• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD Navi 23 ‘NVIDIA Killer’ GPU Rumored to Support Hardware Ray Tracing, Coming Next Year

Status
Not open for further replies.
ALXAndy said:
I doubt many have thought about it too much before commenting, but this could go two ways. Firstly, facts are facts Lisa Su clearly said "A 6000 SERIES GPU". Not "Our fastest flagship 6000 series gpu".
Click to expand...

On Twitter they confirmed it was big Navi though didn't elaborate more than that. Given the performance uplift even with refinements to the arch and process node it would be stretching credibility this is a 64 CU part so probably 72 or 80 unless the arch is very different to RDNA1.
 
ALXAndy said:
On 0 evidence apart from three benchmarks of games.

I doubt many have thought about it too much before commenting, but this could go two ways. Firstly, facts are facts Lisa Su clearly said "A 6000 SERIES GPU". Not "Our fastest flagship 6000 series gpu".

When pressed by GN they would not answer the question. Also, Scott Herkelman was asked to validate the same question and this was his response.

https://www.overclock3d.net/news/gp...st_rdna_2_big_navi_gpu_at_their_ryzen_event/1

Pretty much exactly the same thing Lisa replied with. In a comment to PC World, AMD's Scott Herkelman also stressed that AMD never said which Radeon RX 6000 series GPU they were running.

Now this could go two ways.

1. They are sandbagging and trolling Nvidia, trying to get them to pull their next move first (IE more VRAM, possibly more cards)
2. They are just trying to limit damage by showing that "hey, this card ain't the 3090 matcher people are leading you to believe".

However, both of those are still very possible because there is no clarity here *at all*.

One thing I do know is that if they are doing a number one (lol) then it would be a stroke of genius. Get every one a little disappointed, then come along on review day and friggin smash it, getting every one to pay lots of attention.

Now going on the past and being completely brand ignorant? I would not stick my neck out and risk a number 1. I would say it's more likely number twos. But, at least this time they have been a bit more honest about it instead of totally lying and then releasing a bag of crap like Fury and Vega.
Click to expand...

I like your reasoning but i would add that no one apart from wild speculating AMD fans are suggesting with certainty that AMD have a 3090 contender. Rather the more serious rumours were that big Navi was competing with the 3070. If they published those benchmarks to quash rumours i would guess that it was those ones.

That still doesn't answer the question of whether it was the top die or not, could still be either

But I'll just leave this here anyway
https://twitter.com/TDevilfish/status/1314556821282205697?s=19




Chooooo chooooo

.
 
Rroff said:
On Twitter they confirmed it was big Navi though didn't elaborate more than that. Given the performance uplift even with refinements to the arch and process node it would be stretching credibility this is a 64 CU part so probably 72 or 80 unless the arch is very different to RDNA1.
Click to expand...

We think there are two different big Navis. - Navi 22 and Navi 21.

DDH said:
I like your reasoning but i would add that no one apart from wild speculating AMD fans are suggesting with certainty that AMD have a 3090 contender. Rather the more serious rumours were that big Navi was competing with the 3070. If they published those benchmarks to quash rumours i would guess that it was those ones.

That still doesn't answer the question of whether it was the top die or not, could still be either

But I'll just leave this here anyway
https://twitter.com/TDevilfish/status/1314556821282205697?s=19




Chooooo chooooo

.
Click to expand...

Right:

 
Rroff said:
On Twitter they confirmed it was big Navi though didn't elaborate more than that. Given the performance uplift even with refinements to the arch and process node it would be stretching credibility this is a 64 CU part so probably 72 or 80 unless the arch is very different to RDNA1.
Click to expand...

They seem to be eager to match Ampere's TDPs though... as some leakers have suggested
 
ALXAndy said:
On 0 evidence apart from three benchmarks of games.

I doubt many have thought about it too much before commenting, but this could go two ways. Firstly, facts are facts Lisa Su clearly said "A 6000 SERIES GPU". Not "Our fastest flagship 6000 series gpu".

When pressed by GN they would not answer the question. Also, Scott Herkelman was asked to validate the same question and this was his response.

https://www.overclock3d.net/news/gp...st_rdna_2_big_navi_gpu_at_their_ryzen_event/1

Pretty much exactly the same thing Lisa replied with. In a comment to PC World, AMD's Scott Herkelman also stressed that AMD never said which Radeon RX 6000 series GPU they were running.

Now this could go two ways.

1. They are sandbagging and trolling Nvidia, trying to get them to pull their next move first (IE more VRAM, possibly more cards)
2. They are just trying to limit damage by showing that "hey, this card ain't the 3090 matcher people are leading you to believe".

However, both of those are still very possible because there is no clarity here *at all*.

One thing I do know is that if they are doing a number one (lol) then it would be a stroke of genius. Get every one a little disappointed, then come along on review day and friggin smash it, getting every one to pay lots of attention.

Now going on the past and being completely brand ignorant? I would not stick my neck out and risk a number 1. I would say it's more likely number twos. But, at least this time they have been a bit more honest about it instead of totally lying and then releasing a bag of crap like Fury and Vega.
Click to expand...

You've said an awful lot there.

All I asked was how can something be underwhelming if it matches a 3080. I didn't ask how good the presentation was. It wasn't even for the graphics cards.

I'm just going to wait and see what happens. I don't really care who is "fastest".
 
Nablaoperator said:
They seem to be eager to match Ampere's TDPs though... as some leakers have suggested
Click to expand...

I don't fully understand it but the firmware details seem to have configurations for 230 and 250 watt - not enough detail and/or understanding on my part there though it is possible that doesn't include PCI-e slot power, etc.

If you compare the Borderlands result against the 5700XT though even with some pretty decent optimisations you are looking minimum a ~4608 shader part to get that kind of uplift.
 
Rroff said:
I don't fully understand it but the firmware details seem to have configurations for 230 and 250 watt - not enough detail and/or understanding on my part there though it is possible that doesn't include PCI-e slot power, etc.
Click to expand...

There are 2 rumors from independent sources:
1 liquid cooled card in early 2021
2 a 340w TDP part

I am just connecting the dots.. and that TDP is total board power not the classical TDP definition.. don't know how firmware is interpreted
 
Parsley said:
You've said an awful lot there.

All I asked was how can something be underwhelming if it matches a 3080. I didn't ask how good the presentation was. It wasn't even for the graphics cards.

I'm just going to wait and see what happens. I don't really care who is "fastest".
Click to expand...

there is no evidence for it being like anything yet. What they showed was about as fast as a 3080 yeah, but there was no clarification at all as to exactly what it was.

Not sure why they did it tbh. We'll find out which reason it was soon enough.
 
ALXAndy said:
there is no evidence for it being like anything yet. What they showed was about as fast as a 3080 yeah, but there was no clarification at all as to exactly what it was.

Not sure why they did it tbh. We'll find out which reason it was soon enough.
Click to expand...

Maybe to stop the trolling that Big Navi will be as fast as the 2-year-old RTX 2080 Ti. Just saying.
 
4K8KW10 said:
Maybe to stop the trolling that Big Navi will be as fast as the 2-year-old RTX 2080 Ti. Just saying.
Click to expand...

There's trolling going on from every angle. Nvidia fans don't want it to be much faster than a 3070, and AMD fans want it to be as fast as a 3090. It's all trolling.

This is the point where it pays to have a clear head, and not be brand loyal. As it filters out an awful lot of BS from both sides.

I am just concentrating on the facts. Which TBH? AMD haven't really helped with here. Especially with 0 clarification whatsoever.
 
Phate said:
Yeah who knows. Maybe its the 20 pages of discussion it caused in 24hrs on this thread alone? :p
Click to expand...

I doubt that. If they wanted to give out information then they should have gotten specific.

But. I have seen this behaviour from AMD many times before, and I have always been hugely disappointed. I started whiffing the BS coming out of RTG when they said 4gb HBM was enough. I found out the hard way (by buying two) that it wasn't. The whole big b*****ks about "Vega being nice compared to the 1080Ti" too was just a joke.

If, however, that is the top end card then they just admitted defeat. Which TBH? something here isn't adding up. I mean sure, the CPU benchmarks looked very realistic and honest to me. Why would you be so forwardly honest about your GPU being cack? lol. That is the part that isn't adding up to me.

Whatever it is they are doing here has gotten Nvidia written all over it. That has to be about the loosest teaser I have ever seen in my life with absolutely no facts at all apart from three games. Which if they did cherry pick? would make it even worse. And I am sure they know that too. But. There was no Hitman, a game historically they have done very well against Nvidia at.
 
Have AMD said which process is used within Big Navi? As I know they confirmed the 5000series processor were all plain 7nm, I just wonder if AMD have made use of a better process for the GFX, rather than processor, so they can attempt to leap up and leap frog the oppositions.
If they manage to step up and trade with nvidia after being quite far behind, I'll be delighted.
 
ALXAndy said:
AMD haven't really helped with here. Especially with 0 clarification whatsoever.
Click to expand...

I disagree.
Compare it with the old flagship - the Radeon VII, for example:

Gears of War 5: 45 FPS https://www.gpucheck.com/compare-ga...-8700k-3-70ghz-vs-intel-core-i7-9700k-3-60ghz

RX 6000 series: 73 FPS or +62%

Borderlands 3: 27.9 FPS https://www.gpucheck.com/game-gpu/borderlands-3/amd-radeon-vii/amd-ryzen-7-3800x/ultra

RX 6000 series: 61 FPS or +119%

Call of Duty: Modern Warfare: 48.9 FPS https://www.gpucheck.com/game-gpu/call-of-duty-modern-warfare-2019/amd-radeon-vii/amd-ryzen-5-1600/

RX 6000 series: 88 FPS or +80%
 
The 7 wasn't a gaming GPU though. It was always going to be very easily beaten by something intentionally designed for gaming and not everything else.

I think you missed the context in what I meant too mate. What I mean is if that really is their best card they have just admitted defeat, 20 days before launch. IE - if that really is their fastest card? then any one who was hoping it would be faster and maybe wanted to buy it may now look elsewhere.

It's either misleading or it's just an honest confession IMO.

The rumour mill has been rife with "Much faster than the 3080, as fast as a 3090" and so on. So either they are having some fun or they are basically confessing that what the rumours say are false and it's actually slower. Which is what I don't get. Why do that now, 20 days before release, unless you wanted to push up pre orders of the 3080 even higher just to annoy Jen lmao.

BTW. I watched a bit of "Not an Apple fan but totally a AMD fan" yesterday live after the announcement and he looked absolutely gutted lol. I did enjoy that a bit, even if I don't dislike him. I think he's a genuine enough guy but to be totally convinced by rumour and speculation is something very dangerous. Especially when talking about AMD and GPUs.
 
Status
Not open for further replies.
Back
Top Bottom