1,175
Voltage@GPUz (V)	0,825	0,9	1,05	1,162
Max wattage@AMD catalyst (W)	90	130	175	210

3D Mark Time Spy GPU points#2

10957

12305

13608

13735

Performance ratio vs max #2

79,8

89,6

99,1

100,0

FYI, the 2,6GHz profile has the VRAM at fast timings which helps in getting too close to the fastest of my profiles that the auto-uv profile is.

I've tried to do some test with mine and i was able to run Furmark at 2400MHz with 1.125V and don't draw more than 175W in Radeon Adrenalin. I've never tried undervolt before so i don't know if that's ok.

 

[HD64G]

I've tried to do some test with mine and i was able to run Furmark at 2400MHz with 1.125V, can't use more than 175W in Radeon Adrenalin. I've never tried undervolt before so i don't know if that's ok.

Never run furmark as a benchmark. You won't get valid results that resemple those of gaming. It is a virus-like program that GPU makers have made special settings in thei BIOS for it to not allow to load the GPUs to its max.

 

[RJARRRPCGP]

The 170s for watts? Even older games with the DH-RTGI shader in GShade, can push the GPU harder than that! The graphics are booooring in Furmark, anyways!

 

[Kyan]

Never run furmark as a benchmark. You won't get valid results that resemple those of gaming. It is a virus-like program that GPU makers have made special settings in thei BIOS for it to not allow to load the GPUs to its max.

Any other free benchmark software you can recommend then ? I only know about cinebench for processor. I doubt TPU have a download page for a virus. I can just run a game in full settings to see how it goes.

 

[Super Firm Tofu]

Any other free benchmark software you can recommend then ? I only know about cinebench for processor. I doubt TPU have a download page for a virus. I can just run a game in full settings to see how it goes.

FINAL FANTASY XIV: Dawntrail Official Benchmark

The official site for the FINAL FANTASY XIV: Dawntrail benchmark.

eu.finalfantasyxiv.com

FINAL FANTASY XV WINDOWS EDITION Benchmark | SQUARE ENIX

benchmark.finalfantasyxv.com

UNIGINE Benchmarks

Performance benchmarks by Unigine

benchmark.unigine.com

There are threads here on this site to compare your results.

 

[Deleted member 185158]

I've made an almost complete list on the google docs.
Cards in order 6500 XT, 6600, 6700, 6700 XT
46 benchmarks for comparison
Enjoy.

6000 series benchmarks

docs.google.com

edited so everyone can see it.

 

[AusWolf]

I've made an almost complete list on the google docs.
Cards in order 6500 XT, 6600, 6700, 6700 XT
46 benchmarks for comparison
Enjoy.

6000 series benchmarks

docs.google.com

edited so everyone can see it.

That Speed Way score of 2379 with the 6700 XT is quite impressive, considering that my bog standard 6750 XT scores around the 2330 range.

 

[Deleted member 185158]

That Speed Way score of 2379 with the 6700 XT is quite impressive, considering that my bog standard 6750 XT scores around the 2330 range.

It's definitely overclocked to 2750/2150mhz Gpu/Mem. At most it will do is 2779 on the 4k and 1440p benchmarks. So for most of them, dropped and left it 2750mhz. Certainly would be a higher score with a gigabyte card capable of 2900mhz, I'll have to contemplate that. Ths 6700XT is a week old at 389$ brand spanking new.

 

[HD64G]

Any other free benchmark software you can recommend then ? I only know about cinebench for processor. I doubt TPU have a download page for a virus. I can just run a game in full settings to see how it goes.

Unigine Valley benchmark is one of the best as the ultra low load on the CPU removes it from being a bottleneck and allows the GPU to be pushed to the max. As for Furmark, it isn't a virus that destroys files but it destroys any GPU that lacks the security lock in its BIOS against that type of load that the newest ones have.

 

[Macro Device]

ghetto is the way to go

It definitely is. Decided not only to swap fans but also to make it even more of Frankenstein.

Pic 3 is what I had.
Pic 4 is what I have now.

Completely happy despite no luck applying liquid metal and it became more frightening than I wanted it to be.

"Overstrøm" is Danish for "overcurrent."

 

[RJARRRPCGP]

Unigine Valley benchmark is one of the best as the ultra low load on the CPU removes it from being a bottleneck and allows the GPU to be pushed to the max. As for Furmark, it isn't a virus that destroys files but it destroys any GPU that lacks the security lock in its BIOS against that type of load that the newest ones have.

In 3D Mark 2001 SE, that's more like the meme of Dragothic, unlike the usual Car Chase, (high detail) it's mainly a GPU test, with system bandwidth possibly mattering.

Car Chase, OTOH, is the meme of being related to CPU. (the first map that is benchmarked) Nature is definitely the meme of GPU-related power, shader power matters here, possibly system bandwidth as well.

That FPS action one? "Lobby"? That one seems to be more of a GPU and system-bandwidth test.

Note that I only mentioned these as examples of what emphasize what. It's a very old benchmark now, from the AGP days. And I remember that one the most.

Unigine, both Heaven and Superposition, seem to be much like the "Nature" benchmark in 3D Mark 2001 SE, where CPU doesn't really matter.

 

[siluro818]

RE4 needs 16GB VRAM if you want to run the game with max settings + RT at 1440p.
My 6700XT maintains a 60fps lock easily, but eventually it just crashes to desktop...

 

[Kodehawa]

RE4 needs 16GB VRAM if you want to run the game with max settings + RT at 1440p.
My 6700XT maintains a 60fps lock easily, but eventually it just crashes to desktop...

I wouldn't really run RT tbh.