Big Navi, on par with 2080Ti/RTX3070 or head to head with 3080?

Stoly

FPS Regular
Joined
Sep 18, 2020
Messages
1,338
Points
113
For a while there has been numerous rumors on Big Navi performance. They go from slightly slower than a RTX2080Ti to even faster than a 3080.

We don't really know the specs, but are pretty certain on memory size and bus width. Supposedly Big Navi will sport 16GB of ram on a 256bit bus.
We also know that at least for now, nvidia has exclusivity on GDDR6X

So best case scenario Big Navi has 16Gbps GDDR6 (more likely 15, but lets say 16) for a bandwidth of 512GB/s, that's higher than the bandwidth of the RTX3070 (reportedly 448GB/s), but still shy from the 2080Ti and farther down from the RTX3080.

Even if Big Navi is a more efficient architecture there's just not enough bandwidth to match the RTX3080, keep in mind that historically AMD has been more bandwidth hungry than nvidia. Heck even 1,024GB/s wasn't enough for Vega VII.

IMO Big Navi could be somewhat faster than the RTX3070/RTX2080Ti but far below the RTX3080, specially at 4K.

So what's you take?
 
I believe that Big Navi will probably exceed the performance of the RTX 2080 Ti and 3070. However, I feel it will fall well short of the 3080. Of course, how its positioned price wise will determine how successful the card will truly be. I suspect, AMD will do alright with that.
 
AMD has a ton of community goodwill right now, and this is a huge opportunity for them to gain some more market share.

They seem to be cautious, avoiding hype etc and I think that IF they have a top notch product in the works they should be a bit more aggressive about promoting it. **** the torpedoes etc...

Most ppl seem to agree on AMD falling just short of the 3080 in performance as a best case scenario, I really can't disagree much as AMD certainly have not indicated otherwise.
 
I dunno, I'm going to give the performance nod to Nvidia - mostly because they've been the performance leader for so long now. However, I'm expecting Big Navi to trade blows with the 3080 on select titles.

Whomever gives me the best bang/buck (and stock!) at 4k60fps will likely get my money this year.
 
I believe that Big Navi will probably exceed the performance of the RTX 2080 Ti and 3070. However, I feel it will fall well short of the 3080. Of course, how its positioned price wise will determine how successful the card will truly be. I suspect, AMD will do alright with that.

For decades (I literally waited years to say that :D:D:giggle::giggle: ), nvidia has set the price and AMD has been forced to follow suit. Big Navi better clearly beat the RTX3070 if it spects to sell at the same price.
But hey, if this creates a price war, fine by me.
 
For decades (I literally waited years to say that :D:D:giggle::giggle: ), nvidia has set the price and AMD has been forced to follow suit. Big Navi better clearly beat the RTX3070 if it spects to sell at the same price.
But hey, if this creates a price war, fine by me.

I'm not sure AMD would come out on top in a price war.

Problem with Radeon as of late is that while great bang for the buck, its overall underwhelming for enthusiasts.
But then enthusiasts are only a small slice of the pie.

Time will tell if Radeon will continue to be the budget darling or if they step up in performance and price.
 
If the price is right I’m totally ok with “just” 2080Ti performance.

Right now I admit I have my eye on a 3070, pending reviews. But I am totally open to a Red card - the 5700XT just wasn’t quite a big enough jump from what I am currently running (at the money I wanted to spend) to get me to bite.
 
If the price is right I’m totally ok with “just” 2080Ti performance.

Right now I admit I have my eye on a 3070, pending reviews. But I am totally open to a Red card - the 5700XT just wasn’t quite a big enough jump from what I am currently running (at the money I wanted to spend) to get me to bite.
Does RT and DLSS type stuff matter to you or you just looking for best raster for the buck?
 
Does RT and DLSS type stuff matter to you or you just looking for best raster for the buck?
Does not matter. I've been running a 980 for a while at 4K, if that tells you anything. I have been ok with turning settings down and running at "just" 30-60 fps. Most of what I play aren't heavy action titles anyway. ~BUT~ I have been holding out for HDMI 2.1 / DP 2.0 - since whenever I do another round of monitor upgrades (or, as it's looking more likely, another LG OLED), this is what I will be shooting for.
 
Last edited:
If the AMD solution can some in at 3070 performance (which is unknown on both sides at this time, I speak purely hypothetically here), and is equal in price to the 3070, I'll go with the AMD solution hands-down.
 
