AikanikuluksiI could hardly care less about the RTX 4060 as a product. Based on how the RTX 4000 series has been positioned and priced so far, I don't feel the slightest interest towards it, and the RTX 4060 will obviously follow in the same tracks. Details are irrelevant.

But just as a phenomenon, I do find it interesting/strange to see how much discussion the topic generates.

Why is it strange, anything related to money is going to fuel debate. If there’s a want, need, and money involved, it’s going to be discussed.

john_AMD is strong and cheaper under $600 and people still buy Nvidia cards that sell over MSRP.
So....

I can't fix stupid, so........

AusWolfExcept when they don't (Batterygate).

I think buying a 3070 or a 6700 XT today is a much better value than waiting half a year for a card with a similar performance and MSRP.

You can get a 6800XT for not much more than $400. I'm sure 3080's too are crazy prices. 4060 will not be $400 based on Huang's ludicrous prices so far. I'm not sure where the OP came up with a 10% price increase. I'll bet a 7600XT will kick the crap out of a 4060 too.

john_Well, in the example I said that, if it comes as an RTX 4070 and is offering the same performance (obviously in today's titles, because Nvidia could just stop optimizing for the 3000 series) as the RTX 3070, I did said that then it will be a matter of comparing new features and efficiency.

But we have to have a basic way to compare cards from generation to generation. And average game performance at the same price should be an indication. Features are not for everyone and are not present in every game, or needed in every game. Then again we shouldn't start paying for those in advance. We shouldn't treat any new feature as an excuse for a price increase. These are gaming cards. Manufacturers should use those features to differentiate themselves from the competition. Paying extra for a card because in integrates a faster AV1 encoder for example, if we are never going to use that feature, is just let's say, not smart. They are gaming cards. Sold as gaming cards. If RTX 4060 can run a weather simulation faster than an RTX 3060 is irrelevant in gaming. Have we payed extra for Vulkan support or DirectX 12 support? have we payed extra for video encoding support? Should we? Should we have to pay extra for DP 2.1 support? Did we payed extra for HDMI 2.1 support? HDMI 2.0? HDMI 1.4?
If we go by that logic, RTX 4060 should be costing $2000, RTX 3060 over $1500, etc. because manufacturers are adding new features the last 20+ years.

Well, I wasn't saying we should be ok with increasing prices no matter what.
You said that it makes sense to get a previous generation card for about the same amount of $$$, I said that, while that makes sense, it may not make sense for everyone, because of the reasons I mentioned.

for me, this is kind of a good news but also not the best. At that point most if not all the folks who are looking to upgrade their GPUs are likely getting a 3070Ti or a 3080 should the prices fall even more for those segment. The only thing I see the 4060 being good at is the DLSS 3.0, better RT performance and (hopefully) better rasterization performance over the 3070 it was supposedly meant to compete/replace.

Hey look another push to sell those many Ampere cards still at some premium.

Will you fall for it?

Vayra86Hey look another push to sell those many Ampere cards still at some premium.

Will you fall for it?

I will......

When hell freezes over.

EasoI am sure I will be seeing this in EU for 600 EUR... xD

The dollar is already ~1:1 with euro, good luck with that.

bugWell, I wasn't saying we should be ok with increasing prices no matter what.
You said that it makes sense to get a previous generation card for about the same amount of $$$, I said that, while that makes sense, it may not make sense for everyone, because of the reasons I mentioned.

I think I haven't said that. Either you confused someone else's post with mine, or I forget what I write.
I am only comparing raw performance (not DLSS/(soon)FSR 3.0 software tricks), to come up with a conclusion about generational jump, stagnation, or even regression in a certain price point. And instead of features or efficiency or model names, I am using price points as the main parameter. I don't say "buy the old one at the same price". That doesn't make sense anyway.

"What fps was I getting on average with a $200 card in 2016? In 2020? In 2022?".

Well, you know something? Let's answer that, but let's go far in the past. Someone posted somewhere that people shouldn't stay in the past, but focus on present and the future. Considering things are changing, with new generations coming every two years instead of one and usually offering more features than more raw performance, people remembering how the market was 10-15 years ago, are a problem for AMD, Intel and Nvidia today. Because they are more difficult to be impressed by a 10% generation jump in raw performance and a software trick like DLSS/FSR that where called "cheating" 15 years ago.

