confirmed: AMD's big Navi launch to disrupt 4K gaming

So, how much as the 8800 Ultra?

Sure, pick the largest outlier of the bunch:

1c9a8251-8039-4dc6-9e84-40f92178c220.png


Fact is, even with inflation a $500 card in 2005 should only be about $650 today. The actual prices for the top end are not even close.
 
If they released something at or above 2080TI levels for anything less than a 2080TI their shareholder will take them to pound town. nVidia got theirs to market first and got to set the price they can disrupt it if they want but they sure as shit better have the yields to supply that demand otherwise they are just shorting themselves for temporary and marginal market share gains. AMD is not large enough, rich enough, and does not command the manufacturing capabilities to cause a disruption. At best they might cause a wake in the waters.
 
If they released something at or above 2080TI levels for anything less than a 2080TI their shareholder will take them to pound town. nVidia got theirs to market first and got to set the price they can disrupt it if they want but they sure as shit better have the yields to supply that demand otherwise they are just shorting themselves for temporary and marginal market share gains. AMD is not large enough, rich enough, and does not command the manufacturing capabilities to cause a disruption. At best they might cause a wake in the waters.

If they release a card at or near 2080ti performance at 2080ti prices then there is no discussion. Threads like this are meaningless. If your assertion is true then we should not root for competition, in fact we should pray for AMD and NVIDIA to join forces and treat enthusiasts anyway they want as long as it means they can drop silicon and heatsinks every year for us to throw money at.
 
If they release a card at or near 2080ti performance at 2080ti prices then there is no discussion. Threads like this are meaningless. If your assertion is true then we should not root for competition, in fact we should pray for AMD and NVIDIA to join forces and treat enthusiasts anyway they want as long as it means they can drop silicon and heatsinks every year for us to throw money at.
We do t have competition we have 2 players, that work in tandem for pricing. Either party could drop their prices by 50% and still make a profit but they sure as hell aren’t. Once Intel steps into the ring assuming they actually release a consumer card then we may see something interesting. But until then AMD has to play in a way that satisfies all those shareholders they took a lot of loans to keep afloat and they have a lot of payback to make. If they can only make 100,000 of something then they have to sell each of those for the absolute max they can. AMD is doing great over these last few years but they have a decade of bad designs and decisions to pay for.
 
  • Like
Reactions: Auer
like this
hm

AMD does not seem to be exaggerating lately when it comes to performance, they've pretty much over delivered on every promise for ryzen, ryzen+, and ryzen 2.

RTG has a very bad history of overpromising, but Navi seemed to be on the money and I don't remember any crazy fud about it so hopefully there's some truth in this.

But as I've been constantly reminded by the AMD fanbois (over and over and over and over and over again), AMD CPU hates hates hates the GPU team which hates hates hates the CPU team and both have nothing to do with each other and don't even sleep in the same room, much less the same bed as each other... or some such.
 
AMD won't price it at $599, assuming rumors are true and it's double 5700 XT or close to 2080 Ti performance.

But I also don't see a $1,200 GPU doing well, as everyone that was willing to spend that money already bought a 2080 Ti.

If they didn't want to pay $1,200 then, I don't see that price-point being viable, especially a year later.

Steam says only 0.5% of players have a 2080 Ti, half of 1 percent. AMD will not make any money here.

Investors want a return, but a return does not mean selling a card at a price that prices them out of the market.

Now at $699, or even $799, maybe that would be interesting and actually disruptive.
 
We have Alyx coming out real soon now. If its good enough to move VR sets then people are gonna want halo cards too.

If Valve have cracked that nut then disruption will happen like we haven't seen since Quake and hardware T&L.

Disruption in the video editing space isn't coming from AMD, its coming from Black Magic.
 
  • Like
Reactions: Wat
like this
AMD won't price it at $599, assuming rumors are true and it's double 5700 XT or close to 2080 Ti performance.

But I also don't see a $1,200 GPU doing well, as everyone that was willing to spend that money already bought a 2080 Ti.

If they didn't want to pay $1,200 then, I don't see that price-point being viable, especially a year later..
Exactly, nobody would buy a card that is close to the 2080ti for 2080ti prices, except for rabid fans. Who had that kind of money already bought a 2080ti.
There is no point in spending 1200 on 2080ti performance, when you could've done that already long ago.

$1200 price would only be viable if it came out not long after the 2080ti.
 
Sure, pick the largest outlier of the bunch:

View attachment 218859

Fact is, even with inflation a $500 card in 2005 should only be about $650 today. The actual prices for the top end are not even close.

The “top end gpu” isn’t a comparable good over time like an apple or a loaf of bread. The definition of top end in terms of engineering complexity and raw performance has changed dramatically since 2005.

Everybody likes cheap stuff but using inflation as an argument for cheaper GPUs is kinda pointless.

We aren’t entitled to $650 RTX Titans just because a 8800 GTX was $500.
 
Literally for almost ten years (2008-2017) the high end of the consumer cpu market was quad cores with SMT. Each generation slightly faster than the last, which is why so many held on to those Sandy Bridge chips for so long. There was no significant performance to be gained from upgrading even 5 generations later.

First gen Ryzen doubled that. Two and a half years later 3rd gen did it again, now to 16 with SMT. Intel’s high end finally increased the core counts and they refined 14nm to such a sharp edge the latest ones are an actual improvement on the first ones. Consumers can buy high end laptops with 8 cores and 64gb of RAM.

it is the very definition of disruptive.

as for whether they can do it in the gpu space - I’d love for all these rumors to be true, but aside from complaining about pricing, I don’t think anyone could really argue that a Fermi-based GPU is still “good enough”. Nvidia has been pushing performance improvements each generation, noticeable ones at that, and AMD has been keeping up in the mid-range at least (though a year or two behind).

at this point all they’d need to do to disrupt the market is release a competitive part high end nvidia hardware within a few months of nvidia release, instead of two years later.

Spot on

Disruption means as little as raytracing at this point.

Just a buzzword.

Go AMD and Nvidia I need a new video card (mine blew last night) and I’m waiting for this release before I pull a trigger on either team.

ftfy ;)
 
Last edited:
While it wasn't the top end card, I got a 8800GTS 512MB for $239 when it was new.

Granted, it was mid-range at the time, but it played very well. I could rock Crysis maxed at 720P (with a few graphic mods) and it was beautiful.

That was what I would call a reasonable price for reasonable performance.
 
There is no such thing as "reasonable" and "unreasonable" prices. It's all about what the market will or will not support.
You can call it whatever you want it is the same thing. Prices that are too high for the market conditions are unreasonable. And bringing out 2080ti performance at 2080ti prices 1 year late is unreasonable.
 
You can call it whatever you want it is the same thing. Prices that are too high for the market conditions are unreasonable. And bringing out 2080ti performance at 2080ti prices 1 year late is unreasonable.

We normally expect pc GPUs to be more powerful than the GPU priced inside a console whose total price matches the GPU's price

The xbox x is expected to be 56/60 CU RDNA 2 Gpu and it is expected to cost $600 or so while performing at the level of RTX 2080 super

So we can expect an AMD 5800 XT costing $550 with performance of RTX 2080 super, I am guessing

Of course there would be a higher tier 5900 XT, hopefully 15%-30% better than RTX2080 ti & probably costs less than $800
 
The “top end gpu” isn’t a comparable good over time like an apple or a loaf of bread. The definition of top end in terms of engineering complexity and raw performance has changed dramatically since 2005.

Everybody likes cheap stuff but using inflation as an argument for cheaper GPUs is kinda pointless.

We aren’t entitled to $650 RTX Titans just because a 8800 GTX was $500.

I was going to post this but you beat me to it. People using feelings rather than reality. Pricing is based off of costs / target margins and if the market won’t support that it’s not made at all.
 
