AMD Radeon RX 6950 XT Reference Design

[80-watt Hamster]

I found myself wondering how the RX 6600 fell so far down the P/P charts, then noticed that it's listed at the same price as the 6600 XT ($380). 6600 is USD300 if going by the lowest on Newegg.

 

[HD64G]

Kudos to both @W1zzard and the viewer that provided you with this monster to review. AMD did a good job improving their top GPU without losing efficiency. AIBs almost always are lazy and just turn the notch to 11-12 to get as much performance as possible. I would prefer the BIOS of this ref model in a better cooler as Sapphire's. I wonder why the AIB's don't all have BIOs switches with one of those having the ref one installed.

 

[TheEndIsNear]

The borderlands 3 results can't be right unless there is a huge difference between dx 11 and 12. I've used 12 since it came out. I get 130fps on badass settings at 2k on average on a 6900xt. It never dips as low as it is in the chart. I wonder what could be the difference. Oh I have a 12900k but that shouldn't make a difference like that.

 

[rusTORK]

@W1zzard, are there any chance to see veeery tiny comparison\review of Navi 21 XTXH vs Navi 21 KXTX especially after new "preview" driver from AMD will be WHQL? It's look's like boost few DX11 games.
I just saw Hardware Unboxed video

Also, interesting to see them with Ryzen 7 5800X3D in Test system.

 

[Oberon]

Just woke up and retested this for you on the RX 6900 XT

http://imgur.com/oyWZKbi

(ignore the multi-monitor numbers, I was too lazy to plug in an out the 2nd cable)

You must have a bum card, then. Other outlets with similar testing (mainly igor's lab) don't find the same results even at smaller time scales and no other cards you have tested demonstrate the same behavior.

 

[jesdals]

Great test

Love to se the comparison to the AIB models clearly the better choise if one have the space for a 3-4 slot 32cm long card.

Would like to see the actual backplate taken of because that on my reference card i did find that adding termal pads did lower the temps 3-7c

Not my card (I used Thermal Grissly pads) but a reference with pads

 

[R-T-B]

You must have a bum card, then. Other outlets with similar testing (mainly igor's lab) don't find the same results even at smaller time scales and no other cards you have tested demonstrate the same behavior.

Cards vary. We still have to report what we find.

 

[Oasis]

...kinda, you're just confused about who's who and what...

so 19% yes, assuming you base your math on the chart in the table which I pasted below for reference, but that says that it is the the 3090ti being 19% less efficient than the 6950XT, not the other way around.

So the one being "19% less efficient" is 3090ti, for 100% - 81% = 19%, ezpz

On the other hand if you want to talk about the "6950XT being more efficient than" then you must go beyond addition and subtraction and use multiplication and division too, for now you must assume that 3090ti is at 100% efficiency scale in the chart and then multiply that 100% by the ratio between the previous 6950XT 100% and the previous 3090ti 81%, so

100 * ( 100 / 81 ) = 123,46%

so

the 6950XT is 23.46% more energy efficient than the 3090ti

---
Ref Table from the article:

Yeah I mixed it up

the 3090ti is 19% less efficient than the 6950XT and the 6950XT is 1.234567901 (1.23) 23% more efficient than the base 3090 ti

I would be interesting to see how low of a PSU power spec you can get in a modern gaming pc with the 3090 ti/6950XT without tripping the PSU or shutdowns

Like a 5800x/i7 12th gen and the GPUs

 

[Agent_D]

I'm curious to know which game test you're using when you do the temperature/fan noise charts.

My 6950XT reference fans hit 1800-1950rpm with a steady 1150-1200mv during BL3, Time Spy, and Cold War. Your chart shows ~1031mv on this card you tested, so I'm just curious as to what the testing conditions are.

 

[kapone32]

Yeah I mixed it up

the 3090ti is 19% less efficient than the 6950XT and the 6950XT is 1.234567901 (1.23) 23% more efficient than the base 3090 ti

I would be interesting to see how low of a PSU power spec you can get in a modern gaming pc with the 3090 ti/6950XT without tripping the PSU or shutdowns

Like a 5800x/i7 12th gen and the GPUs

MSI did a live stream using a 12900KS and 3090TI. In some benchmarks the power draw was 900+ watts. It even went over 1000 for a few seconds. They were using a 1000W PSU.

 

[KatyaKassin]

Yesterday i recieved my 6950XT from AMD and i am very happy with it. It has an incredible performance, is quiet and the heat output ist not as much as everywhere said. Hope to have it working in my pc for the next two years.

 

[57er6uigyuholn]

Hi,

Now that some time has passed, does anyone have a PCB shot of this reference 6950 XT? Would like to see if there are any changes to capacitors/filtering and VRM setup to accommodate larger power usage.