Review: AMD Radeon RX 5700 and 5700 XT

Why would anyone pair this with an i3 when the R5 3600 exists? You can build a 3600/5700xt computer right now for $900, and that PC will play everything in QHD for less than a single 2080ti.

 

Yes and this can give you RTX 2080TI performance level!
https://i.**********/R0BB6G67/ris.jpg

 

Radeon Image Sharpening: I wonder why not enable for non-Navi cards?

https://www.pcgameshardware.de/Radeon-RX-5700-XT-Grafikkarte-274775/Specials/Radeon-Image-Sharpening-Test-1293637/

 

What we have here?
https://pastebin.com/iUmDXTVP
(save as .fx file and put in a reshade-shaders folder)

 

That video is absurd and if you meant it as some kind of evidence for "where it counts". Or "real budget systems" w/ "review on a 9900k", then it is even absurdly stupid.

1st, Games tested on 4K and 1440p. Averaging from under 30fps range to up to 60fps while having ugly dips.
2nd, That's not real world experience players will get from those cards, because people would either use them on lower resolution or on lower details as values shown range from unplayable to frustratingly bad gameplay experience.
3rd, Gameplay on budget systems with 9900K... Considering how low fps they shown, they could have as well use some much older chip.
4th, Your statement that nVidia's GPUs benefit more from lower end CPUs is product of your imagination till proven otherwise on variety of games
5th, You are mistaken or victim of validation bias of troll-posts claiming that old FX chips have same gaming performance as low end Ryzen.
6th, side note: Those mentioned FX 8C/8T chips today deliver better gaming performance than their contemporary intel's 4C/4T i5s including OCed ones in newer games that actually max out 4 cores or more.

Edit:
I took Paladins as little "Competitive example" - Main menu (rendered animation) Ryzen 2700X:
@4.0GHz -> 530fps
@3.7GHz -> 530fps
@3.2GHz -> 520fps
@2.6GHz -> 490fps
@2.0GHz -> 350fps
@1.6GHz -> 270fps
@1.2GHz -> 220fps

Paladins - Testing Range 6 characters on screen 2700X:
@4.0GHz -> 380fps
@3.2GHz -> 380fps
@2.6GHz -> 340fps
@2.0GHz -> 250fps
@1.2GHz -> 140fps

Spoiler: And here is Test Range @2.0GHz as example.

 

I said it before. There are a few users on here that likely need to be medicated.

 

I have a 1060, so regardless of which one I go to, it wont be a downgraded experience. In fact I wont feel the difference between a 5070xt or 2080s because I will never have a 2080s with those prices. I will enjoy saving 600$ for more important things in life.

 

I just wanted to point out in the reddit /r/ AMD thread about this video - there is an ongoing discussion that the 4K DLSS images don't have DLSS enabled and/or are using different settings. You can tell by the hatch on top of the tank clearly having lower geometry/lod/poly count. Multiple users posted 4K DLSS enabled screenshots of the same hatch without that issue occurring - and in those screenshots the quality is far better than the video. So either hardware unboxed screwed it's test up or something else is going on.

https://imgur.com/a/7wQ1NCi

Compare the above hatch to the one in hardware unboxed video. They clearly are running the game at lower settings or something.

 

Noticed that too. I think they had texture LOD issue/bug. And screws are bump/parallax mapped. Because some textures were more blurry than others.

 

Just curious, what benchmarks are there? All reviews are done with high end CPUs so they dont apply to the point he is making. Any benches done on lower end CPUs that may prove or disprove his point?
(p.s. I have no position in this argument, only curiosity).

 

I ported FidelityFX CAS to ReShade so anyone can use it, with nearly any game
https://www.reddit.com/r/Amd/comments/cc0575/i_ported_fidelityfx_cas_to_reshade_so_anyone_can/

 

Tbh even I saw a performance gain going from 3770k to 1800x. At that point one could have thought 3770k to be weak CPU. I saw a big increase for my 980ti then.

AMD drivers being bad is simply old tales at this point. Sure regarding dx11 they could maybe improve something. But alas that takes resources which they for sure didn't have in the past and do not have now. It won't sell any more cards if they do that it's not a selling point.

 

Now I will repeat what I said a while back..................... Thank God Raja left. Never liked him. Good to see AMD progressing higher and higher as time goes by. Well done Lisa, well done!

 

Yeah, there is something wrong on the video.
Still, the ground on the zoomed picture looks awful on the DLSS image

 

Yep, I'm done here.

I've got far better things to do with my time than wading through page upon page of troll posts.

 

Raja wasn't the reason for the RTG floundering, Navi is as much his work as anything.

Lisa Su's more agressive at putting the money where it needs to go, the last AMD president was not looking broadly enough.

I found this
https://www.gpucheck.com/compare/nv...re-i3-6300-3-80ghz/ultra?lang=en&currency=usd

on paper, the RX 580 is the superior part yet it loses to the 1060 in various games where it should definitely come out ahead, it should be more comparable to a 1070.

 

580 is doing better than the 1060 lul

Edit: where are the base line numbers to compare?

 

Maybe because different RTX GPUs were used? On the HU video a 2070 or 2070S is used.
If someone uses RTX2080Ti maybe would be able to see better processing due to more tensor cores and shaders?
Do you have the link of the discussion to see the specs of the GPU posted that DLSS image?

Update : The DLSS image you posted is from 2080Ti while HU uses 2070.
Which means DLSS results arent the same for all GPUs, however if thats the case, anything lesser than RTX2080Ti looks low quality mess!!!!!!!

 

Putting the numbers into perspective,

with the i3 6300, the 580 trades blows with the 1060.

now compare it to the 1070 with the i3 6300
https://www.gpucheck.com/compare/nv...re-i3-6300-3-80ghz/ultra?lang=en&currency=usd

Now, we change the Rx580's cpu pairing to a 9900k
https://www.gpucheck.com/gpu/amd-ra...e-i9-9900k-3-60ghz/ultra?lang=en&currency=usd

Do you see the problem?

the rx 580 with a 9900k easily compares with a 1070 FE on the i3 6300.

Also, line up the 3 of these at 2560x1440 and take a look

https://www.gpucheck.com/gpu/nvidia-geforce-gtx-1070/intel-core-i3-6300-3-80ghz/ultra/
https://www.gpucheck.com/gpu/amd-radeon-rx-580/intel-core-i3-6300-3-80ghz/ultra/
https://www.gpucheck.com/gpu/amd-radeon-rx-580/intel-core-i9-9900k-3-60ghz/ultra

Good hardware, held back by drivers.

Statistically, the people buying RX cards aren't using them with 9900k's, they are getting paired with Fx and i3's.
There is also some sort of perception that AMD gpu's run best with amd cpu's regardless of the age/performance of the chip

 

What about the rest of the test bench? Do we know what they used? Mb, mem, hd... (I dont see it)

There are too many blanks to fill to take those numbers seriously.

They could be right...or not.