But given that just 1.49 percent of Steam users own a GTX 1080 Ti — and that GPU is $500 less than the RTX 2080 Ti — it's clear
that money is an object for the vast majority of gamers."

Steam says only 0.5% of players have a 2080 Ti, half of 1 percent. AMD will not make any money here.

If we went by Steam's numbers, then there is no point in AMD continuing to stay in business. Their #1 most popular card on Steam comes in at a laughable 1.29%. Similarly, CPU core counts above 4 must be the dumbest thing ever, right? 4 core CPUs represent 55% of users on steam. 10 core CPUs? 0.06%. Nothing says fanboi like using Steam numbers to declare Nvidia stupid while completely missing the fact that AMD's numbers are a rounding error in comparison.

Maybe you guys have a point though. All displays above 1080p should be canceled and thrown directly into the trash because the sum of all resolutions above 1080p on Steam is only 10%.


Investors want a return, but a return does not mean selling a card at a price that prices them out of the market.

Investors do want a return, that much is correct. How they get that return is pretty variable. If AMD said they're taking a 6mos sabbatical from GPU sales, Nvidia shares would spike and investors would get a massive return - all without Nvidia having to release anything at all. If crypto suddenly made a comeback, more positive share movement. There are a ton of things that move a stock price and which the company can control. Unit sales is only one of them.

Don't forget that the 2080 Ti is also a halo product. Those aren't intended to move truckloads of units. Maybe the Corvette should be killed off because the $123k ZR1 model is too expensive compared to the $13k Spark? This is how halo products work. Simply having them available creates warm and fuzzies in fans which in turn gets them to buy the product they can afford, even if that isn't the halo product. Returning to your steam numbers, AMD's top selling card surely must be the $200 RX580. So why on earth did they launch a new line of GPUs where 5 of the 7 models are priced above $200? Why even have that 50th Anniversary Edition? Because halo models work.
 
It seems many people are confused about how economics works. Every product doesn’t have to be affordable for everyone. The market for a 2080 Ti is not the same as the market for a 1660.

When people say GPUs are overpriced I laugh because there’s literally a card for every budget nowadays. They’re like $50 apart.

If your budget is $500 then buy a $500 card. Don’t cry just because you want the $1500 card and can’t or refuse to afford it. If the market didn’t exist at that price then the card wouldn’t exist at that price.
 
Last edited:

I'll believe it when I see it. AMD has over hyped under-performing video cards one too many times to be trusted until the product is in the hands of unbiased reviewers.

Navi has been spot on. I get it in those Raja days, but its been fairly well since he has been gone. I don't remember Navi being overhyped. It was exactly as they described. I don't see them repeating and over hyping things like they did before when Raja was there.
 
I'll believe it when it shows up, but it won't be disruptive. To be disruptive it would need to sneakily show up after the 3080Ti launch at 30-40% off the 3080Ti MSRP and match very closely to the 3080Ti performance. If AMD did that I would then say absolutely they will disrupt the GPU market. But if it matches the 2080Ti at a lower price and Nvidia can respond to it with a 3070 at the same/lower price then nothing is going to change much from how it is now.
 
apparently that does seem to be the case:

" Is the 2080 TI worth the price?
But it does so with a mammoth cost increase that is completely unjustifiable. If money is no object, the 2080 Ti is a great GPU.
But given that just 1.49 percent of Steam users own a GTX 1080 Ti — and that GPU is $500 less than the RTX 2080 Ti — it's clear
that money is an object for the vast majority of gamers."

https://www.extremetech.com/gaming/...d-rtx-2080-ti-review-you-cant-polish-a-turing

LOL. Stupid article, must be AMD fanboi's.

Steam has 90 million users. Using December 2019's Steam Hardware Survey numbers:
0.50% 2080Ti = 450,000 cards. That's over a half bilion dollars in sales btw.
1.34% 1080Ti = 1,206,000 cards.
The much beloved (by the aforementioned fanbois) AMD 5700Xt is at 0.22% = 198,000 cards. Sales of about 79 million using $400 price.
Top card with a crazy 20.91%! GTX1060 = 18,819,000 cards!