Let's see, with prices at around $149-$249(there was huge competition in the past and prices where fluctuating rapidly) and keeping it simple by looking only at Nvidia options and based on TechPowerUp's database.
2007: 8600 GTS, $199
2008: 9800 GT, $160, +225% performance increase over 8600 GTS
2009: GTS 250, $199, +24% performance increase.
2010: GTX 460 768MB, $199, +51% performance increase.
2011: GTX 560, $199, +25% performance increase.
2012: GTX 660, $229, +39% performance increase.
2013: GTX 760, $249, +20% performance increase.
2014: ?
2015: GTX 960, $199, +10% performance increase.
2016: GTX 1060 3GB, $199, +70% performance increase.
2017: ?
2018: ?
2019: GTX 1660, $219, +30% performance increase.
2020: ?
2021: ?
2022: RTX 3050 4GB, $199, +3% performance increase.

Looking above we see a steady increase in performance year after year from 2007 to 2013. The huge performance increase from 8600GTS to 9800GT was a combination of 8600GT/S being junk cards for their prices, while 9800 GT was a great card. Also AMD was pushing really hard, forcing Nvidia to drop prices. After 2013 the only good card was the GTX 1060, even in it's 3GB form. That's why this card is still relevant today. What have we seen in a period of 7 years in that price category after 2016? Honestly? Nothing. About 35% increase and new features that could be useful in games, but not necesserily helpful. I mean, RT support in a 4GB RTX 3050? Maybe DLSS 2.x support is the only thing here worthy of mentioning.

john_I think I haven't said that. Either you confused someone else's post with mine, or I forget what I write.
I am only comparing raw performance (not DLSS/(soon)FSR 3.0 software tricks), to come up with a conclusion about generational jump, stagnation, or even regression in a certain price point. And instead of features or efficiency or model names, I am using price points as the main parameter. I don't say "buy the old one at the same price". That doesn't make sense anyway.

"What fps was I getting on average with a $200 card in 2016? In 2020? In 2022?".

Well, you know something? Let's answer that, but let's go far in the past. Someone posted somewhere that people shouldn't stay in the past, but focus on present and the future. Considering things are changing, with new generations coming every two years instead of one and usually offering more features than more raw performance, people remembering how the market was 10-15 years ago, are a problem for AMD, Intel and Nvidia today. Because they are more difficult to be impressed by a 10% generation jump in raw performance and a software trick like DLSS/FSR that where called "cheating" 15 years ago.

Let's see, with prices at around $149-$249(there was huge competition in the past and prices where fluctuating rapidly) and keeping it simple by looking only at Nvidia options and based on TechPowerUp's database.
2007: 8600 GTS, $199
2008: 9800 GT, $160, +225% performance increase over 8600 GTS
2009: GTS 250, $199, +24% performance increase.
2010: GTX 460 768MB, $199, +51% performance increase.
2011: GTX 560, $199, +25% performance increase.
2012: GTX 660, $229, +39% performance increase.
2013: GTX 760, $249, +20% performance increase.
2014: ?
2015: GTX 960, $199, +10% performance increase.
2016: GTX 1060 3GB, $199, +70% performance increase.
2017: ?
2018: ?
2019: GTX 1660, $219, +30% performance increase.
2020: ?
2021: ?
2022: RTX 3050 4GB, $199, +3% performance increase.

Looking above we see a steady increase in performance year after year from 2007 to 2013. The huge performance increase from 8600GTS to 9800GT was a combination of 8600GT/S being junk cards for their prices, while 9800 GT was a great card. Also AMD was pushing really hard, forcing Nvidia to drop prices. After 2013 the only good card was the GTX 1060, even in it's 3GB form. That's why this card is still relevant today. What have we seen in a period of 7 years in that price category after 2016? Honestly? Nothing. About 35% increase and new features that could be useful in games, but not necesserily helpful. I mean, RT support in a 4GB RTX 3050? Maybe DLSS 2.x support is the only thing here worthy of mentioning.

