They will sell ever card on release I bet you.
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Rumor: 3080 31% faster than 2080 Ti
- Thread starter imsirovic5
- Start date
My 1080Ti remains "good enough for what I ask of it". When the time to upgrade comes, I'm hoping I'm not in the middle of a 20xx cycle where gen-1 features are tacked on (and unavoidable) while using those features as justification for an almost comical price-hike.
/dev/null
[H]F Junkie
- Joined
- Mar 31, 2001
- Messages
- 15,182
I'm at the same point. 1080Ti is more than fast enough for my linux/3900x box @ 1080p & my 1440p i5-8400/win10 box. I think it's going to be quite a while before I upgrade.
Delicieuxz
[H]ard|Gawd
- Joined
- May 11, 2016
- Messages
- 1,668
That didn't happen with the 2000 series. And people are poorer now than they were when the 2000 series released.
It did? Sure it slowed down now. Nvidia is not stupid. They will milk every dollar out of people they can. If they weren't happy with sells you would of think they would of dropped prices by now. People that buy 2080ti are not people that hurting. Contrary to what media want you to believe. Not everyone is suffering right now.
![Zarathustra[H]](https://cdn.hardforum.com/data/avatars/m/9/9298.jpg?1707758471){kind=avatar}
Zarathustra[H]
Extremely [H]
- Joined
- Oct 29, 2000
- Messages
- 38,866
I was sitting through a long phone conference with people droning on about things not relevant to me, so I decided to run the numbers.
I pulled benchmarks from 9 titles (because they were easily available), all in 1080p, Crysis 3, Bioshock Infinite, Tomb Raider, Battlefield 4, Metro Redux, Sleeping Dogs, Thief, Watch Dogs and Dragon Age Inquisition.
The mean performance increases for these titles, generation over generation for every x80 GPU since the beginning of the DX11 era is as follows:
GTX980 -> GTX1080: 63.0%
GTX780 -> GTX980: 31.9%
GTX680 -> GTX780: 26.8%
GTX580 -> GTX680: 41.3%
GTX480 -> GTX580: 16.3%
Mean for all generations: 35.9%
The impressive increase going to the 1080 is really quite amazing, considering that gains get much more difficult with time, not easier, so I consider it an outlier, not something we should expect to see repeated.
Part of the difficulty here is that these are marketing names, and they don't always mean the same thing from generation to generation. That, and some of these generations are real generations, others are the same designs, but released as a new generation due to yield increases and better binning. Nvidia tries to package it up to tell a nice story in marketing language, but behind the scenes it is different.
480->580->680->780->980
looks nice and clean, but behind the scenes we are talking about:
GF100 -> GF110-375-A1 -> GK104-400-A2 -> GK110-300-A1 -> GM204-400 -> GP104-400-A1
Different size and tier dies within a generation all based on how that gen turned out when it actually made it to silicon.
It's not like they start out with a chip design and say, "this is going to be the next x80 chip.
It's more like, they design a full range of chips within a gen. xx100 chips are usually the largest full chips that get used in Tesla cards. Then the higher the number, the smaller the chip. Then they pick and choose based on yields, costs and market conditions which chip becomes x80, which becomes x70, etc. etc.
Last edited:
Zarathustra[H]
Extremely [H]
- Joined
- Oct 29, 2000
- Messages
- 38,866
That's a good question.
I mean, these are rumors. For all we know some youtuber pulled them out of his or her ass, so they should be taken with a truckload of salt.
They probably don't mean anything at all.
That said, rumor is that Nvidia is - if possible - even more all in with RTX in this upcoming gen than with the previous one, so expect that to be their focus.
imsirovic5
Limp Gawd
- Joined
- Jun 21, 2011
- Messages
- 341
I agree, most people who are in higher paying professional sector are able to telecommute. For those people the bank accounts are swelling at the moment as they are unable to spend a lot of their discretionary income on eating out or traveling. Unfortunately, the people that are impacted the most by COVID are in lower paying service jobs who are typically not the consumers of highest end hardware. So I would not be so sure about any weakness at the very top end.
Last edited:
Delicieuxz
[H]ard|Gawd
- Joined
- May 11, 2016
- Messages
- 1,668
Apparently, it didn't. Nvidia is being sued by investors for misrepresenting to them the importance of crypto-mining to their sales, resulting in Nvidia having made far more GPUs than they could actually sell following the crypto-bubble bursting, which resulted in Nvidia having a "glut of unsold GeForce inventory", according to the lawsuit.
https://regmedia.co.uk/2020/05/15/nvidiashareholdersuit.pdf
"November 15, 2018: Investors Learn New Information Regarding NVIDIA’s Reliance on Crypto-Miners, Exposing a Glut of Unsold GeForce Inventory"
Last edited:
/dev/null
[H]F Junkie
- Joined
- Mar 31, 2001
- Messages
- 15,182
I think if they dropped all $400+ gpus $100, I think they'd sell a lot better. But that is just me...