...Steam says only 0.5% of players have a 2080 Ti, half of 1 percent. AMD will not make any money here....

See my post above. nVidia did pretty well, so anything is possible :)

If we went by Steam's numbers, then there is no point in AMD continuing to stay in business. Their #1 most popular card on Steam comes in at a laughable 1.29%. Similarly, CPU core counts above 4 must be the dumbest thing ever, right? 4 core CPUs represent 55% of users on steam. 10 core CPUs? 0.06%. Nothing says fanboi like using Steam numbers to declare Nvidia stupid while completely missing the fact that AMD's numbers are a rounding error in comparison...

Lol, yup.

Someone go add up all of the AMD gpus' listed and come up with a total %, use the December column: https://store.steampowered.com/hwsurvey/videocard/
 
LOL. Stupid article, must be AMD fanboi's.

Steam has 90 million users. Using December 2019's Steam Hardware Survey numbers:
0.50% 2080Ti = 450,000 cards. That's over a half bilion dollars in sales btw.
1.34% 1080Ti = 1,206,000 cards.
The much beloved (by the aforementioned fanbois) AMD 5700Xt is at 0.22% = 198,000 cards. Sales of about 79 million using $400 price.
Top card with a crazy 20.91%! GTX1060 = 18,819,000 cards!



See my post above. nVidia did pretty well, so anything is possible :)



Lol, yup.

Someone go add up all of the AMD gpus' listed and come up with a total %, use the December column: https://store.steampowered.com/hwsurvey/videocard/

Fanboi math
 
I'm all for next gen cheaper prices and competition, but confirmed is a strong word. I keep reading "rumors" in these articles. I have a 1080ti I'm holding out upgrading until Cyberpunk release, it's either going to be a 3080 or potentially an AMD if this holds true about their new performance at the high end. Either way, I hope it's close just for pricepoint sake.
 
Would love a solid 4k gaming option under $400 CAD.

Probably won't happen, but a man can dream.
 
Which games are straining a 1080 Ti or "old" cards?
Not many per se, but if you include Ultra settings, there are quite a few AAA titles that start straining the 1080Ti. Of course all are more than playable at lesser settings, even high, but who wants to be held back by that? :D And nvidia surround personally for me is a big reason to upgrade to the next gen after the 2080's.
 
We do have to give Lisa Su and team credit for digging AMD out of the complete hole they were in with the FX/Bulldozer designs and turning things around to the point where they are in the game again, but still, without Intel stumbling, Ryzen would have been an OK chip, but 10-20% behind Intel.

In short, AMD's resurgent position is largely due to luck that Intel messed up. In a way, the last time they were successful in 1999 to 2005 the same thing happened. Intel messed up with P4/Netburst/RAMBUS allowing AMD an opening.

What a strange take. You're willing to downplay AMD's recent success, which is due to shipping an actual product, because Intel would have shipped a better competing product if they hadn't "stumbled." So now AMD's new products are competing against projected performance of theoretical Intel products.

So now it's "luck" that benefits AMD - Intel's bad luck at that - and the only reason AMD was competitive back in the day was because Intel "allowed them an opening." I guess AMD was quickly put back in its place once Intel pulled themselves up by their bootstraps and made superior products once more - nothing to do with crippling unfair competition practices.
 
The surveyed users obviously?
You should be more specific. If you are accusing me of 'fanboi math', you may need to look in a mirror...
Ok, more specifically, taking the survey is a manual process or a pop-up prompt when selected. To assume 100% of the user base has done the survey is presumptuous to say the least. Multiple boxes, multiple accounts, never taken a steam hardware survey. And no, it's not behind the scenes and automatic. Your assumptions are flawed and therefore your math is way off base as well.
 
It's not 100% of the user base, never said it was. It is enough to be statistically accurate within the standard deviation. You can't "lie" on your hardware survey, that is automatically pulled if you opt in to the survey. Believe what you want. I'm sure AMD is king of the flat Earth in some reality...
 
What a strange take. You're willing to downplay AMD's recent success, which is due to shipping an actual product, because Intel would have shipped a better competing product if they hadn't "stumbled." So now AMD's new products are competing against projected performance of theoretical Intel products.

So now it's "luck" that benefits AMD - Intel's bad luck at that - and the only reason AMD was competitive back in the day was because Intel "allowed them an opening." I guess AMD was quickly put back in its place once Intel pulled themselves up by their bootstraps and made superior products once more - nothing to do with crippling unfair competition practices.

It had lots to do with Intel's illegal business practices. It starved AMD of the R&D resources necessary to keep up a competitive footing.

Intel's mistakes with P4/Netburst/RAMBUS are what gave AMD the opening, and because of work they had done internally, their acquisition of NexGen, and their licensing of technology from and hiring of many engineers from Compaq after they shuttered DEC, AMD was ready to take advantage of that opening.

What they were unable to do was get over Intel's entrenched position in the market, especially with their illegal and unethical business practices in locking them out, sabotaging their performance in their compiler, etc

I just hope history doesn't repeat itself.

This time around, Intel's fuckup with 10nm has given AMD an opening yet again. An opening they were able to take advantage of because of their skillful financing moves, singing large contracts with console manufacturers in order to sustain the bottom line, stay alive long enough and pay for the likes of Jim Keller and other engineers to design a new architecture.

The fact - however - remains, that without Intel's massive fuckup, they would have been on 10nm++, ready to transition to 7nm by now and would have had much better chips in the marketplace, and we would not have had the close race we have today which is greatly benefitting the consumer.
 
I It is enough to be statistically accurate within the standard deviation. ..
Really? More assumptions based on your insight into steam, how the survey is performed and your knowledge of statistical analysis? :ROFLMAO:
 
Last edited:
4K is the only resolution that really needs power today or for the next number of years.

I picked up a 5700xt and am playing on a ultrawide so a few more pixels to push over 1080p... and very little isn't running over 100fps at ultra settings. AC odyssey averages 80 somethingish. Perhaps cyberpunk pushes our cards a bit more the end of this year... but really it sounds like no developers are going to be targeting the PS5 or XboxNextOneNext only. So games are not likely going to really push things much further then where they are now for another 4 or 5 years at least.

Games have stagnated... the only real disruption that is possible at this point. Is if more people really buy into 4k. At this point where we need the disruption is in the Monitor space. lol Facts are unavoidable according to Valve there are more people playing at 1024x768 and 1280x1024 is about = to the number of people playing at 4k. Almost 3 quarters of gamers are playing at 1080p... and the ultrawide resolutions are almost in the double digits (9% for the ultra wide resolutions vs 2% for 4k).

So at the top the only possible disruption that is possible requires a lot of people to buy 4k monitors with high enough refresh rates for a new card to even be of benefit.

I'm sure at some point 4k will become more popular... and things like Ray tracing may make new cards worthy investments. Neither seems to be happening fast enough for anything to even have a chance to be disruptive. Games people where waiting on to show off RT are dropping it... or scaling it back. And the AAA developers have all basically said they are going to be targeting last generation consoles for a long time yet, I wouldn't hold my breath that the AAA houses are going to even bother with higher resolution texture packs ect.

On the rumors... I have no doubt AMD is going to release a card 15-20% or so faster then a 2080ti... the 5700 xt is already an equal for the 2070 super, even if they just bump the shader cores up they will get there. Hopefully there RT engine is fantastic... but I would have to believe if the developers knew something fantastic was coming there they would be working harder on RTX stuff. On the NV front... 50% faster and 50% less power that is a pipe dream that we all know form history is impossible. NV can choose 50% faster or 50% less power but not both die shrink or not. I would expect 20% better on both fronts. Or perhaps 50% more power at the same power draw as the 2000s... die shrinks aren't magic.
 
Back
Top