Don't care if it has HBM or not, as long as yields are good and availability is good. HBM always seems to be a stumbling block in terms of availability, they've been producing it for 5 years so surely it's at a point now where it can be churned out en-mass with low failures.
-
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
- Thread starter Gregster
- Start date
- Status
Soldato
- Joined
- 8 Jun 2018
- Posts
- 2,827
I would be surprised if AMD didn't use an HBM variant of some kind.
Soldato
- Joined
- 21 Apr 2007
- Posts
- 2,594
HBM 2e is almost certainly going to be there on the higher end parts in fact if you don't see it in the specs I'd wager there is more to come
Soldato
- Joined
- 8 Jun 2018
- Posts
- 2,827
This rumor is sponsored by "pinch of salt". Make sure you use your pinch today!!!
Current rumors is that RDNA 2 won't beat a 3090/3080ti. But it will beat a 3080. No real confirmation on this as of yet. And, it's not clear which sku was used to present this. It's assumed it's a 6900 xt but no solid info on id of the card.
Current rumors is that RDNA 2 won't beat a 3090/3080ti. But it will beat a 3080. No real confirmation on this as of yet. And, it's not clear which sku was used to present this. It's assumed it's a 6900 xt but no solid info on id of the card.
Last edited:
This rumor is sponsored by "pinch of salt". Make sure you use your pinch today!!!
Current rumors is that RDNA 2 won't beat a 3090/3080ti. But it will beat a 3080. No real confirmation on this as of yet. And, it's not clear which sku was used to present this. It's assumed it's a 6900 xt but no solid info on id of the card.
So its still slower than a 2080ti using DLSS 2.0.That's a shame.
The Ray Tracing stuff is interesting though, we need more info on that - but if its true thats great news for the future of ray tracing in games. if the performance hit is just 5% to 10% then there is 0 reason that every game developer would not use ray tracing.
If I'm trying to find an explanation for the RT performance, I'm assuming that what AMD has done is reserving something like 1000 or 2000 shader cores just for Ray Tracing calculations and the rest for raster performance. This would explain why the raster performance is disappointing - the 6900XT appears to only be 50% faster than the 5700XT on average using those numbers you posted.
and where is the source? I can't see this rumour reported on videocardz or wccftech.
Last edited:
Seriously?
It's not faster at *native resolution* than the fastest card on the market, running a *lower* resolution?
That's disappointing?
I just need it to be priced well.
*Edit*
Appears to be a fake:
https://www.pcgamer.com/amd-big-navi-rdna-2-leak-fakery/
Last edited:
He has been building up to use that line for ages
Soldato
- Joined
- 8 Jun 2018
- Posts
- 2,827
Yup, interesting how they respond isn't it? They don't disappoint.
What is being said is that those aren't the results of this big navi. Problem is some reports out there are confusing big navi vs nvidia killer. Which are 2 different products.
Last edited:
Nvidia have been charging whatever some people are prepared to pay ever since AMD stopped competing at the high end. This has led to a significant part of the market blindly 'accepting' that the performance improvements that used to come with each new generation of card (typically 25-40%) free at the same price point must now be accompanied by a 25-40% increase in price. This price creep is facilitated by your defeatist logic that says 'I accept this' when in fact a big portion of the market doesn't accept it and simply slowed their upgrade cycle/refused to buy. AMD will not regain significant market share without offering either significant price reductions for equivalent performance or significantly better performance at a given price. Look at the CPU market if you don't understand the playbook and what works. Offer 90-95% of the performance at 60% or thereabouts of the price not 95% of the performance at 89% of the price. Or offer 120% of the performance at 95% of the price. It's how you regain mindshare and mass market acceptance/adoption rather than clinging on to less than 20% market share. If it's just a 5-10% difference the brute mass will always buy Nvidia, it's what they're programmed to do and AMD have to reprogram them.
Yeah. This makes a hell of a lot more sense than “AMD are not a charity” approach which will lead them accepting whatever price nvidia sets (which will keep going up each gen) and then either charging 5-10% less or maybe 5-10% more (they are now a premium brand, didn’t ya know). Worst part is they always seem to bring said performance months if not a years later at the top end leading to nvidia always having the best performance helping them keep/further extend mindshare.
If I was them I would keep profits to a minimum for one or two gens at least to gain some market and mindshare. They can still sell for much bigger profits to data centres, professionals etc. Also if they are able to beat nvidia by bringing a card out that is the best, they should go ahead and charge a silly price for it to make it look ultra premium and milk the I want the best crowd just like Nvidia do with the Titan.
How did you work that out? There are no numbers of substance on those slides. The first slide shows improvements 6900XT over 2080 Ti, the 2nd slide shows RT impact to itself. So how did you determine "slower than 2080 Ti with DLSS"? But then, reducing image quality has always gained FPS, so if it takes turning on upscaling crap to stay ahead of the competition running at native then enjoy shifting your goalposts.
That article is from June and refers to the spec leaks, not those bar charts. That's not to say the bar charts aren't fake too of course
Well, those charts are clearly labelled 6900XT
As for rumours, there's a rumoured belief that AMD are sandbagging a little in it's them saying RDNA 2 will only beat the 3080 to control leaks. Since Lisa Su fired the previous RTG marketing team, AMD have become pretty tight sealing up the leaks. Information is apparently more controlled now, and "only beating 3080" is part of that information control.
The rumoured 50-60% faster than 2080 Ti has apparently come from inside Nvidia. Take that one for what it's worth.
I've just finished a massive 10Kg tub of sourdough pretzels so I literally have a huge bucket of salt I can share with you all
Associate
- Joined
- 24 Jun 2016
- Posts
- 852
- Location
- Hartlepool
I've just finished a massive 10Kg tub of sourdough pretzels so I literally have a huge bucket of salt I can share with you all
Was that in one sitting?
Was that in one sitting?
That's one way to get so salty
The article does discuss the two performance slidesThat article is from June and refers to the spec leaks, not those bar charts. That's not to say the bar charts aren't fake too of course
Not sure why ECH is posting such old speculation....
It might as well have been! The other half randomly bought me this big-ass tub and, not being that much of a pretzel fan, though they'd last me forever. Dip in once in a while, have a nibble.Was that in one sitting?
They've gone in a touch over a month.
It seems to have come out of the blue though.
But after Fury X , Vega 64 and Radeon VII
AMD do seem to love high-end HBM gaming cards so there may be as you say a "Halo" card produced.
This.
AMD need to regularly feature high up in the benchmarks. It's excellent marketing. I know people who've looked at the them and bought a lower spec Nvidia card because of the performance of the X080Ti cards.
Unfortunately it appears Nvidia will be filling slots in the benchmarks with 3080Ti, 3080 and possibly 3070 cards before AMD release Big Navi cards. AMD performance will look extremely weak for a (short) period of time.
Last edited:
- Status