I'm not sure AMD would come out on top in a price war.

Problem with Radeon as of late is that while great bang for the buck, its overall underwhelming for enthusiasts.
But then enthusiasts are only a small slice of the pie.

Time will tell if Radeon will continue to be the budget darling or if they step up in performance and price.
I think AMD could stand a pricewar. They kind of are already, as they probably expected Big Navi to be on the $600~$700 price bracket and are now supposedly go as low as $499. AFAIK BigNavi should be cheaper to produce, because of the tried and true 7nm process and simpler design.
 
According to Videocardz Big Navi will have a performance of 22.5 TFLOPs. RTX3080 has 29.8 TFLOPS while RTX3070 is 20.3.

Historically, AMD has better on paper flops performance but that doesn't translate in real world gaming performance, this doesn't bode well for Big Navi
 
According to Videocardz Big Navi will have a performance of 22.5 TFLOPs. RTX3080 has 29.8 TFLOPS while RTX3070 is 20.3.

Historically, AMD has better on paper flops performance but that doesn't translate in real world gaming performance, this doesn't bode well for Big Navi
I agree that this is a concern relative to AMDs GPU history. They do tend to lean more toward raw compute performance that's useful beyond gaming than Nvidia does when comparing both companies' consumer-focused parts, which does tend to result in theoretical compute performance that doesn't correspond directly to performance in games.

Big Navi (or, RDNA 2.0) could be a departure from that, which would be a pleasant surprise for gamers at least!
 
I agree that this is a concern relative to AMDs GPU history. They do tend to lean more toward raw compute performance that's useful beyond gaming than Nvidia does when comparing both companies' consumer-focused parts, which does tend to result in theoretical compute performance that doesn't correspond directly to performance in games.

Big Navi (or, RDNA 2.0) could be a departure from that, which would be a pleasant surprise for gamers at least!

To be fair, it seems Ampere seems to have much more compute performance than Turing. As evidenced by Brent's tests.
 
Newegg let something loose for a short time last night. 5120 stream processors on the top card with16GB of 16 Gbps GDDR6.

 
If the AMD solution can some in at 3070 performance (which is unknown on both sides at this time, I speak purely hypothetically here), and is equal in price to the 3070, I'll go with the AMD solution hands-down.
because..

I'd go nvidia because of 2nd gen RTX and DLSS. BTW I've read rumors that the 3070 will have 16gbps ram instead of 14gbps. 3rd party cards maybe?
 
I'd go nvidia because of 2nd gen RTX and DLSS.
DLSS and RTX voice, on top of better content-creation application compatibility. I wouldn't buy a GPU without some form of RT now regardless, but I honestly can't say that it'll make a difference. We're still looking at hardware likely not being fast enough for RT outside of a limited set of circumstances.
 
I haven't had faith in team AMD when it comes to performance over nvidia for sometime now -- this time though, even when paper numbers may differ, AMD will edge out or match 3080, land maybe between 3080 and 3090, or even trade a blow here and there with the 3090 -- For now nVidia is having a hard time getting the 3000 series to perform as great as they look on paper and Jensen says they are (though they are solid performers), that may change with Super's, then again maybe AMD will be able to respond with 6950 or 50 series when that time comes.
 
I haven't had faith in team AMD when it comes to performance over nvidia for sometime now -- this time though, even when paper numbers may differ, AMD will edge out or match 3080, land maybe between 3080 and 3090, or even trade a blow here and there with the 3090 -- For now nVidia is having a hard time getting the 3000 series to perform as great as they look on paper and Jensen says they are (though they are solid performers), that may change with Super's, then again maybe AMD will be able to respond with 6950 or 50 series when that time comes.
Not with the latest rumored specs. Lower TFLOPS performance and lower Memory bandwidth than the 3080 makes me think it may beat the 3070, but that's it.
 
Not with the latest rumored specs. Lower TFLOPS performance and lower Memory bandwidth than the 3080 makes me think it may beat the 3070, but that's it.

They took a panel of 11 games to test games on shading performance only, no DLSS, and no RTX. There wasn't a particular methodology to picking the titles - they just benched the games had lying around -- and found that the RTX 3090 was on avg 33% faster than the RTX 2080 Ti. This means, for the most part, the card is acting like a 23.5 TFLOPs GPU.

 
Become a Patron!
Back
Top