AMD Radeon RDNA2 Navi (RX 6500, 6600, 6700, 6800, 6900 XT)

[eastmen]

Well of course. Vanilla 6800 Btw is not looking as the good bangs for the buck with the actual pricing, too. We know it is because of the 16 Gig of VRAM but it definitely would have looked better with a 40-50$ lower price.

Yea 16 vs 8 gigs of ram , performance would be close .

Wonder why they choose these prices with ray tracing being bad

 

[sonen]

Thoughts about the stock cooler?

For all Nvidia's innovations and bragging about cooler it doesn't seem to be any worse than 3080's cooler. Best AMD cooler ever.

 

[Kaotik]

Haven't watched yet.

Steve doesn't like it, but pretty much everyone else so far has liked it.
6800 XT is definitely more silent than 3080 FE at least near GPU based on Bitwits video (timestamped)

 

[CarstenS]

Thoughts about the stock cooler?

One of the quietest ref card air coolers yet under load. Not 100% sure if I didn't miss something, but might be going back to GeForce 7900 GTX to find something even quieter.

video (timestamped)

Thanks for that.

 

[Deleted member 2197]

One of the quietest ref card air coolers yet under load. Not 100% sure if I didn't miss something, but might be going back to GeForce 7900 GTX to find something even quieter.

You get some differences among review setups.

Spoiler

https://www.guru3d.com/articles_pages/amd_radeon_rx_6800_xt_review,8.html

 

[Leoneazzurro5]

Yea 16 vs 8 gigs of ram , performance would be close .

Wonder why they choose these prices with ray tracing being bad

Maybe because in "their" upcoming titles they expect ray tracing would not be so bad? This may be an explanation.

 

[CarstenS]

These are the robs not the rasterizer. Rasterizer are only 4 but in driver it is mentioned 8 if you look in the driver.

Performance figures do not look like double the rasterizers. Thanks to very high clocks in those synthies (slightly north of 2500 MHz), it's on par with 2080 Ti and it's six rasterizers, but slightly less than RTX 3080.

 

[j^aws]

From the driver breakdown, please can someone clarify these entries:

num_sc_per_sh 1 1 1 1 1 1 1 1
num_packer_per_sc 2 2 2 2 4 4 4 4

Is SC per SH number of Primitive Units per Shader Array?
And Packer per SC the maximum number of triangles culled per Primitive Unit?

 

[trinibwoy]

So reading all these reviews do you guys think they would have done much better with reviews and impressions if they did

$600 6800xt and $500 6800 ?

Of course cheaper is better but AMD already has the bang for buck badge. I get the impression some reviewers were looking forward to AMD convincingly beating Nvidia like they did to Intel. They came roaring back but didn’t quite pull off the win and lower prices wouldn’t have helped with that sentiment.

 

[CarstenS]

You get some differences among review setups.

Spoiler

https://www.guru3d.com/articles_pages/amd_radeon_rx_6800_xt_review,8.html

Apparently. I can only speak for us: We are measuring with calibrated equipment and our anechoic chamber can reliably go down to 15 dB(A). In that environment, we're seeing 1.3 Sone and 31 dB(A).

 

[CarstenS]

Of course cheaper is better but AMD already has the bang for buck badge. I get the impression some reviewers were looking forward to AMD convincingly beating Nvidia like they did to Intel. They came roaring back but didn’t quite pull off the win and lower prices wouldn’t have helped with that sentiment.

That Intel comparison is a good point. Over all the deserved praise, Ryzen 5000 is getting, people seem to forget that it's main opponent is something like the contingency plan for the continengy plan for the contingency plan should something happen to Intel's original plan. If Intel had executed as planned, Zen 3 would still be a great product, but it would not stomp over the competition like it does in unfolded reality. Nvidia otoh did not screw up that badly, even though their success with Samsung N8 and the accompanying power consumption is debatable at least.

 

[eastmen]

https://www.reddit.com/r/Amd/comments/jwji3n

Ask me anything from AMD

 

[Deleted member 2197]

Apparently. We are measuring with calibrated equipment and our anechoic chamber can reliably go down to 15 dB(A). In that environment, we're seeing 1.3 Sone and 31 dB(A).

That is quiet. ComputerBase.de has slightly higher readings at 37 dB and TechPowerUp matches your readings at 31 dB.

 

[Esrever]

Wonder if faster GDDR would offset the cache misses at 4k.

Raster performance seems pretty good but the drop off at 4k really hurts vs the 3080.

 

[Digidi]

I asked the question about the Rasterizer at Hot Hardware. It was a realy good interview. They realy asked the questions from the audiance this was greate!

 

[techuse]

6800xt has ~50% the compute and bandwidth of a 3090 and offers 85% the performance. The efficiency of this GPU is quite high.

 

[troyan]

Hm... Navi21 has 50% more transistors than GA104 and offers only 25% (6800XT vs 3070) more performance. GA102 has 7% more transistor than Navi 21 on the other hand...

I'm a little shocked that Navi doesnt offer better performance. Will be interesting to see the slower versions and how they will compete against GA104 and GA106.

 

[pjbliverpool]

6800xt has ~50% the compute and bandwidth of a 3090 and offers 85% the performance. The efficiency of this GPU is quite high.

It arguably has more bandwidth with the IC. Also a lot more fill rate.

 

[CarstenS]

It arguably has more bandwidth with the IC. Also a lot more fill rate.

The 288 GPix/s AMD is quoting are Fermi-style.

To clarify: I'm seeing numbers that align extremely close with 4 rasterizers á 16 pixels/clk, only at the end of the pipeline you get 308G depth writes/clk and 612 depth rejects/clk.

 

[techuse]

Hm... Navi21 has 50% more transistors than GA104 and offers only 25% (6800XT vs 3070) more performance. GA102 has 7% more transistor than Navi 21 on the other hand...

I'm a little shocked that Navi doesnt offer better performance. Will be interesting to see the slower versions and how they will compete against GA104 and GA106.

You should wait for the 6900xt numbers before doing the GA102 transistor comparison.