• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

NVIDIA GeForce RTX 5060 8 GB

This isn't the card for you if you play at 4K. You are going to absolutely want something with more VRAM. The 9060 XT is going to get you double the VRAM for only $50 more plus it has more raw horsepower to boot.

Assuming we can believe AMD in house benchmarks which I'm not holding my breath, I'm also not holding my breath that it'll actually be 350 usd becuase if it doesn't suck it'll get scalped by retailer's #2025

That is the silver lining with this GPU its so bad its not hard to get close to msrp.
 
I hate the fact that it performs better than the B580, and yet still has 8GB of VRAM.
 
This isn't the card for you if you play at 4K. You are going to absolutely want something with more VRAM. The 9060 XT is going to get you double the VRAM for only $50 more plus it has more raw horsepower to boot.
@MatthewGTX

Yeah, as with his suggestion above, either the upcoming RX 9060 XT or the RX 9070 non-XT would be your best bet if you are always running at 2160p resolution and are sticking with Linux.
 
I hate the fact that it performs better than the B580, and yet still has 8GB of VRAM.

That's the thing though the B580 has been scalped by retailers above its MSRP if this card was a 12GB card from Nvidia it wouldn't be 299 even if that was the official price so while I think this card is a joke it being a joke is the only reason it'll probably stick close to it's msrp and for 2060 6GB or older users that might be a win.
 
The 5060 Super 12 GB (assuming it releases with the 3 Gb chip config on the same bus width) is what the 5060 should be from the get-go.

Sure, an 8 buffer on a gaming GPU would suffice for eSports and light gaming... But at that price, and the relative raw power of the chip... 8 GB restricts it to a small order of tasks, 12 GB isn't great, but it will allow the card to stretch it's usability, so that's still better than what released.
 
But his argument still makes sense because that just means NVIDIA should've kept the memory bus size the same alignment for the entire series (e.g. 60-series = 192-bit, 70-series = 256-bit, 80-series = 256-bit+, etc.) instead of regressing. They would've been able to keep the ~$330 pricing with minimal public backlash and we would have a minimum of 12GB standard for the lower-middle class video cards now.

That and the performance minimum would've been higher today.
Bigger bus is expensive, though. From a 3060 Launch Day review (https://www.techpowerup.com/review/palit-geforce-rtx-3060-dual-oc/39.html): "Just like on the RTX 3060 Ti, NVIDIA uses GDDR6 memory, this time with a bus width of 192-bit (RTX 3060 Ti is 256-bit). Having to route just 192 memory chip signals instead of 256 does make a difference, the PCB design is much less complex, which means it's cheaper to manufacture. The same goes for the GPU chip, which can make do with a simpler memory controller." The cost of more VRAM isn't the chips, it's the card design.

Nvidia has always favored memory bandwidth over capacity, which seems to have worked pretty well for them compared to AMD who's done the opposite. And now Nvidia can even do both with 3GB chips on the Super refresh.

As for setting a standard, the standard is 8GB. That's what people buy. The 4060 is as popular as the 3060. All that Nvidia would do by setting 12GB as the "standard" is losing money by refusing to make cheaper GPUs for people willing to buy them. This is basic economics.
 
That's the thing though the B580 has been scalped by retailers above its MSRP if this card was a 12GB card from Nvidia it wouldn't be 299 even if that was the official price so while I think this card is a joke it being a joke is the only reason it'll probably stick close to it's msrp and for 2060 6GB or older users that might be a win.
Why did Nvidia give 12GB on the relatively recent 3060 then?
 
Why did Nvidia give 12GB on the relatively recent 3060 then?
They had more than one braincell back then...

They knew it's performance would outpace the 6GB config. That would've been the limiting factor.
 
Why did Nvidia give 12GB on the relatively recent 3060 then?
They designed it to have 6GB. That's why the 3070 has 8GB and the 3080 has 10GB. At the very last minute, they decided 6GB wouldn't be enough for ray tracing, so they had to do 12GB instead (that, or cut down the bus to 128-bits and give it 8GB, but that would cost $ and cost significant performance due to lost memory bandwidth).
 
Becuase the alternative was 6GB due to the bus width.
Well, there was also an 8GB variant of the 3060. It was rare but did exist.
 
Well, there was also an 8GB variant of the 3060. It was rare but did exist.
There were 2 models, a GA106 and a GA104 die

Only the GA106 model was produced, with a total of 15 AIB variants

It sucked.
 
Well, there was also an 8GB variant of the 3060. It was rare but did exist.

It came later because they wanted to clear stock that they over ordered due to the pandemic/crypto boom.

They made so many to this day you can still buy the 3060 12GB
 
Bigger bus is expensive, though. From a 3060 Launch Day review (https://www.techpowerup.com/review/palit-geforce-rtx-3060-dual-oc/39.html): "Just like on the RTX 3060 Ti, NVIDIA uses GDDR6 memory, this time with a bus width of 192-bit (RTX 3060 Ti is 256-bit). Having to route just 192 memory chip signals instead of 256 does make a difference, the PCB design is much less complex, which means it's cheaper to manufacture. The same goes for the GPU chip, which can make do with a simpler memory controller." The cost of more VRAM isn't the chips, it's the card design.

Nvidia has always favored memory bandwidth over capacity, which seems to have worked pretty well for them compared to AMD who's done the opposite. And now Nvidia can even do both with 3GB chips on the Super refresh.

As for setting a standard, the standard is 8GB. That's what people buy. The 4060 is as popular as the 3060. All that Nvidia would do by setting 12GB as the "standard" is losing money by refusing to make cheaper GPUs for people willing to buy them. This is basic economics.
By this point, then they can just use their previous xx50 lineup and make an 8 GB 128-bit RTX 5050 and sell it at $250 to $280, which would hit that price point. They could have kept the xx60 at 192-bit and continue selling it at around $330.

NVIDIA is currently the market leader, so they should be the ones to keep pushing the envelope. While it's good that the performance is still "better" than the previous generation due to architectural improvements, they should've included increasing memory capacity, which would be in-line with the RT features they are marketing. As we can see with the general benchmarks the RTX 5060 suffers when RT is enabled in some games due to lack of VRAM, even at 1080p.
 
And since it's the best price to performance card as of now it's not THAT bad. Sure, 8 GB for 300 bucks is horrible. Sure, what NVIDIA did to the reviewers is unforgivable. But the SKU exists and is available. And nothing, I repeat, NOTHING dares to compete with it.

Best price to performance at 1080p and if you ignore the fact that you could have gotten more for your money two years ago, that you can't run RT or PT, that you are going to have to lower graphics settings, and that the card won't last long.

Assuming we can believe AMD in house benchmarks which I'm not holding my breath, I'm also not holding my breath that it'll actually be 350 usd becuase if it doesn't suck it'll get scalped by retailer's #2025

That is the silver lining with this GPU its so bad its not hard to get close to msrp.

I haven't bothered to look at AMD's in house benchmarks, HWUB has already hinted in regards to the card's price to performance.

The pricing argument applies to both AMD and Nvidia so it's a moot one unless you think there'll be a notable difference. There might be on the RX 9060 XT, given 16GB of VRAM at $350 is likely to be popular.

As for setting a standard, the standard is 8GB. That's what people buy. The 4060 is as popular as the 3060.


People buy 8GB because that's what's offered in the price bracket most buy in at. Not because they get excited for and want 8GB.

All that Nvidia would do by setting 12GB as the "standard" is losing money by refusing to make cheaper GPUs for people willing to buy them. This is basic economics.

Why give customers anything at all? It's all costing Nvidia money at the end of the day.

You are arguing like an Nvidia shareholder and not a gamer. The benefit to having say 12 or 16GB to customers is obvious, it would allow this card to run RT / PT, it would improve 1% and .1% lows, it would improve performance in a variety of games, it would enable higher resolution gaming, and it would increase the longevity of the card.

Have you been so abused by capitalism that you forget that the customer should be considered when making a product?
 
By this point, then they can just use their previous xx50 lineup and make an 8 GB 128-bit RTX 5050 and sell it at $250 to $280, which would hit that price point. They could have kept the xx60 at 192-bit and continue selling it at around $330.

NVIDIA is currently the market leader, so they should be the ones to keep pushing the envelope. While it's good that the performance is still "better" than the previous generation due to architectural improvements, they should've included increasing memory capacity, which would be in-line with the RT features they are marketing. As we can see with the general benchmarks the RTX 5060 suffers when RT is enabled in some games due to lack of VRAM, even at 1080p.

100%

But in 2025 99% of their focus is on the AI market I honestly believe they fully believe releasing this gpu at 299 is a charity to gamers.
 
charity to gamers.
"If it can run MS Paint, it's good enough for gamers"
"The good stuff will bring more cash from Ai farm operators"
 
Have you been so abused by capitalism that you forget that the customer should be considered when making a product?
The customer gets a cheaper GPU than a 12 or 16GB GPU would cost. That's the benefit to the customer. It's entirely up to the customer to decide if it's worth it for them, Jensen isn't breaking into people's houses and forcing them to buy 8GB GPUs at gunpoint. People buy it because it's worth the money to them, simple as that. If it's not worth it, they won't buy it.

Maybe in 2025, more people decide 8GB for $300 isn't worth it for them. I am 100% open to that being a possibility. If that happens, Nvidia will lose money and and not sell GPUs. So Nvidia has a very, very strong incentive to sell what people want. Either way, I'm not bothered, because (a) it's not my money, I just buy what is worth it to me, and (b) I'm pretty sure a $3 trillion company has put at least a little bit of thought into what the market is willing to pay for GPUs (i.e. they've thought about what people want to buy).
 
The customer gets a cheaper GPU than a 12 or 16GB GPU would cost. That's the benefit to the customer. It's entirely up to the customer to decide if it's worth it for them, Jensen isn't breaking into people's houses and forcing them to buy 8GB GPUs at gunpoint. People buy it because it's worth the money to them, simple as that. If it's not worth it, they won't buy it.

Maybe in 2025, more people decide 8GB for $300 isn't worth it for them. I am 100% open to that being a possibility. If that happens, Nvidia will lose money and and not sell GPUs. So Nvidia has a very, very strong incentive to sell what people want. Either way, I'm not bothered, because (a) it's not my money, I just buy what is worth it to me, and (b) I'm pretty sure a $3 trillion company has put at least a little bit of thought into what the market is willing to pay for GPUs (i.e. they've thought about what people want to buy).
Actually, yes.

If others think the product is worth it to them (whether or not it's legitimately, not a great product) then it's worth the price

More informed people know what the bigger picture is. Problem is more informed people are less common.

We can't change people's minds if they're making a bad choice because in their head it seems good

Bet if I was a casual gamer hearing the 5070 was a 4090 but cheaper, and I didn't know anything about anything with no knowledgeable opinions to hear, I'd buy it and be happy.
 
100%

But in 2025 99% of their focus is on the AI market I honestly believe they fully believe releasing this gpu at 299 is a charity to gamers.
Well I'll disagree that its not charity for sure, but at $300 (and yes I'm including today's inflation too) this RTX 5060 should be at least 12GB of VRAM.

Or at least the idea that RTX 5060 = 8 GB and the RTX 5060 Ti = 16GB only. Having the RTX 5060 and the RTX 5060 Ti 8GB still existing together is the stupid part. $300 product along with a $380 so close to it is just saturation. The RTX 5060 Ti 16GB being MSRP'd at $430 (unfortunately $480 ugh) is not that bad of a price.
 
Well I'll disagree that its not charity for sure, but at $300 (and yes I'm including today's inflation too) this RTX 5060 should be at least 12GB of VRAM.

Or at least the idea that RTX 5060 = 8 GB and the RTX 5060 Ti = 16GB only. Having the RTX 5060 and the RTX 5060 Ti 8GB still existing together is the stupid part. $300 product along with a $380 so close to it is just saturation. The RTX 5060 Ti 16GB being MSRP'd at $430 (unfortunately $480 ugh) is not that bad of a price.

100% agree even if that meant 329 or at least giving the consumers the option with having both varients like what AMD is doing with the 9060XT.

Actually, yes.

If others think the product is worth it to them (whether or not it's legitimately, not a great product) then it's worth the price

More informed people know what the bigger picture is. Problem is more informed people are less common.

We can't change people's minds if they're making a bad choice because in their head it seems good

Bet if I was a casual gamer hearing the 5070 was a 4090 but cheaper, and I didn't know anything about anything with no knowledgeable opinions to hear, I'd buy it and be happy.

The problem is the majority of buyers are going to be people buying prebuilts who have no idea what they are actually getting only that it says Nvidia on the box so it must be good.

I think people in the diy market for the most part know better unless they can't spend a dime over 300 and need a gpu becuase their's died etc.
 
"Excellent price/performance" "$300 low price point"
I don't understand how this card is supposed to be excellent price to performance when it doesn't beat out the previous gen 4060Ti 8GB.
The card would be better price to performance at $200-250, not at $300, and the 5060 will probably remain close to MSRP because of how mediocre it is.
 
Last edited:
Back
Top