Can't say I disagree with any of that. After all, I'm still using my 1060 6GB. But I wasn't upgrading every generation anyway. I think my upgrade path was 460 -> 660Ti -> 1060 6GB.

bugCan't say I disagree with any of that. After all, I'm still using my 1060 6GB. But I wasn't upgrading every generation anyway. I think my upgrade path was 460 -> 660Ti -> 1060 6GB.

I don't upgrade often either. Still using an RX 580 for whatever gaming I find time to do.

But when there isn't constant generation improvement, when the time comes that someone decides to upgrade, they will be in front of a dilemma. Buy something that it is not much better in raw performance compared to the old card, or just pay much more?
Someone having bought a 9800GT in 2008, after 5 years of use, will pay (probably?) $199 for a discounted GTX 6600 in 2013, or $249 for a GTX 760. The performance jump is HUGE and of course plenty of new features like DirectX 11/12 support, Vulkan support, better OpenGL support, better media codec support, better efficiency, better media encoding, current drivers, lower idle power consumption, modern video out connectors. The same for someone upgrading from GTX 560 from 2011 to a GTX 1060 3GB in 2016, with G-Sync/VRR support and (funnily) FSR 2.x support.

But what are the options for you now? Buy a GTX 1660? Why? Pay extra and buy an RTX 3050 8GB? Why? Go for an RX 6600? If you care about RT performance and after all those years you are convinced like so many others that AMD's drivers are bad, again, why?
So you will (for example) end up paying much more for an RTX 3060 or an RTX 4060, or try your luck in the second hand market for an RTX 3070. You will get probably that same generational performance increase with an RTX 4060, but for probably double cost of what you had payed 5 years ago to go to GTX 1060. And in 5 years that cost will increase even further, because Nvidia is moving prices up and it seems to tie the performance increase with the card's price. The pricier the model, the higher the performance increase from generation to generation.

john_I don't upgrade often either. Still using an RX 580 for whatever gaming I find time to do.

But when there isn't constant generation improvement, when the time comes that someone decides to upgrade, they will be in front of a dilemma. Buy something that it is not much better in raw performance compared to the old card, or just pay much more?
Someone having bought a 9800GT in 2008, after 5 years of use, will pay (probably?) $199 for a discounted GTX 6600 in 2013, or $249 for a GTX 760. The performance jump is HUGE and of course plenty of new features like DirectX 11/12 support, Vulkan support, better OpenGL support, better media codec support, better efficiency, better media encoding, current drivers, lower idle power consumption, modern video out connectors. The same for someone upgrading from GTX 560 from 2011 to a GTX 1060 3GB in 2016, with G-Sync/VRR support and (funnily) FSR 2.x support.

But what are the options for you now? Buy a GTX 1660? Why? Pay extra and buy an RTX 3050 8GB? Why? Go for an RX 6600? If you care about RT performance and after all those years you are convinced like so many others that AMD's drivers are bad, again, why?
So you will (for example) end up paying much more for an RTX 3060 or an RTX 4060, or try your luck in the second hand market for an RTX 3070. You will get probably that same generational performance increase with an RTX 4060, but for probably double cost of what you had payed 5 years ago to go to GTX 1060. And in 5 years that cost will increase even further, because Nvidia is moving prices up and it seems to tie the performance increase with the card's price. The pricier the model, the higher the performance increase from generation to generation.

Tbh, I used to upgrade more regularly because I could get the card at close to MSRP and sell my previous card for a good price, meaning I got to upgrade for anything between $0 and $100. That's also a factor.
And I really, really hope that with the mining craze over and Nvidia's latest financials, they will start looking into lowering prices. I do not share the opinion that Nvidia is somehow conspiring to raise prices. Rising prices are simply a consequence of increased pressure on leading manufacturing nodes and of GPU chips getting bigger. Historically, large dies and PCBs with support for >256 bit memory bus have always been expensive. While there is little Nvidia can do about fab capacity, I hope they will start looking at their chip designs from new angles.

bugTbh, I used to upgrade more regularly because I could get the card at close to MSRP and sell my previous card for a good price, meaning I got to upgrade for anything between $0 and $100. That's also a factor.
And I really, really hope that with the mining craze over and Nvidia's latest financials, they will start looking into lowering prices. I do not share the opinion that Nvidia is somehow conspiring to raise prices. Rising prices are simply a consequence of increased pressure on leading manufacturing nodes and of GPU chips getting bigger. Historically, large dies and PCBs with support for >256 bit memory bus have always been expensive. While there is little Nvidia can do about fab capacity, I hope they will start looking at their chip designs from new angles.

Increasing node costs are definitely an important factor, but Nvidia's ridiculous margins are the most important factor. This graph has been posted before, and the accompanying article is worth the read too. Another graph shows R&D as a percentage of revenue, and it has been almost constant since 2010 when margins were close to 40%.

bugTbh, I used to upgrade more regularly because I could get the card at close to MSRP and sell my previous card for a good price, meaning I got to upgrade for anything between $0 and $100. That's also a factor.
And I really, really hope that with the mining craze over and Nvidia's latest financials, they will start looking into lowering prices. I do not share the opinion that Nvidia is somehow conspiring to raise prices. Rising prices are simply a consequence of increased pressure on leading manufacturing nodes and of GPU chips getting bigger. Historically, large dies and PCBs with support for >256 bit memory bus have always been expensive. While there is little Nvidia can do about fab capacity, I hope they will start looking at their chip designs from new angles.

Yes, in the past we had more excuses to upgrade to the next series.

In Nvidia's latest financial report that was released a few days ago, they are predicting their profit margin to jump back to 65%.

They might drop prices in case AMD forces them, but I doubt AMD will do so. AMD only announced 2 models with prices of $800 and $900. We might have to wait many more months until we see something new and with a price tag under $400. And even then it will be questionable, in my opinion, if that new card will be bringing meaningful performance improvement over the older model and not just DLSS/FSR 3.0 marketing tricks.
As for Nvidia raising prices. It's not a conspiracy theory. It's a business decision. And it has nothing to do with GPUs getting bigger. CPUs also got bigger. We where buying 4 cores in 2010 and even after a period of stagnation because of AMD's Bulldozer, we buy today 12-16 bigger, more complex cores at the same prices(not talking about E cores here). Bus had gone mostly down, from 256bit-384bit even 512bit for the high end cards to 128bit-256bit, while power consumption in mid range models is lower than that of hi end models of the past. In the end the price of a high end GPU card is now much higher than the combination of an expensive motherboard, 32GB of RAM and a top 16-24 cores CPU. I doubt a top motherboard, top ram and top CPU model cost more to manufacture than a top graphics card. And I doubt an RTX 3050 with 8GB of RAM, cost more to manufacture than a b660 mATX motherboard, 16GB or DDR4 and a i3 12100 CPU.
Prices of GPUs have gone up much faster than anything else.

AnotherReaderIncreasing node costs are definitely an important factor, but Nvidia's ridiculous margins are the most important factor. This graph has been posted before, and the accompanying article is worth the read too. Another graph shows R&D as a percentage of revenue, and it has been almost constant since 2010 when margins were close to 40%.

Gross margins are largely irrelevant. Net/EBITDA margins are relevant, I haven't seen those yet.

Edit: Also, even if not directly comparable, AMD's gross margin isn't significantly lower: www.macrotrends.net/stocks/charts/AMD/amd/gross-margin
It's almost 50%, even if it dipped a little recently.

bugGross margins are largely irrelevant. Net/EBITDA margins are relevant, I haven't seen those yet.

Edit: Also, even if not directly comparable, AMD's gross margin isn't significantly lower: www.macrotrends.net/stocks/charts/AMD/amd/gross-margin
It's almost 50%, even if it dipped a little recently.

Net margins have increased dramatically too. AMD's gross margins have only reached 50% recently. AMD is no saint, but their GPUs, made on a more expensive node than Ampere, are now available for much more reasonable prices.

john_I don't upgrade often either. Still using an RX 580 for whatever gaming I find time to do.

But when there isn't constant generation improvement, when the time comes that someone decides to upgrade, they will be in front of a dilemma. Buy something that it is not much better in raw performance compared to the old card, or just pay much more?
Someone having bought a 9800GT in 2008, after 5 years of use, will pay (probably?) $199 for a discounted GTX 6600 in 2013, or $249 for a GTX 760. The performance jump is HUGE and of course plenty of new features like DirectX 11/12 support, Vulkan support, better OpenGL support, better media codec support, better efficiency, better media encoding, current drivers, lower idle power consumption, modern video out connectors. The same for someone upgrading from GTX 560 from 2011 to a GTX 1060 3GB in 2016, with G-Sync/VRR support and (funnily) FSR 2.x support.

But what are the options for you now? Buy a GTX 1660? Why? Pay extra and buy an RTX 3050 8GB? Why? Go for an RX 6600? If you care about RT performance and after all those years you are convinced like so many others that AMD's drivers are bad, again, why?
So you will (for example) end up paying much more for an RTX 3060 or an RTX 4060, or try your luck in the second hand market for an RTX 3070. You will get probably that same generational performance increase with an RTX 4060, but for probably double cost of what you had payed 5 years ago to go to GTX 1060. And in 5 years that cost will increase even further, because Nvidia is moving prices up and it seems to tie the performance increase with the card's price. The pricier the model, the higher the performance increase from generation to generation.

That just means that upgrading every 2-3 generations isn't necessary anymore. I mean, if we assume that you're happy with a 1060, then why would you look for upgrade options? It's pointless.

Oh, and by the way, AMD's drivers aren't as bad as they used to be. In fact, RDNA 2 drivers are as stable as Nvidia's. ;)

Edit: typo

AusWolfThat just means that upgrading every 2-3 generations isn't necessary anymore. I mean, if we assume that you're happy with a 1060, then why would you look for upgrade options? It's pointless.

Oh, and by the way, AMD's drivers aren't as bad as they used to be. In fact, RDNA 2 drivers are as table as Nvidia's. ;)

I can't speak for anyone else, but my 290X was rock stable throughout its life in a couple of different computers. The Vega 64 had a lot of problems about a year after I bought it and it turned out to be my motherboard. Since I replaced the motherboard, it has been as well behaved as the 290X ever was. If you're still rocking a 1060 or 580, this holiday season might be the right time to upgrade. The 6600 XT has been sold at $250 and the 6600 has dropped below $200 at times.

AnotherReaderI can't speak for anyone else, but my 290X was rock stable throughout its life in a couple of different computers. The Vega 64 had a lot of problems about a year after I bought it and it turned out to be my motherboard. Since I replaced the motherboard, it has been as well behaved as the 290X ever was. If you're still rocking a 1060 or 580, this holiday season might be the right time to upgrade. The 6600 XT has been sold at $250 and the 6600 has dropped below $200 at times.

The "AMD driver instability" argument is mostly used by people who have never touched an RDNA 2 card. I, personally, never had any such problem with any AMD card except for the 5700 XT. RDNA 1 drivers were truly a bit immature, but AMD has come a long way since then. The GCN era was awesome, too. It's mind-blowing to see the lengths people go to justify their bad choice of spending more on an Nvidia card than they should have.

The 6600 series are an amazing deal. People upgrading from a 1060-like card should look there, and not wait for an overpriced 4060, imo.

rv8000Correct, wrong terminology.

What I really meant was that in comparison to previous gen, the value proposition gets worse throughout the stack. Upon the release of either FE 4090 and 4080, the cost per frame at 4K was relatively close, if you already had $1200 for a GPU and can swing another $400 for the 4090, the 4080 is absolutely pointless from a pure value standpoint; different case as it seems the MSRP for the 4090 was a limited thing…

Strange gap between 4080 and 4090, agree. 4090 performance would be ok at MSRP (my opinion). Hope to get one the next 6-12 month.

That does not change the fact that everything is dramatically sliding up in price generation to generation (4080 is a horribly good example), to the point we’re getting tremendously worse value for what we’re paying than just a generation ago.

Yes. But is there any reason, to upgrade every generation? I upgrade any 2nd generation and I am fine with the performance jump.

Seriously though, how blind does someone have to be to call a 4080 at almost double the msrp of a 3080 a good product, let alone what BS is going to happen amongst lower tier 4000 cards.

