NV 4060 / 4060TI reviews

[blckgrffn]

Eh, I remember enough of us being skeptical but either way, this isn't two years ago. But if you're playing current or older games the ram won't matter. Then again, if it doesn't matter that much why not buy one of the older cheaper 8GB cards that are two years old now? Frame generation I guess.

Few complained because no one could actually buy them. By the time the 3070 actually launched it was full ETH & Covid. 8 GB was fine for mining. The size of the frame buffer when it related to gaming was largely academic.

But I've been skeptical of them being 8GB the whole time. If they had been for gamers and not miners, there would have been more feedback IMO, but we'll never know.

The 3070ti, which launched later was the one I was most critical off because... it launched later. I don't give a darn about it having GDDRX, as I've already said. It could have gotten a solid 20% more bandwidth just using higher tier GDDR6 AND had 16GB too. But no.

 

[Heartbreaker]

You seem to conflate everything under the Ultra umbrella, as if Ultra textures can be downgraded just as easily as Ultra Particle FX.

We've had multiple examples now where 8GB of VRAM leads to a different experience in games. Downgraded textures, missing textures, texture pop-in. The alternative is stutter or some other performance issues. Even Digital Foundry mentions the 4060 needs to drop textures to High in Spider Man Miles Morales once you aim for actual gaming sessions where VRAM allocation tends to vary:

Not sure how this is so damming. He's trying to run 4K with RT (with DLSS) at Very high textures, runs into problems turns the textures to high, and backs off RT a bit to solve his issues. That seems perfectly reasonable for a $300 GPU. 8GB GPUs are not useless because they can't run smoothly at 4K or Ultra + RT.

Last but not least, I don't understand your stance on the "worth" of the extra 8GB for $100. You don't find the base 8GB card acceptable either, yet you write an entire post disagreeing with me because reasons. During this post you make it clear the extra 8GB is not worth $100 because the benefit is marginal. And then a bit later you write this:

So when the 4060 Ti is $300, the extra $100 for another 8Gb would "work for you"?! What happened to buying the 16GB Mac Mini Instead?

I don't have a problem with $100 for 8GB more VRAM, but I wouldn't spend it for gaming, because a couple of setting can easily fix it for gaming.

I do have a problem with the base price of the 4060 Ti.

$400 4060 TI 8GB, and $500 4060 TI 16GB are both unacceptable. It's not the extra $100 that makes the 16GB 4060 Ti 16GB unacceptable.

$300 4060 TI 8GB, and $400 4060 TI 16GB are both acceptable.

I would still buy a low end $300 or less 8GB card for gaming, but I would not buy a 8GB Mac Mini because that is too little RAM for a whole computer IMO.

Exactly where are you seeing a contradiction.

 

[coercitiv]

Exactly where are you seeing a contradiction.

When the 16GB model is 25% more expensive, you find it unacceptable. When the 16GB model is 33% more expensive, you find it acceptable. In both situations the absolute price delta is $100.

Somehow the marginal benefit of 16GB is worth $100 when the base model becomes cheaper. That's a contradiction.

 

[Saylick]

When the 16GB model is 25% more expensive, you find it unacceptable. When the 16GB model is 33% more expensive, you find it acceptable. In both situations the absolute price delta is $100.

Somehow the marginal benefit of 16GB is worth $100 when the base model becomes cheaper. That's a contradiction.

FWIW, I don't think there's an issue with @Heartbreaker's line of thinking, even if the relative price increase for the extra 8GB is higher if both models are $100 cheaper.

Acceptable vs. unacceptable doesn't have to be purely black and white. There are varying degrees of acceptable. If the 4060 Ti 8GB was $50 and the 4060 Ti 16GB was $150, I don't think many people would complain about it. Yeah, sure, the 8GB model is far more bang for buck, but at $150 for the 16GB model, it would be a worthwhile price if you wanted to buy the fancier version of the two and be done with it.

Said another way, if I offered someone their choice of a $50 4060 Ti 8GB or a $150 4060 Ti 16GB (assuming they could only buy one), I don't think anyone on these forums should or would say that getting a heavily discounted 4060 Ti 16GB for $150 is a bad decision given today's prices.

 

[coercitiv]

Said another way, if I offered someone their choice of a $50 4060 Ti 8GB or a $150 4060 Ti 16GB (assuming they could only buy one), I don't think anyone on these forums should or would say that getting a heavily discounted 4060 Ti 16GB for $150 is a bad decision given today's prices.

$0 4060Ti or $100 4060Ti 16GB. You only get to choose one. If you're willing to spend $100 to upgrade from a FREE card, there better be a solid reason there. And if there's a solid reason there, that reason applies for real world pricing as well.

 

[Heartbreaker]

When the 16GB model is 25% more expensive, you find it unacceptable. When the 16GB model is 33% more expensive, you find it acceptable. In both situations the absolute price delta is $100.

Somehow the marginal benefit of 16GB is worth $100 when the base model becomes cheaper. That's a contradiction.

That looks you are being willfully obtuse. I was VERY clear.

I consider $100 for 8GB VRAM GPU upgrade an acceptable price. Period. The price of the upgrade is NOT the issue.

But if the base price of the card is bad, it's still going to be bad after the paying for an upgrade.

Likewise if the base price of the card is good, it's still going to be good after the upgrade.

Because, repeat until you understand: The price of the upgrade is NOT the issue.

The original price is the problem with these cards.

 

[Saylick]

$0 4060Ti or $100 4060Ti 16GB. You only get to choose one. If you're willing to spend $100 to upgrade from a FREE card, there better be a solid reason there. And if there's a solid reason there, that reason applies for real world pricing as well.

I agree. In that hypothetical situation, paying the $100 needs to be worthwhile over the free card for one to choose it, but if there is a valid reason for it, no one here should give that person crap for paying $100 when they could've had a free card (assuming the idea wasn't that you simply resell it for profit).

 

[coercitiv]

That looks you are being willfully obtuse. I was VERY clear.

I consider $100 for 8GB VRAM GPU upgrade an acceptable price. Period. The price of the upgrade is NOT the issue.

[...]

Because, repeat until you understand: The price of the upgrade is NOT the issue.

The original price is the problem with these cards.

Sigh.

For gaming I wouldn't pay $100 for that extra 8GB, and it's not because $100 is that overpriced for 8GB added in a product, but because it makes little difference.

You made it abundantly clear the extra 8GB is not worth the money because it makes little difference. The price of the base cards does not change the performance and IQ improvement the extra 8GB provides, the perceived benefit is always the same.

As for being willfully obtuse, remember all this started with me saying TPU will become aware of testing for IQ problems in certain games based on the work of other outlets. This prompted a reaction from you in which YOU decided to talk about optimized settings in games and the $100 price of the extra 8GB. I made zero reference to either of these talking points before you started, you basically used me as a prop to steer the discussion. Now you're annoyed I'm paying attention and calling out inconsistencies.

One last time: you can call the $400 4060 Ti and the $500 upgrade unacceptable, while both cards become acceptable when base card drops to $300. That's fine, we have common ground here. What doesn't work though, is saying at the same time that the extra 8GB makes little difference. This invalidates the idea that the base card price is the problem, because it makes the 16GB not worth at any price point as long as the base card is ~$100 cheaper.

If the 16GB card is a (theoretical) good buy at $400 then all the discussion about cherry picking game settings in reviews to make the 8GB card look bad is moot, you acknowledge the difference is worth your money.

 

[BFG10K]

Behold, teachings from the ancient scripture of "reasons why 8GB is never NV's fault!"

HWUB is essentially cherry picking...

It's the reviewer's fault.
It's a conspiracy against nVidia.

...the most VRAM hungry games...

It's the game's fault.
It's the developer's fault.
Wrong games were tested.

...then running at the Ultra settings, plus RT (where available)...

Wrong settings were tested.

...looking for situations that will break 8GB.

Wrong area was tested.

Since Ultra Settings aren't needed, you can just turn the settings down a notch.

It's completely acceptable for 3070/3070TI/3080 owners to turn down settings while 3060/6700XT users have no issue.
It's completely acceptable to disable ray tracing on nVidia cards while AMD users have no issue.

All hail the glorious 8GB gospel, may it live another nine years!!!

Most people agree that Ultra setting are dumb, including HWUB:

Yeah, most people do agree ray tracing is dumb, except NV who parades it as the raison d'etre for their GPUs. There are instances where AMD beats them at their own game because NV ships nine year old VRAM capacities @ $400.

Oh, you mean the other(tm) ultra settings like textures? Except textures provide one of the highest IQ benefits, they work on any graphics card, and there's zero performance penalty with enough VRAM.

 

[Heartbreaker]

One last time: you can call the $400 4060 Ti and the $500 upgrade unacceptable, while both cards become acceptable when base card drops to $300. That's fine, we have common ground here. What doesn't work though, is saying at the same time that the extra 8GB makes little difference. This invalidates the idea that the base card price is the problem, because it makes the 16GB not worth at any price point as long as the base card is ~$100 cheaper.

If the 16GB card is a (theoretical) good buy at $400 then all the discussion about cherry picking game settings in reviews to make the 8GB card look bad is moot, you acknowledge the difference is worth your money.

Saying a price is acceptable. Doesn't mean I must run out and buy it, or even choose it over another acceptable option if both were available. Saying the $300 8GB and $400 16GB cards are both acceptable doesn't mean I have to choose the 16GB version. I can still chose the 8GB version, or I could choose the 16GB, that's what "both" acceptable means.

My point all along is that the price of the upgrade isn't the problem, with these cards is not the price of the upgrade, it's the base price of the cards. Many people have made a big deal out of the $100 upgrade cost like it's a rip off, and I was making it clear that I don't think that is the issue.

8GB doesn't make $100 worth of difference for gaming for me(there may be non gaming use cases that make it worth it...) , because you can just easily adjust settings and not have issues. I've basically tweaked settings in every 3d game I've ever played, so it's not like this is some kind of major egregious problem, and as already discussed by HWUB, Playing at Ultra is just dumb, yet they turn around and act like the sky is falling if you can play Ultra+RT on every game. Show me the game that can't be fixed with a couple of settings, or you know just starting at high instead of Ultra, as suggested by HWUB in their previous video. IMO, for lower end cards ($300 and below), this molehill has been made into a mountain. Just adjust the settings.

Right now, I'm kind of in limbo because all choices are kind of unpalatable. I would have preferred the 4060 Ti had been 12GB with a wider bus from the start( probably would outperform 3070), and have been better for both gaming and potentially other uses like Machine Learning. A better all rounder, to make the $400 price acceptable.

As it stands the 4060 Ti pricing is just too high for what you get, 4060 is lower on the performance tier than I really want to go, and 4070 is very nice but really beyond what I wanted to pay... (I've never paid more than $300 CAD for a GPU - 4070 is $800 CAD).

I have all these cards and more in an Amazon wish list that I check daily looking for something to make one of the options more palatable.

I had hoped for AMDs N32 cards to be here and competitive by now, to force NVidia to lower prices... 😈

 

[dlerious]

I have all these cards and more in an Amazon wish list that I check daily looking for something to make one of the options more palatable.

I had hoped for AMDs N32 cards to be here and competitive by now, to force NVidia to lower prices... 😈

Nvidia is laughing all the way to the bank. They're not going to lower prices, just look at the 7900XTX/4080 and 7900XT/4070Ti . Unless you're willing to hit them in the pocketbook by buying AMD, they're not going to change.

 

[Heartbreaker]

Nvidia is laughing all the way to the bank. They're not going to lower prices, just look at the 7900XTX/4080 and 7900XT/4070Ti . Unless you're willing to hit them in the pocketbook by buying AMD, they're not going to change.

Nah it doesn't have to be me that buys AMD, just more of you guys buying AMD.... Go AMD...

 

[SteveGrabowski]

nm

 

[DAPUNISHER]

The 4060 has exceptional performance per watt. I see that as enough to carve out its niche. The TIE models not so much.

 

[Thunder 57]

I had hoped for AMDs N32 cards to be here and competitive by now, to force NVidia to lower prices... 😈

And this is why Nvidia has a near monopoly in the GPU market. And as stated above, they will not lower prices. Until they start losing market share, they will not lower prices.

 

[Heartbreaker]

And this is why Nvidia has a near monopoly in the GPU market. And as stated above, they will not lower prices. Until they start losing market share, they will not lower prices.

If AMD has something very competitive, NVidia will move before they actually start losing market share.

IIRC, they timed the 2060 Super and 2070 Super to launch the same week AMD launched 5700/XT.

 

[Saylick]

And this is why Nvidia has a near monopoly in the GPU market. And as stated above, they will not lower prices. Until they start losing market share, they will not lower prices.

And it's because of that same mentality, that people treat AMD as a means to cheaper Nvidia prices, that AMD now refuses to compete by lowering margins.

 

[DAPUNISHER]

And it's because of that same mentality, that people treat AMD as a means to cheaper Nvidia prices, that AMD now refuses to compete by lowering margins.

I have been watching the current TPU poll on GPU ownership since it was under 10K votes. We are up to almost 53K votes and AMD has held a steady 33% share throughout. I would be willing to wager it won't go under 30% even if they get 100K responses.

Since OEM, S.I., and Asia, is where they are undoubtedly getting slaughtered. AMD may as well make as much as they can off of their customer base. Which evidently is primarily the enthusiast market in N.A. and Europe.

 

[Saylick]

I have been watching the current TPU poll on GPU ownership since it was under 10K votes. We are up to almost 53K votes and AMD has held a steady 33% share throughout. I would be willing to wager it won't go under 30% even if they get 100K responses.

Since OEM, S.I., and Asia, is where they are undoubtedly getting slaughtered. AMD may as well make as much as they can off of their customer base. Which evidently is primarily the enthusiast market in N.A. and Europe.

Yeah, if I were Lisa Su, I would pursue the same strategy as well. Businesses' number one goal is to make as much profit as they can given a finite amount of resources. With finite resources, you allocate efforts towards the most profitable endeavors first and work your way down from there.

Expanding GPU market share by lowering margins in the short term to hopefully gain more profit via raising margins in the long term is not even remotely close to being low hanging fruit for them since they have other more profitable markets to pursue.

It's not until you get to 2005-2015 era Intel where you're so dominant (>90% market share) in your more profitable market segments that you start diversifying your revenue sources because investors and shareholders require you to keep growing quarter after quarter.

Edit: Now that I think about it some more, the current consumer GPU space can be described very accurately as being in a Nash equilibrium. AMD and Nvidia play out their hand knowing how the other will react, and the current best course of action for both is to not deviate from whatever they're currently doing.

 

[DAPUNISHER]

Edit: Now that I think about it some more, the current consumer GPU space can be described very accurately as being in a Nash equilibrium. AMD and Nvidia play out their hand knowing how the other will react, and the current best course of action for both is to not deviate from whatever they're currently doing.

That makes a lot of sense to me; I dig it. Hopefully Intel (Carol) is disrupting that. Forcing both to evaluate what the best response is when including C.

 

[dlerious]

And it's because of that same mentality, that people treat AMD as a means to cheaper Nvidia prices, that AMD now refuses to compete by lowering margins.

How low do they have to go? The 6950XT is currently $350 below MSRP, the 6800XT, 6800 and 6700XT are over $100 under. 7900XT are $100-$150 under MSRP.

I've been waiting to see price/perf for the 7800XT - and getting a little upset it hasn't been released yet. AMD isn't going to be able to charge too much with the 7900XT at $750 currently and maybe less by the time the 7800XT is released (Sept. last I heard). Maybe I'll be able to grab a 7900XT for under $700 on Black Friday.

 

[Aapje]

I've been waiting to see price/perf for the 7800XT - and getting a little upset it hasn't been released yet. AMD isn't going to be able to charge too much with the 7900XT at $750 currently and maybe less by the time the 7800XT is released (Sept. last I heard). Maybe I'll be able to grab a 7900XT for under $700 on Black Friday.

I wouldn't get your hopes up too much. RDNA2 is at pretty good prices due to oversupply, but I don't see them going as low for RDNA3.

 

[igor_kavinski]

I wouldn't get your hopes up too much. RDNA2 is at pretty good prices due to oversupply, but I don't see them going as low for RDNA3.

Yup. They would rather restrict the supply of RDNA3 cards than to dilute its value by flooding the market with it.

 

[kondziowy]

I wouldn't get your hopes up too much. RDNA2 is at pretty good prices due to oversupply

yeeeaa but... RDNA2 is priced perfectly between RDNA3 RX7600 <-> RX7900XT. That means, RDNA2 cards still carry price premium of new generation products, and have room to go down. I think we will see real price cuts after 7700/7800 cards arrive. Right now we are seeing RX6800/6900/6950xt at their full premium price. Tell me I'm wrong and why (pandemic MSRP prices were ludicrous and mean nothing to me)

 

[Heartbreaker]

yeeeaa but... RDNA2 is priced perfectly between RDNA3 RX7600 <-> RX7900XT. That means, RDNA2 cards still carry price premium of new generation products, and have room to go down. I think we will see real price cuts after 7700/7800 cards arrive. Right now we are seeing RX6800/6900/6950xt at their full premium price. Tell me I'm wrong and why (pandemic MSRP prices were ludicrous and mean nothing to me)

Where else would it be priced but between them? They aren't going to price a 6800 XT under a RX 7600.

I do think supplies of RDNA2 are dwindling. On shop AMD.com in Canada, the only RDNA2 they list are 6750 and 6950, and the 6950 is out of stock.

By the time 7700/7800 arrive, there really won't be much competing RDNA2 stock left to cut more.