Nvidia GeForce RTX 2080 Founders Edition review

https://www.techpowerup.com/reviews/NVIDIA/GeForce_RTX_2080_Ti_Founders_Edition/33.html
relative-performance_3840-2160.png
 
2080ti fail. 30% performance increase over 1080ti with a huge $1200 price tag.

I smell a price drop incoming and a lot of cancelled pre orders.

It's been a 30% increase almost every generation for the flagship card with a ~$700 price. Why this card is $1200 with the same performance increase is absurd. Nvidia can go F themselves.

The 2080 is essentially a 1080ti with less memory and Ray Tracing all while costing hundreds more.
 
Last edited:
Not buying now either. As a raster card the 2080 is barely any faster than the 1080Ti that wasn't enough of an upgrade over my 1080.

If/when DLAA is available for a decent set of games I'll reconsider. It looks really promising from the handful of tests publicly available, but until NVidia makes models for each game and packages them in the drivers it's not doing anything.

If dev's are able to get something useful out of raytracing at 4k, I'd reconsider too; but when the NVIdia demos struggle at 1080p I'm not holding my breath. Raytracing at 4k is probably going to end up an aspirational goal until at least 5nm. (4x as much area would require a bigger die at 7nm even if 100% of the extra area was spent on RT.)

Disappointing, even though it means I'm probably going to be getting 3 years out of my last card instead of the 2 I'd planned on. Not a huge surprise though, the raster core counts and clock rates remained about the same; which is why I didn't pre-order.
 
2080ti fail. 30% performance increase over 1080ti with a huge $1200 price tag.

2080 Ti is a big chip, considering the size of the chip in a niche market, the 2080 Ti will sell as well as any Titan level part. The 2080 is a lot more a question mark though.
 
Last edited:
2080ti fail. 30% performance increase over 1080ti with a huge $1200 price tag.

I smell a price drop incoming and a lot of cancelled pre orders.

It's been a 30% increase almost every generation for the flagship card with a ~$700 price. Why this card is $1200 with the same performance increase is absurd. Nvidia can go F themselves.

The 2080 is essentially a 1080ti with less memory and Ray Tracing all while costing hundreds more.
Not sure I'd call this a fail. It's what is expected. There is no linear increase in performance vs price at the extreme high end. This applies to pretty much all computer parts. There's a certain price premium for having the very best.
 
2080ti fail. 30% performance increase over 1080ti with a huge $1200 price tag.

I smell a price drop incoming and a lot of cancelled pre orders.

It's been a 30% increase almost every generation for the flagship card with a ~$700 price. Why this card is $1200 with the same performance increase is absurd. Nvidia can go F themselves.

The 2080 is essentially a 1080ti with less memory and Ray Tracing all while costing hundreds more.

Halo price for a halo card. I dont't know about you, but previous flagship cards are titans and those usually run above 1k. I wish Nvidia didn't go about renaming a Titan to a Ti but that is what happened.
 
Don't let Hameeedo see that chart his head will explode when he sees Fury X over 980 TI .

I agree 2080 and 2080Ti just dont seem like a worthwhile upgrade for the price Nvidia expects.
 
I think the 2080ti should really be branded as the 2080...
and the 2080 should actually be branded as the the 2070

Then they could unveil a "REAL" 2080ti in six months like normal.

I believe this would all be the case except one word = profits
 
In 3840x2160 Wolfenstein 2 the 2080 got 91.3 FPS and the 1080 Ti got 70.5. So we need more Vulkan games?
 
I think the 2080ti should really be branded as the 2080...
and the 2080 should actually be branded as the the 2070

Then they could unveil a "REAL" 2080ti in six months like normal.

I believe this would all be the case except one word = profits

Profits make the world go round.
 
I think the 2080ti should really be branded as the 2080...
and the 2080 should actually be branded as the the 2070

Then they could unveil a "REAL" 2080ti in six months like normal.

I believe this would all be the case except one word = profits

my guess is they did it due to the confusion they caused by releasing the titan V.. that was a terrible name to use on it since it's not even a gaming card which then completely screwed the whole 2080 line up.
 
I think the 2080ti should really be branded as the 2080...
and the 2080 should actually be branded as the the 2070

Then they could unveil a "REAL" 2080ti in six months like normal.

I believe this would all be the case except one word = profits

I do have a sneaky suspicion that may happen, since there is no change in pricing for the Pascal stack.
 
my guess is they did it due to the confusion they caused by releasing the titan V.. that was a terrible name to use on it since it's not even a gaming card which then completely screwed the whole 2080 line up.

No more confusing than releasing two Titans during Pascal days, Nvidia could have easily call it Titan T.
 
Not sure I'd call this a fail. It's what is expected. There is no linear increase in performance vs price at the extreme high end. This applies to pretty much all computer parts. There's a certain price premium for having the very best.

1080ti disagrees! You cant compare GPUs to all other computer parts. After 2 years you should get 70-80% performance increase. Check out Pascal, it was just that.
 
time to upgrade my 980 to a 1080ti,,,,,,oh wait, still a thousand bucks in canada...
 
1080ti disagrees! You cant compare GPUs to all other computer parts. After 2 years you should get 70-80% performance increase. Check out Pascal, it was just that.

How do you figure that? We're getting close to the end of the road on process with GPU's just like with CPU's. What happened with past performance isn't an indicator of future performance anymore unfortunately.

Also, 12nm isn't a node jump at all. Just a refined process. That doesn't make for much uplift there. NV is getting all of their uplift from a bigger chip. Of course they could have used more of that real estate for rasterization, but they see RT (and so do most developers) as the next step in graphics. We're not going to see 70% rasterization uplift in a 7nm Turing successor either IMO.
 
1080ti disagrees! You cant compare GPUs to all other computer parts. After 2 years you should get 70-80% performance increase. Check out Pascal, it was just that.

Pascal was a massive node shrink from 28nm to 16nm. The next bigger node shrink will be next gen Nvidia cards from 12 to 7nm plus any architectural improvements they come up with.

I think we are unlikely to see as big a performance difference anytime soon as going from Maxwell to Pascal.
 
2080ti fail. 30% performance increase over 1080ti with a huge $1200 price tag.

I smell a price drop incoming and a lot of cancelled pre orders.

It's been a 30% increase almost every generation for the flagship card with a ~$700 price. Why this card is $1200 with the same performance increase is absurd. Nvidia can go F themselves.

The 2080 is essentially a 1080ti with less memory and Ray Tracing all while costing hundreds more.

I highly doubt they will see many cancellations. For the people that do cancel, someone else will happily take their place. Considering that the upcoming cards are on a completely new architecture, the performance it brings to the table is a good starting point. It's only going to get better as the drivers mature & optimizations are implemented. The 10 series will go down as one of the best releases in Nvidia's history but the massive performance leap that it offered over previous generations has created a lot of spoiled gamers. We shouldn't expect anywhere near the performance leap that the 10 series had for a very long time. If anything, that expectation should be placed on AMD since they have the most ground to make up.
 
The more I look, the more I see these series as a demonstration of all the future technologies to be implemented and potentially widely used in game titles starting 2020 and beyond, released in limited quantities right now. Just to stir up the interest and sell for huge markups, while no competition exists. Are these nice cards? For sure. Do we need or can properly utilize them any time soon? Not really even in limited scenarios yet. Are they worth buying? Hell no, unless you're a collector.
 
Not sure I'd call this a fail. It's what is expected. There is no linear increase in performance vs price at the extreme high end. This applies to pretty much all computer parts. There's a certain price premium for having the very best.
I saw predictions of 40-50% increase. After reading the article it said the Founders Edition cards have an OC on them this time and all other cards tested were reference cards so the gap between an OC 2080Ti and OC 1080 Ti might be even less than 28%. For the price and for CURRENT games I say it is a fail.
 
I sold my 1080ti and backed down to a 1070.
I play mostly Civ, and I'll dabble with betas of whatever AAA title is running but I'm basically playing solitaire.

I don't do any meaningful creative work besides hobby fumbling around.

There's nothing this gen for me to waste $ on.
 
I saw predictions of 40-50% increase. After reading the article it said the Founders Edition cards have an OC on them this time and all other cards tested were reference cards so the gap between an OC 2080Ti and OC 1080 Ti might be even less than 28%. For the price and for CURRENT games I say it is a fail.

I won’t call it a fail, but it’s an expensive chicken with no eggs to sit on. Way too ahead of the game (no pun intended).
 
I highly doubt they will see many cancellations. For the people that do cancel, someone else will happily take their place. Considering that the upcoming cards are on a completely new architecture, the performance it brings to the table is a good starting point. It's only going to get better as the drivers mature & optimizations are implemented. The 10 series will go down as one of the best releases in Nvidia's history but the massive performance leap that it offered over previous generations has created a lot of spoiled gamers. We shouldn't expect anywhere near the performance leap that the 10 series had for a very long time. If anything, that expectation should be placed on AMD since they have the most ground to make up.

Not so sure about people picking up all the cancelled cards. Most 2080TI’s are pre-ordered by hardcore gamers. If those are the ones cancelling, who’s buying those cards? Not the smart people waiting for reviews. They’re going to be as disapointed with the reviews as the guys who preordered. I’m looking at myself. I planned on upgrading my 1080ti with an rtx2080 because the rtx2080ti is outside of my budget ($800). The rtx2080 is not a big enough performance stepup from my 1080ti. This is a marketing cluster frack of epic proportions. I have $800 burning in my pocket that Nvidia is not getting this round unless the 2080ti drops to that pricepoint.
 
Not so sure about people picking up all the cancelled cards. Most 2080TI’s are pre-ordered by hardcore gamers. If those are the ones cancelling, who’s buying those cards? Not the smart people waiting for reviews. They’re going to be as disapointed with the reviews as the guys who preordered. I’m looking at myself. I planned on upgrading my 1080ti with an rtx2080 because the rtx2080ti is outside of my budget ($800). The rtx2080 is not a big enough performance stepup from my 1080ti. This is a marketing cluster frack of epic proportions. I have $800 burning in my pocket that Nvidia is not getting this round unless the 2080ti drops to that pricepoint.

The 2080 is 'ok' for people that are wanting to upgrade from a 1080 or below. The amount of 1080Ti owners out there and the subset of those that were wanting to upgrade to a non-TI 2080 is far less than the people most likely to upgade to standard 2080; those with a 970/980ti/1070/1080.
 
The 2080 is 'ok' for people that are wanting to upgrade from a 1080 or below. The amount of 1080Ti owners out there and the subset of those that were wanting to upgrade to a non-TI 2080 is far less than the people most likely to upgade to standard 2080; those with a 970/980ti/1070/1080.

I am one of those with a 1080Ti that was going to move to a 2080. The operative word being "was". I cancelled my pre-order. From the several reviews I have looked at now, the 2080 barely keeps pace with the 1080Ti. $800 for a card that barely matches a card that has been out 1.5 years. No thanks Nvidia. You won't get my money this time.
 
Hoping for some deep learning reviews of these. It could be a monster budget workstation card.

Gaming? Not so much. 30% over my 1080ti isn't very compelling
 
I am one of those with a 1080Ti that was going to move to a 2080. The operative word being "was". I cancelled my pre-order. From the several reviews I have looked at now, the 2080 barely keeps pace with the 1080Ti. $800 for a card that barely matches a card that has been out 1.5 years. No thanks Nvidia. You won't get my money this time.

It is about equal at worst and in some games quite a bit faster (See Hardware Unboxed's Wolfenstein numbers). Is it a good value for someone that already has a beastly 1080ti? Nope, but then it's still far faster for those with a lower end card that wanted a 1080ti while having more features that may (or may not) pay off. I can still return the 1080 I bought almost a month ago and upgrade but MicroCenter's 1080tis are still $700-ish. If the 2080 is $800 I'd make the $100 jump just in case; the real problem is that they have an open-box 1080ti for $612 which is a little tougher sell.
 
The 2080 is 'ok' for people that are wanting to upgrade from a 1080 or below. The amount of 1080Ti owners out there and the subset of those that were wanting to upgrade to a non-TI 2080 is far less than the people most likely to upgade to standard 2080; those with a 970/980ti/1070/1080.

This is [H]ard|OCP, not Smart|OCP lol. The cards will sell well because despite how a handful of people feel about the cards, there will be people that want the latest & greatest. Since AMD will have no answer for some time, there is nothing really stopping these cards from flying off shelves other than those that choose to wait for the next big performance leap. They will be waiting awhile for that.
 
Looking at the 1080ti vs 2080 (non ti), I can make a few simple conclusions in regards to upgrading:

1. The price difference between selling your used 1080ti and then buying a 2080 is roughly about $300 that you need to kick in.
2. That $300 "upgrade fee" buys you about the same FPS, unproven and unimplemented RayTracing, and DLSS (Anit-Aliasing).
3. There may be a surprising number of people that skip this Generation
 
  • Like
Reactions: Savoy
like this
The 2080 is 'ok' for people that are wanting to upgrade from a 1080 or below.

That was my first thought. But, all this hinges on DLSS and how much of a real world performance boot it brings and the number of games that support it. Otherwise as a 970 owner myself I'd be better off buying a 1080Ti and save a couple hundred dollars to put towards the 2180 ~2 years from now.
 
I'll hold onto my pre-order. I have buddies who want my 1080ti and the 2080ti should be a nice boost at 3440x1440. ~35% increse is solid for me. It will let the games I play ~70fps (yes with a 1080ti 3440x1440 there are a lot of them) to lock into a solid 100Hz.
 
2080ti fail. 30% performance increase over 1080ti with a huge $1200 price tag.

I smell a price drop incoming and a lot of cancelled pre orders.

It's been a 30% increase almost every generation for the flagship card with a ~$700 price. Why this card is $1200 with the same performance increase is absurd. Nvidia can go F themselves.

The 2080 is essentially a 1080ti with less memory and Ray Tracing all while costing hundreds more.

1080ti disagrees! You cant compare GPUs to all other computer parts. After 2 years you should get 70-80% performance increase. Check out Pascal, it was just that.
#1: It's been less than 2 years from top card to top card.
#2: We only see that kind of performance jump with a node shrink. TSMC 12nm is really a refined 16nm with the transistor gates narrowed, hence why we didn't see any real perf/watt increase.

GTX 285 -> GTX 580 = +80% (node shrink, 23 months from Tesla 2.0 to Fermi 2.0)
GTX 580 -> GTX 780 Ti = +80% (node shrink, 36 months from Fermi 2.0 to Kepler 2.0)
GTX 780 Ti -> GTX 980 Ti = +30% (node refinement, 20 months from Kepler 2.0 to Maxwell 2.0)
GTX 980 Ti -> GTX 1080 Ti = +70% (node shrink, 21 months from Maxwell 2.0 to Pascal)
GTX 1080 Ti -> GTX 2080 Ti = +30% (node refinement, 18 months from Pascal to Turing)
I saw predictions of 40-50% increase. After reading the article it said the Founders Edition cards have an OC on them this time and all other cards tested were reference cards so the gap between an OC 2080Ti and OC 1080 Ti might be even less than 28%. For the price and for CURRENT games I say it is a fail.
FE is not overclocked, it just has a higher reference clock. Given that NVIDIA locked the power on the RTX cards I imagine the performance gap is going to be about the same when both are overclocked.
Looking at the 1080ti vs 2080 (non ti), I can make a few simple conclusions in regards to upgrading:

1. The price difference between selling your used 1080ti and then buying a 2080 is roughly about $300 that you need to kick in.
2. That $300 "upgrade fee" buys you about the same FPS, unproven and unimplemented RayTracing, and DLSS (Anit-Aliasing).
3. There may be a surprising number of people that skip this Generation
A 2080 is a sidegrade to a 1080 Ti. You should be going to a 2080 Ti from a 1080 Ti.
 
yeah i cancelled my pre-order
i guess i'm sticking with my titan x pascal for another year or two
 
I feel like DLSS will have to be the saving grace for these cards. If it does what they claim and it's pretty widely implemented, the prices start to look a little less ridiculous. Until then, I feel like they're going to have a hard time moving these after the initial rush.
 
#1: It's been less than 2 years from top card to top card.
#2: We only see that kind of performance jump with a node shrink. TSMC 12nm is really a refined 16nm with the transistor gates narrowed, hence why we didn't see any real perf/watt increase.

GTX 285 -> GTX 580 = +80% (node shrink, 23 months from Tesla 2.0 to Fermi 2.0)
GTX 580 -> GTX 780 Ti = +80% (node shrink, 36 months from Fermi 2.0 to Kepler 2.0)
GTX 780 Ti -> GTX 980 Ti = +30% (node refinement, 20 months from Kepler 2.0 to Maxwell 2.0)
GTX 980 Ti -> GTX 1080 Ti = +70% (node shrink, 21 months from Maxwell 2.0 to Pascal)
GTX 1080 Ti -> GTX 2080 Ti = +30% (node refinement, 18 months from Pascal to Turing)

FE is not overclocked, it just has a higher reference clock. Given that NVIDIA locked the power on the RTX cards I imagine the performance gap is going to be about the same when both are overclocked.

A 2080 is a sidegrade to a 1080 Ti. You should be going to a 2080 Ti from a 1080 Ti.

Any device out of the box set to higher than reference clocks is overclocked.
 
Back
Top