GoodBoy
2[H]4U
- Joined
- Nov 29, 2004
- Messages
- 2,768
Flawed thinking. I have tons of games in Steam. Tons of pc gamers are just going to give up their existing software libraries, and buy all new on the next console because they don't like GPU prices?
That would be kinda stupid.
Lol, what? So because some people cannot afford something due to covid, they should just lower prices??? So we can expect to see $10,000 cars too?
You do not need the next GPU to be priced where you think it should be, to survive covid, or life in general. you do not even need the next GPU, period. If someone is unemployed, next-gen GPU pricing is the furthest thing from their minds. Utterly ridiculous statement.
That is a entire different issue. I believe it was about 10xx blue and not the 20xx that they had a over abundance of left over stock. They ramped up production to fill crypto demand and the market crashed soon after. Was hard to sell new stock of 10xx cards when the used market got flooded with them. The 20xx were still flying off the shelves.
DooKey
[H]F Junkie
- Joined
- Apr 25, 2001
- Messages
- 13,560
Yeah, the 10XX series was overproduced and NV got stuck with having to take a bunch back from their partners if I remember correctly.
Delicieuxz
[H]ard|Gawd
- Joined
- May 11, 2016
- Messages
- 1,668
That's not really thinking things through. In fact, that's absurd.
People don't have unlimited disposable income and many people will be choosing between buying a console or upgrading their PC. If the new consoles are a much more attractive deal, then GPU sales won't be as strong. Also, buying a console doesn't mean that people lose their Steam libraries. I'm not sure why you would think otherwise.
What's utterly ridiculous is that comment of yours. Over 45 million people in the US have lost their jobs due to covid-19. That's a lot of people who are watching their spending and not buying things they otherwise would have. And many more millions of people have had their hours cut due to covid-19. Also, the stock market is down and the uncertainty of how long its recovery will be and what ups and downs it will have in the process makes people more cautious about spending money.
It seems to me you didn't factor-in the actual situation that's been going on. Maybe you have been unaffected and don't pay attention to the news, but the reality is very different than what you've assumed. And arguing that there isn't a reason for over-priced GPUs to have their prices reduced and making a pretentious 'you don't need the next GPU, period' argument when discussing the inarguably hyper-inflated prices of the 2000 series makes it sound like you're on Nvidia's payroll.
Last edited:
Auer
[H]ard|Gawd
- Joined
- Nov 2, 2018
- Messages
- 1,972
How many of those 40mil in the U.S that lost jobs are upgrading their GPU every 2 years?
Your numbers mean nothing, sorry
GoodBoy
2[H]4U
- Joined
- Nov 29, 2004
- Messages
- 2,768
You should read my post again.
If you are unemployed, you shouldn't be concerned with next-gen GPU pricing. If you are, your priorities are out of whack, and no one on this forum can likely assist you. Reality-check, counseling, there are avenues to assist you but they are not in the Hardforum.
If you are unemployed, you shouldn't be concerned with next-gen GPU pricing. If you are, your priorities are out of whack, and no one on this forum can likely assist you. Reality-check, counseling, there are avenues to assist you but they are not in the Hardforum.
Delicieuxz
[H]ard|Gawd
- Joined
- May 11, 2016
- Messages
- 1,668
Your comment couldn't be more wrong.
As of June 18th, over 45 million people have lost their jobs due to covid-19.
The employed population in the US is around 158 million: https://www.statista.com/statistics/269959/employment-in-the-united-states/
That's 28.5% of people with incoming disposable income who might not be making unnecessary purchases right now.
That's a huge loss to potential profits for new products right now.
I read your post correctly and responded to it accordingly. People who shouldn't be concerned with new GPU prices right now make-up a very significant chunk of the normal consumer market - which means that businesses, GPU businesses included, have to take that into consideration when they're pricing their next GPUs.
If you think my response was about what people who can't afford GPUs should be doing, then you read it wrong. It's about what impact all the people who have less money to spend right now will have on the prospects for sales of high-priced GPUs (especially when more affordable consoles will be available).
The idea that such a large decrease in income-earners wouldn't impact business revenues is absurd. It already has impacted business revenues to a large degree.
Gideon
2[H]4U
- Joined
- Apr 13, 2006
- Messages
- 3,558
So they just wont upgrade their PC graphics and if new games come out that they feel their current PC cant handle then perhaps they will go with the console since it's cheaper then upgrading the graphics card at that point. If you want to spend 1000 bucks for a 3080 on the person I quoted then go right on ahead, but very few would be ok with that. Gaming quit being worth that kind of cash some time ago for me as my current 1080 still works just fine.
Last edited:
The glut of unsold inventory were Pascal cards not Turing cards.
Turing cards sold really well.
Gideon
2[H]4U
- Joined
- Apr 13, 2006
- Messages
- 3,558
What do you consider really well as the 1060 has more cards in use on steam then all of Turing.
T4rd
Fully [H]
- Joined
- Apr 8, 2009
- Messages
- 20,112
I think the more relevant question is with DLSS or without, esp. with DLSS 2.0 looking as good or better than native resolution.
Delicieuxz
[H]ard|Gawd
- Joined
- May 11, 2016
- Messages
- 1,668
Can you provide a source for that?
The GTX 1080 Ti was out of production around April 2018 ---- Edit: That was according to one source, but others mention it going out of production in November 2018
I don't know when the other models went out of production, but I'm guessing that it was around the same time.
It was reported in July 2018 that Nvidia was stockpiling GPUs: https://www.pcgamesn.com/nvidia-next-gen-graphics-card-inventory-stockpile
I'm not sure when the RTX 2000 series entered production, but this article reported that GDDR6 would enter production in late June, or early July: https://wccftech.com/nvidia-geforce...july-gddr6-mass-production-timeline-confirms/
Also, the price of Bitcoin crashed from January 2018 until September 2018, when the RTX series released, and then underwent another crash from November to December 2018, after the RTX series had released.
The period in which Nvidia allegedly misrepresented the contribution of crypto-mining sales to their revenue covers the release of Turing and carries on long after.
The class-action lawsuit period is from May 10, 2017, and November 14, 2018.
According to the lawsuit, shortly after the 2000 series released, GPUs were not flying off the shelves. Instead, demand for them was in sore decline:
"On November 15, 2018, the relevant truth behind Defendants’ deception was more fully revealed. Defendants announced that NVIDIA had missed analyst expectations for the third quarter and was revising its revenue guidance for the fourth quarter to reflect a 7% decline year-over-year. Attributing the reversal to a “sharp falloff in crypto demand” for NVIDIA’s Gaming GPUs, NVIDIA revealed that it would make no shipments into the distribution channel of—and thus recognize no revenue for—the midrange GeForce GPUs that miners had favored. The promised demand from gamers simply did not exist, and it became fully apparent to the market that, contrary to Defendants’ earlier representations, NVIDIA’s revenues were unduly dependent on cryptocurrency mining."
Last edited:
Nice work. Glad you found something productive to do
Just want to add to your analysis. If you look through the figures what's and maybe go back a few more years to when the 7800 GTX cards released, you will see a clearer picture. If you notice any time there is a die shrink performance is over 40% and sometimes well over 40%.
Using your figures.
GTX980 -> GTX1080: 63.0% Almost a double die shrink, went straight from 28nm to 16nm skipping 20nm.
GTX780 -> GTX980: 31.9% No die shrink.
GTX680 -> GTX780: 26.8% No die shrink
GTX580 -> GTX680: 41.3% Die shrink.
GTX480 -> GTX580: 16.3% No Die Shrink.
The same happened in earlier years too, with only one exception, going from the 7800 GTX to the 8800 GTX. That was a big jump in performance without a die shrink 70%+. They achieved this with a massive die size increase, over double the size of the top 7 series card.
Nvidia managed to achieve a 30%+ increase going to Pascal to Turing without any die shrink. But now the next gen is going to be after a die shrink, like Maxwell to Pascal and Fermi to Kepler.
Just food for thought when you are trying to work out the performance of the next cards from Nvidia.
I have no doubt that the new Flagship will be faster than the current flagship, but what I don't know is launch date, availability, cost. Regardless I doubt I am replacing my 2080TI this gen, I'm only running 1440p and with my eyes I can barely tell the difference between 60+ fps and 120+ fps. So I think I will be sitting this gen out unless I come into some money or fall in love with a game that requires more power. That said I am a big fan of better performance especially in the lower price brackets, the better the average is the better it is for everybody. So while this is potentially awesome news I am still more excited for what the new Xbox and PS bring to the table, not because I use them but because they drastically move the bar.
There couldn't have been a GLUT of unsold Turing cards as they were only released at the end of September 2018 and only two models the 2080Ti and the 2080 and stocks on release were extremely low. The GTX 2070 was only released near the end of October 2018.
Read the article you are quoting from.
Nvidia hadn't released any mid range Geforce GPUs by November 2018.
The 2060 Turing card didn't arrive until January 2019. The first GTX Turing card didn't arrive until March 2019.
They had a stock pile of cards after the mining bubble went into decline from around February 2018.
https://www.pcgamer.com/crypto-mining-crash-leaves-nvidia-with-excess-inventory-of-pascal-cards/
https://venturebeat.com/2018/11/17/...ed-nvidia-to-lose-23-billion-in-market-value/
Not all of Turing. Just all of Turing RTX cards.
So how does that compare to Navi? RTX 2080 by itself outsells all of Navi, and it's in a much more expensive price class.
Pascal was an anomaly.
It was one of, if not the best generation of NVidia cards ever in terms of the performance gains, it was one of the longest lasting generations, so it had more time to saturate the market, and it happened during the crypto boom so excess inventory was pushed into the channels.
It was the perfect storm. Failing to live up to Pascal is not a failure, it's just a return to normal.
SilverSliver
[H]F Junkie
- Joined
- Feb 23, 2004
- Messages
- 12,526
I'm fine with a 30% increase in performance.
Hagrid
[H]F Junkie
- Joined
- Nov 23, 2006
- Messages
- 9,163
I seriously doubt that. The 3080ti at 30% would be a stretch. There is 0 reason they have to make it way faster. They have no competition.
Anyone that actually believes a low power, peashooter APU in a console is "gunna beat a 2080Ti" is delusional. If AMD had any GPU tech that came anywhere close to matching or beating Nvidia, they wouldn't need to hide it in a console.
The reality is it'll be another console gen that struggles to touch true 4K60 without upscaling and checkerboarding cheats.
Last edited:
When any company price gouges it is very easy to spot in their Financials and if there is any sort of competition unless they are colluding than it can't last for long. Neither AMD nor nVidia show signs of price gouging in their numbers, both AMD and nVidia are similarly priced, this simply comes down to a greatly increased manufacturing cost increase, you want to see prices like we did from 5+ years ago, not a problem just produce the chips on 16nm or larger lots of plants capable of that. If you want to release something competitive you are going to have to use TSMC, and you are going to have to pay for it. TSMC has 4 plants currently operating that is able to produce chips for 7nm and they are booked solid running 24/7 and those are producing chips for a lot of people, Apple, Intel, nVidia, AMD, Tesla just to name a few. And they all have to bid against each other for their time, and once their window closes that's it no more of that lineup until their next window opens up. So that means fewer chips, at a higher manufacturing cost, with limited restocking capabilities which means pricing has to be front loaded. TSMC has one more fab being built now so in a few years when it is complete we could see a cost reduction but until TSMC gets some competition and manufacturing costs go down we can expect to be paying some higher prices for what we are getting.
Delicieuxz
[H]ard|Gawd
- Joined
- May 11, 2016
- Messages
- 1,668
OK, so that makes sense.
However, I still don't see an indication that the RTX 2000 series sold well.
In June 2020, RTX cards account for 8.39% of Steam user GPUS. The 1000 series accounts for 40.31% of Steam user GPUs.
Turing GPUs account for 15.56%, while Pascal GPUs account for 33.14%.
https://store.steampowered.com/hwsurvey
The 2000 series certainly hasn't had nearly the same draw as the 1000 series, likewise for Turing compared to Pascal. So, whatever constitutes selling well appears to be measured by a distinct barometer than Pascal. In which case, I would ask: How are those sales figures determined to be good when they aren't measured as relative to the previous generation of GPUs?
Pascal was out for 2 years and 4 months before Turing released. Turing will have been out for 2 years when Ampere releases. Not too much of a difference there.
That Pascal had a huge performance gain did help it to sell. But saying that RTX 2000 had a relatively poor performance gain and an unjustifiable price-to-performance proposition to go along with it only explains that the 2000 series in-fact didn't sell well. Identifying the reason why it didn't sell well isn't an argument that it actually sold well.
If it's like last gen 5% will be a 30% increase in cost... Which means a 30% performance increase would be 180% increase in cost. Let's hope that's not the case
Over? The 2080 Ti was released 22 months ago. That is less than two years. Have you considered that the benchmark could be from the 3070?
Zarathustra[H]
Extremely [H]
- Joined
- Oct 29, 2000
- Messages
- 38,866
He does have a point. Pricing is not set based on cost. Pricing for products is set based on what the market will bear. In a healthy surging economy, people are willing to spend more than they are in a down economy.
Product pricing works something like this (ignoring th eproduction side for simplicity). We all know that demand changes with price. The cheaper something is, the more people are likely to buy it, but what that demand curve looks like is very different from product to product, and depending on the state of the economy. Marketeers try to predict the shape of this demand curve. Once they have, they use it to set pricing based on what will maximize profits. Qty you expect to sell at a given price X the price = Predicted revenue from the product. They then attempt to maximize this figure. For some products it makes more sense to sell higher volume at lower prices, for others it makes more sense to sell lower volumes at higher prices.
Production figures very much into this. If you add to the equation above, the predicted cost of manufacturing a certain number of units, the maximized point shifts as well. Sometimes you have idle plant capacity, in which case you can easily increase production volumes without much of cost increase. In other cases you need to add extra shifts, or builkd new facilities, which drives the average cost per unit higher.
A few years ago when AMD launched the FX9xxx CPU's are high prices, people thought they had lost their minds, but it all was based on this type of calculation. They knew some fans out there who wanted the fastest chip they could get without buying Intel and were willing to pay for it. These were highly binned chips, so available wuantity would be low, so they set the price high enough that they only expected to sell low quantities, in order to maximize income given that you can't sell product you don't have.
Anyway, long story short, yes, the health of the economy definitely factors in to the demand curves above. it is unclear to what extent it factors in, and where it moves the maximized point on the demand curve (towards higher prices with lower volumes, as there is a small wealthy group who haven't been so impacted so they can still afford expensive GPU's and re willing to pay for them, or towards lower prices to maintain larger volumes.
The yields of this generation and how much volume they are bound by contract with TSMC to buy will likely also factor in. If they are bound to buy X units by contracts signed back in 2019, and now they suddenly think the market is softer due to COVID19, they may decide to take a hit on profit margin and sell them more cheaply, as that is better than paying for silicon and having it sit unsold.
How well AMD's Big NAVI with RDNA2 performs will certainly also factor in a bit.
Lots of things could happen, and you'd need so much data, a lot of it proprietary, in order to even try to make an educated guess at where they are going to land, that it is tough to say. But will the weaker market due to COVID19 have some sort of impacts on their pricing model? You bet.
Arguments about prices dropping because of CV-19 are nonsense. Most parts are increasing in price.
Yeah Covid is driving manufacturing costs up, mining operations are running at reduced capacity so raw material costs from reputable sources are going up, manufacturing facilities producing the chemicals needed to refine the raw materials as well so costs are up there as well. TSMC is running facilities at reduced capacity to comply with social distancing requirements in the countries they operate in which for the most part are stricter than what you are seeing in the US so output is down so supply is down. Costs are up across the board, so I am expecting to see a release that is roughly the same price points adjusted for inflation not a decrease. Yeah because Covid is messing with people's finances by a lot makes luxury products a hard sell, but it's not like nVidia and AMD can afford to sell parts at a loss because of it, it just means they make fewer of them, and scale down R&D expenses to compensate, what I am expecting is that the next generation is going to be stifled because of spending reductions being made now.
Delicieuxz
[H]ard|Gawd
- Joined
- May 11, 2016
- Messages
- 1,668
The second sentence in that post is a non sequitur from the first sentence in that post.
Nvidia doesn't sell GPUs for the price of what they cost to produce. Nvidia sells them at an upcharge in order to make profits, with the difference between the cost to manufacture and distribute and how much they sell for being the profit-margin.
How much people are willing to pay for a graphics card guides how much Nvidia choose to upcharge their graphics cards. If Nvidia think they can charge more and enough people will still pay it, then they'll increase the price. If they think they'll make more money from a lot more people being willing to buy their GPUs at a lower performance-per-dollar point, then they'll lower the prices of their GPUs.
The RTX 2000 series is surely not as exorbitantly and prohibitively expensive as it is because Nvidia kept the same profit-margin on them compared to Pascal but the expense of developing and manufacturing them just cost that much more. It was surely because Nvidia are greedy and thought they could get more for them and so pushed the profit-margin way up - but the Steam hardware stats don't appear to validate Nvidia's gambit. If part costs increase and the market's disposable income simultaneously decreases, then that might mean that Nvidia have to reduce their product cost along with their profit-margins in order to maintain their volume of sales. The new consoles are also a competitor to 3000-series sales and that too surely factors into what price they'll be set at.
Since the 2000 series already wasn't within most people's budgets, and since the consumer market overall has less disposable income now due to covid-19, it makes sense that Nvidia would take that into account when pricing their new products. That's not me saying that Nvidia will price the 3000 series cheaper. But there are a lot of reasons for them to do so.
Last edited: