Wednesday, May 6th 2026

NVIDIA GeForce 10-series "Pascal" Generation Turns 10

NVIDIA GeForce GTX 10-series graphics cards turn 10 today. In May 2016, NVIDIA debuted the GeForce GTX 1080, powered by the "Pascal" graphics architecture on the TSMC 16 nm FinFET foundry node. This would go on to be the final generation of NVIDIA GeForce GPUs to truly comply with Moore's Law. To illustrate, the GeForce GTX 1080, a predecessor to the current RTX 5080, needed a single 150-Watt 8-pin power connector, and could play AAA games of its time at 4K Ultra HD with the game being rendered at native resolution (no upscaling). "Pascal" established phenomenal standards of performance-per-Watt over the GTX 9-series "Maxwell" generation.
Add your own comment

124 Comments on NVIDIA GeForce 10-series "Pascal" Generation Turns 10

#51
Tyler-98-W68
KritYes you can undervolt RTX 5090 but it will never be no the RTX 5070 Ti or RX 9070 XT level noise/heat wise. I'm talking about AIR cooling closed pc case. If fans will no spin faster RTX 5090 will be toasted and whole pc will be on higher temps. You pay a lot to get what worse noise levels!


Not happy about that at all but i can manage my RX 9070 XT @273watts to run at constant @30dBA which is quiet enough under load.
So the RTX PRO 6000 Blackwell Max-Q version which is 300w, is nearly twice as fast as a 5070ti but takes the same amount of power is not as efficient?
Posted on Reply
#52
Krit
Tyler-98-W68So the RTX PRO 6000 Blackwell Max-Q version which is 300w, is nearly twice as fast as a 5070ti but takes the same amount of power is not as efficient?
Very efficient 10000€ with low end turbine cooling.
Posted on Reply
#53
Tyler-98-W68
KritVery efficient 10000€ with low end turbine cooling.
So your only measure of efficiency is cooling? Heat is heat, so we just ignore the GPU with twice the performance producing the same amount of heat?


Half the power but only lose 20% performance?

That's super terrible, also the cooling is mentioned as well...
Posted on Reply
#54
Krit
Tyler-98-W68So your only measure of efficiency is cooling? Heat is heat, so we just ignore the GPU with twice the performance producing the same amount of heat?
Turbine is turbine no silence!
Posted on Reply
#55
Tyler-98-W68
KritTurbine is turbine no silence!
So again you are ignoring the basis of how efficiency is measured which is power consumption, makes sense.

so 62c with the fan at 37% is bad?
Posted on Reply
#56
Krit
Tyler-98-W68So again you are ignoring the basis of how efficiency is measured which is power consumption, makes sense.

so 62c with the fan at 37% is bad?
dBA numbers / rpms.
Posted on Reply
#57
Tyler-98-W68
KritdBA numbers / rpms.
So heat and power consumption has nothing to do with efficiency according to you?
Yes you can undervolt RTX 5090 but it will never be no the RTX 5070 Ti or RX 9070 XT level noise/heat wise.
So 300w on a 5070Ti is a different kind of heat vs 300w on a 6000 Max-Q?
Posted on Reply
#58
Krit
Tyler-98-W68So heat and power consumption has nothing to do with efficiency according to you?


So 300w on a 5070Ti is a different kind of heat vs 300w on a 6000 Max-Q?
How do you think 300w gpu with turbine will be quiet? At that cost what a shit cooling.
Posted on Reply
#59
Warigator
between 1996 and 2006 there was in the U.S. 28.5% inflation (so $299 from 1996 turned into $384.2)
between 2006 and 2016 there was in the U.S. 19.1% inflation (so $299 from 1996 turned into $457.4)
between 2016 and 2026 there was in the U.S. 37.6% inflation (so $299 from 1996 turned into $629.3)

so today, 1996 Voodoo Graphics equivalent would be the cheapest model of RTX 5070 or the most expensive model of RTX 5060 Ti
Voodoo Graphics for $299 had 4 MB, but $399 version had 8 MB and $599 version had 16 MB
5060 Ti 16GB has four thousand times more memory than Voodoo Graphics 4MB thirty years ago
Posted on Reply
#60
Tyler-98-W68
KritHow do you think 300w gpu with turbine will be quiet? At that cost what a shit cooling.
Where exactly is it shit cooling?

62c and 100% gpu load and 37% duty cycle on the fan?

Why did you ignore my question?

300w on a 5070ti is somehow different and better than 300w on a different gpu?

Don't get too tired from the mental gymnastics here
Posted on Reply
#61
Krit
Tyler-98-W68Where exactly is it shit cooling?

62c and 100% gpu load and 37% duty cycle on the fan?

Why did you ignore my question?

300w on a 5070ti is somehow different and better than 300w on a different gpu?

Don't get too tired from the mental gymnastics here
It's about noise that gpu will shit herself when compared to a proper x13 less expensive cooling let's say on RX 9070 XT. 50 to 55 dBA are some shocking numbers.
Posted on Reply
#62
Tyler-98-W68
KritIt's about noise that gpu will shit herself when compared to a proper x13 less expensive cooling let's say on RX 9070 XT.
So the GPU will shit itself? You know this how? Am I talking to a 12yr old amd fanboy? Who uses that kind of argument?

So again I'll reiterate, what exactly about the cooling system is bad?

The card is not noisy, you made claim (a completely baseless one)

Then when presented with facts you move the goal posts and state efficiency has nothing to do with power and heat.

Price is irrelevant, you wanted to talk about efficiency, and I pointed out precisely an extremely efficient GPU on the same level as a 5090, but by your measure now it's not efficient because it doesn't fit your narrative.

Maybe some people don't want a huge 3-4 slot cooler which obviously has better cooling capacity.

You keep talking about noise, but where is everyone complaining about the noise in the GPU I have mentioned?

You do realize blower style coolers aren't all noisy? They also used to be common on consumer gpus.

So you used AI and not any real world examples? Where did AI say it was going to "shit itself"
KritIt's about noise that gpu will shit herself when compared to a proper x13 less expensive cooling let's say on RX 9070 XT. 50 to 55 dBA are some shocking numbers.
So you used AI which was linked up to LinkedIn???

You got any real world tests? Or just relying on AI slop?
Posted on Reply
#63
Rulumi
AusWolfExactly. So what's preventing game devs from using such tech on machines that support it?

I can hardly recall a generation when current consoles were based on cutting edge tech.
That most developers of bigger scale games try to focus and target the most selling console as a baseline, which is why a lot of these things has got little use until recently even if using some don't impact performance negatively in hardware that implements the new features.
WarigatorDoesn't PS5 have RDNA2? At least that's what I heard and read on www. Maybe RDNA 1.5? Certainly not exactly the same RDNA that's in 5700 XT. PS5 GPU is more like desktop PC's 6600 XT.
That's what SIE's PR and marketing said at some point, but it was all just Sony Interactive Entertainmemt marketing. The PS5 Oberon chip is GFX 1013, which is mostly just RDNA 1 µArch with added RT from RDNA 2, but outside the added RT hardware and instructions, support of everything else mostly stays like how it was in RDNA 1 µArch. Could be called RDNA 1.5 and that's what s lot of people use, however it's not an official term.
WarigatorBase PS5 has a custom RDNA 2-based GPU, but not a full desktop/Xbox-style RDNA 2 implementation in terms of features like Variable Rate Shading, Mesh Shaders and Sampler Feedback. Calling it RDNA 1 is certainly going too far. Calling it custom RDNA 2-based, but not feature-complete in the same PC/Xbox/DX12U sense is much closer to the truth.

Also, I would not call VRS, Mesh Shaders and Sampler Feedback “newer instructions”. They are API/hardware features, not just GPU instructions. The distinction matters, because console GPUs have always been custom designs, and Sony does not expose the hardware through DirectX in the same way as Xbox or PC.
It's not going too far, it literally is RDNA 1 with RDNA 2 RT parts added to it, but without changing the rest of the µArch from RDNA 1 specifications, it has been already confirmed by engineers and now directly by the hardware itself. It' easy to directly check and document nowdays thanks to AMD BC-250 and now the ps5-linux project. It is true however that in the softwre part SIE's OS and own graphical APIs are well optimized for the hardware, yes.

gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33116

elektricm.github.io/amd-bc250-docs/drivers/radv/#bazzite
www.phoronix.com/news/AMD-RADV-PS5-BC-250

www.phoronix.com/news/AMD-More-Cyan-Skillfish-2025
Posted on Reply
#64
Visible Noise
JustBenchingPeople don't really care about power
People care about power when it makes their team look good.
Posted on Reply
#65
AusWolf
lexluthermiesterThe 10 year anniversary of release not a good enough reason? Seems like a good reason to me..
Not to me, sorry. Like I said, Pascal isn't supported anymore, so there's nothing to celebrate. Why treat it differently than any other old architecture with no support?
JustBenchingSo surely, with these measly gains from the 4090 to the 5090, there must be a lot of competitors that have far surpassed the 5090 in performance. Right?
Apples to oranges. We're talking about efficiency, not peak performance.
JustBenchingPhysics have a wall. At some point, we hit that wall and performance stopped scaling. Do you think they can offer twice the performance of the 4090 at half the power but they decided not to for "reasons"? Like what are we even talking about here.
Let's say there's a wall. Let's say "next product" won't be any better than current ones. So why should I applaud and buy "next product" then?
JustBenchingThen why are you buying xx70 tier cards that draw over 300 watts? The 1070 was sitting at 140 watts, your 9070xt draws over 300...you voted with your wallet that you don't care about power draw, didnt' you?
Because that's the level x80 and x90 cards used to be at.
JustBenching300 watts is over twice of the pascal xx70 chips you are praising here. So clearly you don't seem to care about power since you bought a card with twice the power draw, right?

You can undervolt and tune the 5090 as well, moot point.
You can undervolt it, but you can't make the card physically smaller.
Posted on Reply
#66
lexluthermiester
AusWolfLike I said, Pascal isn't supported anymore, so there's nothing to celebrate.
Celebrating it's existence? It's like a birthday party, but for GPU's.
AusWolfWhy treat it differently than any other old architecture with no support?
Other than the fact that it was the single greatest jump in performance and features from a GPU line up, possibly ever?

I guess we disagree on this one too. It's all good, no worries.
Posted on Reply
#67
AusWolf
RulumiThat most developers of bigger scale games try to focus and target the most selling console as a baseline, which is why a lot of these things has got little use until recently even if using some don't impact performance negatively in hardware that implements the new features.
But current gen consoles have always been under PC tech in feature set. So how is the present situation any different leading to stagnation?
Posted on Reply
#68
ShrimpBrime
AusWolfNot to me, sorry. Like I said, Pascal isn't supported anymore, so there's nothing to celebrate. Why treat it differently than any other old architecture with no support?


Apples to oranges. We're talking about efficiency, not peak performance.


Let's say there's a wall. Let's say "next product" won't be any better than current ones. So why should I applaud and buy "next product" then?


Because that's the level x80 and x90 cards used to be at.


You can undervolt it, but you can't make the card physically smaller.
Because people really did like those cards back then. A bit faster than Maxwell and felt like you where getting what you pay for, unlike today.
lexluthermiesterOther than the fact that it was the single greatest jump in performance and features from a GPU line up, possibly ever?
Beat me to it, but this is what I'm saying also ^^ TY.

I still break out old hardware and enjoy them occasionally. Why not. Still can play most of the old school to even newer games, so it's not like a complete obsolete graphics card, aside from no more driver updates. NV is probably hoping by now we'd upgraded.

Posted on Reply
#69
AusWolf
lexluthermiesterCelebrating it's existence? It's like a birthday party, but for GPU's.

Other than the fact that it was the single greatest jump in performance and features from a GPU line up, possibly ever?

I guess we disagree on this on too. It's all good, no worries.
Why would I celebrate an object, a product?

I love my passive 1050 Ti to bits, but I'm not gonna throw a party to celebrate buying it because 1. It doesn't care, 2. My life isn't that sad.
ShrimpBrimeBecause people really did like those cards back then. A bit faster than Maxwell and felt like you where getting what you pay for, unlike today.
I still like Pascal a lot even today. But I'm not gonna celebrate an object. False idols and all.
Posted on Reply
#70
ShrimpBrime
AusWolfWhy would I celebrate an object, a product?

I love my passive 1050 Ti to bits, but I'm not gonna throw a party to celebrate buying it because 1. It doesn't care, 2. My life isn't that sad.


I still like Pascal a lot even today. But I'm not gonna celebrate an object. False idols and all.
haha, no, I don't think the OP and title of the thread was aiming for that. You must have picked that up from the gen pop conversations I think (I skimmed this one...)

I stuck with GTX 980 and a 980 TI for many many years. I'm old so got stuck on them old Valve engine games. HL/2/DM/Gary'sMod/ CS/CSS and CSGO. Frame rates went up every time I upgraded a CPU anyways haha. I miss those days (pre 2015 actually).
Posted on Reply
#71
JustBenching
AusWolfApples to oranges. We're talking about efficiency, not peak performance.
Not really? The post I quoted said that the 5090 has a massive power draw for only 35% performance increase, which is a weird thing to say.
AusWolfLet's say there's a wall. Let's say "next product" won't be any better than current ones. So why should I applaud and buy "next product" then?
Then don't, I don't know what exactly you are asking me.
AusWolfBecause that's the level x80 and x90 cards used to be at.
No? Some high end cards did draw 250w, some other drew a lot less. 5850 / 5870 were far below 200w of power. People just don't care about power efficiency, as I've said before the 4060 and 4060ti pushed efficiency but people said "it's not that much faster than previous gen". Well of course it isn't, it draws half the power, lol. How do you expect a huge jump in performance and a reduction in power draw at the same time? That cannot happen.
Posted on Reply
#72
wolf
Better Than Native
Congrats to a legendary generation, I still cherish my launch week GTX1080. Probably about time I busted it out and see how it fares today, which I bet is still rather good for something that has very much earned it's reputation in retirement.
Posted on Reply
#73
JustBenching
AusWolfYou can undervolt it, but you can't make the card physically smaller.
The 5090 FE is smaller than your average xx70 / 9070xt card.
Posted on Reply
#74
lexluthermiester
ShrimpBrimeBeat me to it, but this is what I'm saying also ^^ TY.
It really was a huge step up. Performance plus power efficiency plus feature upgrade and updates. Pascal is NVidia at it's finest. Don't get me wrong, I really dig on raytracing and I still say my RTX3080 is the best card I've ever spent money on.
AusWolfWhy would I celebrate an object, a product?
Because of the enjoyment it brings. Seriously?
AusWolfI love my passive 1050 Ti to bits, but I'm not gonna throw a party to celebrate buying it
Well yeah, but that's not what we're doing here. It's a news article. Or are you being intentionally contrary?
AusWolf1. It doesn't care, 2. My life isn't that sad.
Ok, you're being contrary... LOL!
Posted on Reply
#75
Dr. Dro
AusWolfNot to me, sorry. Like I said, Pascal isn't supported anymore, so there's nothing to celebrate. Why treat it differently than any other old architecture with no support?
At least professional is still being updated on a steady cadence, latest R580 U9 is from 2 weeks ago, but released same day as the latest 596.36 from R595. Though once R580 drops it's done that's pretty much it, it'll be in the same spot of Adrenalin 23.19 for GCN

wolfCongrats to a legendary generation, I still cherish my launch week GTX1080. Probably about time I busted it out and see how it fares today, which I bet is still rather good for something that has very much earned it's reputation in retirement.
FWIW, at 1080p, the only game I play regularly (as in every single day) my P6000 struggles with is Zenless Zone Zero. It really doesn't run amazing, although it does have the 1080p 60 FPS target beat. Roughly the same performance of a stock Titan Xp on tap, but with more memory to work with. Not that it matters, rendering performance is exhausted long before 24 GB is actually needed. Even Apex Legends actually seems to run at 1080p 120 just fine at highest settings, with minor dips here and there.
Posted on Reply
Add your own comment
Nov 17th, 2026 17:48 CST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts