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

If Big Navi is only 15% ahead of 2080Ti & costs $499, I'm in at launch (if linux drivers are available)

*Spiderman Meme* Oh we're using our made up numbers.

Expectations - reality = happiness. It seems like people LOVE to get unrealistically high expectations with AMD. I'd expect 15% over a 2080ti to be around $899.
 
*Spiderman Meme* Oh we're using our made up numbers.

Expectations - reality = happiness. It seems like people LOVE to get unrealistically high expectations with AMD. I'd expect 15% over a 2080ti to be around $899.

Well someone mentioned disruptive, so... :)
 
Well someone mentioned disruptive, so... :)
Oh, the bugs in their drivers are disruptive alright! All the way back from my hd4850 thru 5870 and 290 days when I outright gave up on them. I had nice ones in ancient times like the 9700pro though.
 
Oh, the bugs in their drivers are disruptive alright! All the way back from my hd4850 thru 5870 and 290 days when I outright gave up on them. I had nice ones in ancient times like the 9700pro though.
I means space invaders was kind of disruptive too ;). But, I digress, their drivers do need work.
Odd, I've had almost zero issues with my AMD cards. I still have a 280x, fury x, fury nano, RX 560, RX 570... all running with absolutely no issues. My laptop had an AMD HD onboard that has stuttering issues if I install the latest AMD software... The drivers are fine if I don't install all the add-on crap. But it's a laptop from like 2009 and I think it's probably some setting trying to do something in th background. I wasn't going to spend a lot of time on it since it's not like I need any of the streaming tools, lol. Wish they would finally squash the rest of the bugs people are running into that have been outstanding for way to long though.
 
Oh, the bugs in their drivers are disruptive alright! All the way back from my hd4850 thru 5870 and 290 days when I outright gave up on them. I had nice ones in ancient times like the 9700pro though.

The driver thing is a little over stated for some of the older cards. I had a 4800 series, and an hd 6950, they were both excellent, same with my X1800xt. But yeah, before they were bought by AMD I think they made better products. They're turning around, but it's taken some time. The RX 5700 I have has been flawless since the 2020 release.
 
The driver thing is a little over stated for some of the older cards. I had a 4800 series, and an hd 6950, they were both excellent, same with my X1800xt. But yeah, before they were bought by AMD I think they made better products. They're turning around, but it's taken some time. The RX 5700 I have has been flawless since the 2020 release.
It's less their eventual driver quality, and more how long it takes them to get there after a release -- it's hard to name an AMD GPU release that wasn't a driver shitshow.
 
I bought the Radeon VII at launch, the drivers were fine and everything was good.
Well, if your copy had a good cooling solution and the BIOS wasn't borked by the vendor (or otherwise), then the drivers should have been alright because it was just a shrink of Vega 64 ;)
 
the market will continue to pay the same for an AMD GPU in terms of dollars-per-performance to a comparable Nvidia GPU.

People dreaming of a mythical $499 2080Ti-killer are just not understanding that the world has changed

After doing video card reviews for a few years and gaming for so many more, I can say confidently you can almost always judge a card's overall performance by its real-world price, assuming its features are in-line with the mainstream. This isn't new, it's a return to normal.

Unless it does something specifically different--being the very best counts as something different, but so could a number of other features like special cooling or a non-standard capability or compatibility--and it's not brand-new, the in-store/online pricing will reflect the general performance. The mining era was a freakout that messed with that rule of thumb, but it takes something aberrant in the market to sway prices away from the performance.
 
Why bash AMD peeps, they never were in the halo product line.
For example when I bought my 390X it was within Ti performance for 300 less, value over performance buy.
Nvida issues are huge dies, poor yields. Until one breaks NDA and states exactly what is cost to make and package a 2080 Ti, we will see what the gouge or price margin really is. It is not as high as people think. What is a retailers invoice for a 2080 Ti?
 
Me too long term AMD GPU user, but I believe most people today assume AMD has to compete at the halo end, as long as they stay under 600 and perform well I will buy them. I did not see a 980 Ti at $700 being that much faster for the price diff at 1080p. I settled on a XFX $350 Newegg sale for a 390X and it still runs everything I play at 2K 144 Hz. today.
For a reality check I read a XFX THICC OC 5700XT review that got 45 FPS on Ultra for RD2, my five year old card on same setting 40 FPS. I am not spending $400 to gain 5 FPS @ 2K.
 
Me too long term AMD GPU user, but I believe most people today assume AMD has to compete at the halo end, as long as they stay under 600 and perform well I will buy them. I did not see a 980 Ti at $700 being that much faster for the price diff at 1080p. I settled on a XFX $350 Newegg sale for a 390X and it still runs everything I play at 2K 144 Hz. today.
For a reality check I read a XFX THICC OC 5700XT review that got 45 FPS on Ultra for RD2, my five year old card on same setting 40 FPS. I am not spending $400 to gain 5 FPS @ 2K.

That makes no sense, i get around 70 FPS with a RX 5700, i think there's like 2 settings i turn down to high.
 
It's mostly a brand/image thing.

While AMD hasn't had the top end in a long time, for the low and mid range I would say they are highly competitive in terms of price and performance.

But people still buy Nvidia even while AMD has competent mainstream offerings, e.g. in the sub $400 range.

They need to do something on the level of Ryzen, and even then it took several years and 2 generations to take the lead spot (not to mention Intel's numerous stumbles).
 
It's mostly a brand/image thing.

While AMD hasn't had the top end in a long time, for the low and mid range I would say they are highly competitive in terms of price and performance.

But people still buy Nvidia even while AMD has competent mainstream offerings, e.g. in the sub $400 range.

They need to do something on the level of Ryzen, and even then it took several years and 2 generations to take the lead spot (not to mention Intel's numerous stumbles).


The difference is Nvidia hasn't been complacent for the last decade and been cool with marginal improvements generation after generation. RDNA kind of reminds me of the Ryzen 1000 series chips, but I don't see Nvidia staying complacent long enough for it to matter even when the Zen 2 of GPUs eventually arrive.
 
But people still buy Nvidia even while AMD has competent mainstream offerings, e.g. in the sub $400 range.
And Nvidia can still be the better buy for a variety of reasons. Could just be game support, which they invest heavily in. Could be drivers. Could be the transcoding blocks for streamers. Could be availability!
 
And Nvidia can still be the better buy for a variety of reasons. Could just be game support, which they invest heavily in. Could be drivers. Could be the transcoding blocks for streamers. Could be availability!

Graphics card reviews these days are pretty shit to be honest. They all just focus on graphs of FPS and rarely delve into the actual features and usability of the product. A lot of those things have a huge impact on the user experience and sway purchase decisions more than a few % FPS difference.

In my case I’m a very heavy user of in-home streaming from my desktop to multiple tvs in the house over Ethernet. AMD doesn’t have a competitive alternative so I’m pretty much stuck in Nvidia’s ecosystem.
 
I want to attempt to summarize the argument against AMD pricing lower, open for criticism:

Basically, if AMD prices lower enough, Nvidia will simply lower their own prices to keep AMD in check. From a business standpoint, Nvidia can afford to do this more than AMD can, and is more likely to be willing to do it. So AMD is best off pricing just under Nvidia on the outset and getting better margins.

You mean AMD isn't the tech messiah that will save us from the evil that is Nvidia?

That being said there were a couple times where AMD severely undercut nvidia pricing for the performance offered.

Hd4870/4890
" For now, the Radeon HD 4870 and 4850 are both solid values and cards we would absolutely recommend to readers looking for hardware at the $200 and $300 price points. The fact of the matter is that by NVIDIA's standards, the 4870 should be priced at $400 and the 4850 should be around $250. You can either look at it as AMD giving you a bargain or NVIDIA charging too much, either way it's healthy competition in the graphics industry once again (after far too long of a hiatus). " Anandtech

HD5870
"Let’s be clear here: the 5870 is the single fastest single-GPU card we have tested, by a wide margin. Looking at its performance in today’s games, as a $379 card it makes the GTX 285 at its current prices ($300+) completely irrelevant. The price difference isn’t enough to make up for the performance difference, and NVIDIA also has to contend with the 5850, which should perform near the GTX 285 but at a price of $259. As is often the case with a new generation of cards, we’re going to see a shakeup here in the market as NVIDIA in particular needs to adjust to these new cards. " Anandtech

r9 290
" To get the positive aspects covered first, with the Radeon R9 290 AMD has completely blown the roof off of the high-end video card market. The 290 is so fast and so cheap that on a pure price/performance basis you won’t find anything quite like it. At $400 AMD is delivering 106% of the $500 GeForce GTX 780’s performance, or 97% of the $550 Radeon R9 290X’s performance. The high-end market has never been for value seekers – the fastest cards have always commanded high premiums – but the 290 completely blows that model apart. On a pure price/performance basis the GTX 780 and even the 290X are rendered completely redundant by the 290, which delivers similar-to-better performance for $100 less if not more " Anandtech

Saying AMD hasn't undercut nvidia on pricing isn't accurate, they have. Not certain they will this generation though with how closely price navi is to counterparts.
 
Last edited:
Regardless of all the banter and tragedy these days it's great to be excited about the future of PC gaming still. I'm most interested in performance per watt this round as we all understand at some level less heat is what we really need.
 
Well, the 290X was the fastest when it launched in October 2013 with a terrible cooler but amazing (for its time) architecture, nVidia released the 780Ti which was only slightly faster than the 290X but architecturally inferior. So much so that in a year's time, the 290X was 25% faster than the 780Ti and the gap continued to widen in the coming years. I still believe it was their best card after the 9700 Pro/1950XTX/HD5870, but could've done a lot better had it been released with a good cooler and at a better voltage/frequency curve.

October 2013 still remains the last day an AMD card was faster than nvidia at launch. Let's hope there's renewed competition again this fall.
 
Graphics card reviews these days are pretty shit to be honest. They all just focus on graphs of FPS and rarely delve into the actual features and usability of the product. A lot of those things have a huge impact on the user experience and sway purchase decisions more than a few % FPS difference.

In my case I’m a very heavy user of in-home streaming from my desktop to multiple tvs in the house over Ethernet. AMD doesn’t have a competitive alternative so I’m pretty much stuck in Nvidia’s ecosystem.
Yeah, I'm really hoping AMD does something with their less than stellar IQ for encoding. It's embarassing to the point I'm better off with an Intel CPU with integrated graphics. But, then I'm stuck with an Intel CPU, so no go on that (I'm not paying Intel prices for cores for my home server, it's not a gaming machine). Only other option is to get an nvidia card. But they lock down the # of transcodes you can do (which is easily bypassed, but I don't want to support that). AMD doesn't lock anything down, but the quality just isn't there. Hopefully that money finally rolling in from Zen can get a few more people hired to work on these little issues. But as you mentioned, it's like nobody even tests these things or checks them out (besides one or two niche people), so it doesn't get a lot of attention.
 
Well, the 290X was the fastest when it launched in October 2013 with a terrible cooler but amazing (for its time) architecture, nVidia released the 780Ti which was only slightly faster than the 290X but architecturally inferior. So much so that in a year's time, the 290X was 25% faster than the 780Ti and the gap continued to widen in the coming years. I still believe it was their best card after the 9700 Pro/1950XTX/HD5870, but could've done a lot better had it been released with a good cooler and at a better voltage/frequency curve.

October 2013 still remains the last day an AMD card was faster than nvidia at launch. Let's hope there's renewed competition again this fall.
Yes, and if AMD could have figured out the drivers sooner, that slightly faster --> 25% would have been like 20% at launch and 25% after a year which would have lead to much better sales. When it's close, people will stick with a brand they know (in general). When it's a blowout, people will switch.
 
All AMD would be doing in a hypothetical of lowering price is giving away margin to retailers and gougers - the market will continue to pay the same for an AMD GPU in terms of dollars-per-performance to a comparable Nvidia GPU.

People dreaming of a mythical $499 2080Ti-killer are just not understanding that the world has changed, and it's probably going to require more years before the benefit of enough hindsight for it to click and be obvious.

Why is it in these speculation threads, invariably for the past 20 years (quite literally), there are always a wave of absurdly overly-optimistic people who think the new cards will deliver 10lbs of bacon for the price of a can of Bacos? But no one listens. I've often thought, if I had the time, to go back to the old R600, G70, Vega, RDNA, Maxwell, Pascal etc. threads to catalogue some of the most egregious claims (it's often the same posters...). 2080 TI for $499 is pretty close to the top of the list.

2080TI-level performance in September will cost nearly the same as it does now, which is ~$900. That's the model that nVidia and AMD have instituted over the past 3-4 years, and that's what we'll be getting with RDNA 2 and Ampere. The price/performance mark won't move much further than 10%.
 
People dreaming of a mythical $499 2080Ti-killer are just not understanding that the world has changed, and it's probably going to require more years before the benefit of enough hindsight for it to click and be obvious.

2080TI-level performance in September will cost nearly the same as it does now, which is ~$900.

Ok. I'll take this bait again.

AMD has consistently been touting 50% gain of RDNA2 over RDNA1 for similar parameters

Taking AMD at its word, (because they have had 3 years to work on RDNA 2), let us look at 2 scenarios:

1) A 6700 XT similar in size to 5700 XT but 50% better performance. That should place it between RTX 2080 super & 2080 ti. Based on the size of the chip, it should cost $500 at launch. Of course, you could argue based on value, AMD might price it higher, if NVIDIA's RTX 3070 or equivalent is priced at $700 & above, which I think is unlikely. I think both RTX 3070 & 6700 XT will be around $500 & perform at or better than RTX 2080 super

2) A 6900XT that is exact doubling of non-overclocked 5600 XT. This should be 15%?? better than an RTX 2080 ti & again reasonable to expect $999 unless AMD implements fancy cooling solutions that push the power all the way up, in which case it could be $1200 or more
 
AMD has consistently been touting 50% gain of RDNA2 over RDNA1 for similar parameters

50% gain in perf/watt.

Not 50% in absolute performance.

Looking at XBSX with 52 CU and boosted Clocks, it does NOT appear to be performing significantly better than RDNA2 would be at that clock speed.
 
Not that you're wrong, but if they increase perf/watt by 50% and then keep power draw the same, shouldn't that result in 50% more performance or thereabouts?

It never really works that way. There is pretty much a guaranteed "up to" in there.
 
People dreaming of a mythical $499 2080Ti-killer are just not understanding that the world has changed, and it's probably going to require more years before the benefit of enough hindsight for it to click and be obvious.

After doing video card reviews for a few years and gaming for so many more, I can say confidently you can almost always judge a card's overall performance by its real-world price, assuming its features are in-line with the mainstream. This isn't new, it's a return to normal. Unless it does something specifically different--being the very best counts as something different, but so could a number of other features like special cooling or a non-standard capability or compatibility--and it's not brand-new, the in-store/online pricing will reflect the general performance. The mining era was a freakout that messed with that rule of thumb, but it takes something aberrant in the market to sway prices away from the performance.


Usually yes. And this is the way it has mostly been working in the mid-range market where there has actually been competition. In the high end market, where there has not been competition, pricing has become inflated because they can.

If AMD can release a product that is truly competitive in the high end though, I don't think it is unreasonable that we would see some price reductions in both camps. I mean, look at how much cheaper Intel's high end consumer CPU's are this generation compared to last.

It wouldn't be instant. AMD would likely slightly undercut Nvidia to try to gain back some market share and reestablish themselves in the high end. Nvidia would likely resist lowering prices in the face of a competitive product from AMD, trying to point to features and gimmicks like raytracing to distract from this fact and continue to charge high prices. At that point the market would decide. If people still buy Nvidia GPU's at high prices despite a competitive offering from AMD, they will continue to sell them at that price. If - however - people start shifting to AMD as a result, we would likely see Nvidia start to respond pricing wise. Then there would be a back and forth, with pricing settling in a little bit lower than where it is today.

It is not unreasonable to expect that competition would work out for the consumer. It's pretty much a law of economics. Companies abuse lack of competition by price gouging customers as a rule. No exceptions. Pricing in markets goes down when competition becomes viable.

I would agree with odditory though. A $499 2080ti killer is completely unrealistic. The days of sub $500 flagships are permanently behind us. The costs of manufacturing GPU's have gone up to the point where if that is all they could charge for one, they'd likely just pass on making it all together (which is what AMD has done this far) because they likely just wouldn't be able to make money doing it, and a corporation, any corporations exists solely to make money for it's shareholders. There is a good chance it would lose them money.

I'd imagine break even (not based on gross profit, but with R&D spending included) is somewhere in the $600 to $650 price range for a flagship GPU. I think it is totally possible a 2080ti killer could settle in at somewhere between $749 and $899 depending on how eager AMD is to take market share.

That's cheaper than Nvidias current $1,200 offerings without competition in the high end, but still reflective of the fact that flagship GPU's are more complex products today than they were 15 years ago.
 
Last edited:
The driver thing is a little over stated for some of the older cards. I had a 4800 series, and an hd 6950, they were both excellent, same with my X1800xt. But yeah, before they were bought by AMD I think they made better products. They're turning around, but it's taken some time. The RX 5700 I have has been flawless since the 2020 release.

I never found driver bugs to be a problem with AMD. I haven't used them since the Catalyst era though. I thought the interface was a little clunky and less usable, with fewer options than with Nvidia but bugs? No.

Only time I ever had problems with Nvidia was when I tried to run Crossfire. For many years I thought that was an AMD issue, until I tried running SLI with dual 980ti's in 2015, and found that no, it's just an mGPU thing. Doesn't matter which brand you use.
 
  • Like
Reactions: Axman
like this
50% gain in perf/watt.

Not 50% in absolute performance.

Looking at XBSX with 52 CU and boosted Clocks, it does NOT appear to be performing significantly better than RDNA2 would be at that clock speed.
50% performance per watt, I think that was in mobile parts, may not be completely linear at higher power. But, if they keep the same power level (which is honestly pretty tame on the 5700xt, I wouldn't doubt a slight increase), and it's actually 50% per watt increase (assumed best case, not reality)... perf * new_watts = old_perf * old_watta*1.5. if watts are the same, they cancel out.

I think power remaining the same or going higher is likely, I doubt that 50% scaling works out for top end parts. I'm not suggesting they will actually see 50% on the top end, just explaining his math. He even said giving them the benefit of doubt if it's actually 50%, meaning he doesn't think it'll actually be that high.
 
I think that most speculations will be a bit wrong because AMD will try to switch to something innovative as they leak to us. They said that they will give us card for good 4k experience.
They will try to get "the crown" with "halo" product so this mean that they want market share and this can be gain only with top product. So for all this they need really good product with which one to continue to grow their image.

I don't know how good will be Big Navy but for all this it must be better than 3080 ti (or what ever will be the name). And if they success to catch and Titan, then the prices will be really interesting from both companies :)
 
AMD may not want to start a price war, but they have to compete somehow (and Ryzen has shown they can successfully compete on price and still win).

The Navi won't be $500 but I think in the $700-800 range is a possibility. Maybe more if they have a good lead on Nvidia. I mean, a 2080 Ti killer for $899 could still sell and give AMD good publicity.

Thing is, they can't come to the party 2 years late, with the same performance and the same price. That's a dead end move.
 
Its not gouging, its called what the price the market will bare. For example, shoes are on average 2-3x's the cost in Canada as they are in the US, it isn't that it costs more to get the shoes here, or that we have more disposable income than an average american. It is literally the price that the Canadian market will pay for shoes, so it is the price that businesses charge.

There is no law that tells businesses what they can charge or says that a markup is fair/unfair, it is a free market, and in a free market businesses set their own price.
Actually this is what make Apple the business model we need. If you are an authorized Apple dealer or reseller you must sell at their MSRP or your inventory is pulled. no gouging is allowed. This is why when a phone is released the price is fixed irregardless of demand.
 
The gouging is usually from third party sellers or ebay scalpers.

If I remember correctly, Newegg or Best Buy or whatever will sell at a reasonable price (maybe not Apple-style locked but usually around MSRP).
 
Actually this is what make Apple the business model we need. If you are an authorized Apple dealer or reseller you must sell at their MSRP or your inventory is pulled. no gouging is allowed. This is why when a phone is released the price is fixed irregardless of demand.

what? (gpus sell at msrp in 95% of my exp)
 
Last edited:

Manufacturers' suggested retail price is usually about 10-15 percent above resellers' price before market action. Retail prices vary between 60 and 90 percent of suggested pricing.

This isn't just video cards, this is all MSRP markets.
 
Back
Top