Depends everything on the competition/price/availability of new the AMD generation. No real competition -> high prices for consumers.

SisyphusDepends everything on the competition/price/availability of new the AMD generation. No real competition -> high prices for consumers.

The 3080 is 20% slower than the 4080, but only costs half as much. The 40-series occupy a terrible position within Nvidia's own product stack. AMD has nothing to do with this.

AusWolfThe "AMD driver instability" argument is mostly used by people who have never touched an RDNA 2 card. I, personally, never had any such problem with any AMD card except for the 5700 XT. RDNA 1 drivers were truly a bit immature, but AMD has come a long way since then. The GCN era was awesome, too. It's mind-blowing to see the lengths people go to justify their bad choice of spending more on an Nvidia card than they should have.

The 6600 series are an amazing deal. People upgrading from a 1060-like card should look there, and not wait for an overpriced 4060, imo.

Depends on your income. If you bought rdna1 and lost 2 days to solve driver problems, you can calculate 2 working days income lost. Lets say 2*300$= 600$. If there was an insolvable issue with one of your preferred applications, you will never buy AMD GPUs again for the foreseeable future. Happens to me with the rd 6870. Never bought AMD since then, never had GPU driver problems again since that time.

AusWolfThe 3080 is 20% slower than the 4080, but only costs half as much. The 40-series occupy a terrible position within Nvidia's own product stack. AMD has nothing to do with this.

If they don't compete here, they will never have anything to do with it. :laugh:

SisyphusDepends on your income. If you bought rdna1 and lost 2 days to solve driver problems, you can calculate 2 working days income lost. Lets say 2*300$= 600$. If there was an insolvable issue with one of your preferred applications, you will never buy AMD GPUs again for the foreseeable future. Happens to me with the rd 6870. Never bought AMD since then, never had GPU driver problems again since that time.

What income? I thought we were talking about gaming.

If you don't want to buy an AMD card again just because you got burned once with driver support for one application, that's fine. If you don't want to date again just because one of your exes snored, that's okay, too. You can live your whole life alone if you want to. You can also spend unreasonable amounts of money on an Nvidia Ada card, let me not stop you. ;)

SisyphusIf they don't compete here, they will never have anything to do with it. :laugh:

Did you read my whole post? There are other Nvidia cards out there that are much better deals than the 4080 or 4090!

AusWolfThe "AMD driver instability" argument is mostly used by people who have never touched an RDNA 2 card. I, personally, never had any such problem with any AMD card except for the 5700 XT. RDNA 1 drivers were truly a bit immature, but AMD has come a long way since then. The GCN era was awesome, too. It's mind-blowing to see the lengths people go to justify their bad choice of spending more on an Nvidia card than they should have.

The 6600 series are an amazing deal. People upgrading from a 1060-like card should look there, and not wait for an overpriced 4060, imo.

The driver instability argument is from people than know AMD since before RDNA2.
To get to their stable driver today, AMD basically wiped the slate clean with RDNA, throwing everything that came before it under the bus. As you have noted, even with a clean slate, the first RDNA iteration was still a bumpy ride. Today the drivers are much better, but you can't fault people that were bitten in the past for still having a bitter taste in their mouths.
I'm not denying that there aren't those that simply parrot "driver instability" simply because they don't like AMD, but let's not pretend that's all there is to it.

bugThe driver instability argument is from people than know AMD since before RDNA2.
To get to their stable driver today, AMD basically wiped the slate clean with RDNA, throwing everything that came before it under the bus. As you have noted, even with a clean slate, the first RDNA iteration was still a bumpy ride. Today the drivers are much better, but you can't fault people that were bitten in the past for still having a bitter taste in their mouths.
I'm not denying that there aren't those that simply parrot "driver instability" simply because they don't like AMD, but let's not pretend that's all there is to it.

I know AMD and ATi from the early 2000s, and never had any problem with drivers, except for the 5700 XT.

If people want to continue parroting "driver instability" instead of asking users and doing some research, that's their choice. I accept it, just don't agree with it. They should specify that they're too lazy to look into things before bringing up past issues that are totally irrelevant with today's products and misleading other people as a result.