Nvidia ,Rtx2080ti,2080,2070, information thread. Reviews and prices September 14.

[monkeydelmagico]

This doesn't make X70 buyers feel any better about their upgrade option being priced at $600. Six. Hundred. Dollars. For an X70 card. Anyone trying to rationalize these prices can go right ahead and throw themselves a party, but I'm not coming. I'm grabbing my stuff and I'm walking home, and no, you can't give me a ride.

Don't buy current gen cards? During the release of 1070 a top end brand new 980ti went for around $400.-. If this new card manages the same type of quality/speed increases along with the stack shift described already then I should be able to pick up a 1080ti on blowout for around $450.-. Still a bit spendy for my budget tastes. If the stars align I predict somewhere around black friday.

As a backup plan I already bought a new playstation. There is a better than 50% chance I won't bother to upgrade my PC any more. I have plenty of money to do so, but I just don't see the value of it.

 

[crisium]

Re: $400 980 Ti: Also consider that the $450 FE / $380 MSRP 1070 was faster than a stock 980 Ti and had more VRAM. The $600 FE / $500 2070 is already confirmed to have less VRAM, and speculation is abound that it will perform slower than a 1080 Ti in standard rasterization games (based on Nvidia's own slides no less). It'd be a pleasant surprise but I wouldn't count on new retail 1080 Ti's for $450 anytime soon as there is less downward pressure from the new gen as we are not expecting to see traditional performance-per-$ gains.

 

[realibrad]

https://www.eteknix.com/digitimes-nvidia-forced-aibs-to-absorb-glut-of-10-series-gpus/

This should back up the idea that the current prices are to help clear out old 10 series inventory.

Basically, if the partners want 20 series chips, they have to accept shipments of the old 10 series chips.

 

[alcoholbob]

I assume you factored in the fact that Volta has ~800 more cores than the 2080 Ti. I am not expecting much faster than Titan V, maybe 5% or so.

I assume you factored in anyone with the ability to use a screwdriver could put an aftermarket cooler on the TV and go way beyond this performance level.

 

[PeterScott]

Don't buy current gen cards? During the release of 1070 a top end brand new 980ti went for around $400.-. If this new card manages the same type of quality/speed increases along with the stack shift described already then I should be able to pick up a 1080ti on blowout for around $450.-. Still a bit spendy for my budget tastes. If the stars align I predict somewhere around black friday.

Often after the Reviews come out, the pricing gets adjusted to match the relative perf/$ in favor of old stock so they can clear it. So you often end up getting slightly better deals on old stock if you shop around. But it usually isn't by much, and you may be missing out on newer features.

But to be clear, prices don't just drop by large amounts in isolation, they drop in relation to the perf/$ of the new incoming cards.

Since this release will be something of a hiccup on perf/$, you can likewise expect less price reductions needed on old stock to clear it.

So I also don't think we will see any movement of GTX 1080Ti below $500, and it will probably be higher than $550 for some time.

But definitely any value buyers should be waiting for reviews, to see how it shakes out, before hunting for deals. There isn't much point buying either (new or old stock) until we get a real handle on the nature of performance changes and other improvements.

 

[RichUK]

I have the itch for a 2080ti - get me some gigarays.

 

[LTC8K6]

https://www.eteknix.com/digitimes-nvidia-forced-aibs-to-absorb-glut-of-10-series-gpus/

This should back up the idea that the current prices are to help clear out old 10 series inventory.

Basically, if the partners want 20 series chips, they have to accept shipments of the old 10 series chips.

That story keeps coming up. I've always wondered if it made any sense.

What prevents the partner from making the 10 series cards attractive to get rid of that glut of 10 series chips forced on them?

There is probably more profit in a 10 series card now, than in a new 20 series card, for partners and retailers.

I can clearly see how NV could get rid of too many 10 series chips.

Figuring out how the partners or retailers would do that is a bit murky to me.

 

[MrTeal]

That story keeps coming up. I've always wondered if it made any sense.

What prevents the partner from making the 10 series cards attractive to get rid of that glut of 10 series chips forced on them?

There is probably more profit in a 10 series card now, than in a new 20 series card, for partners and retailers.

I can clearly see how NV could get rid of too many 10 series chips.

Figuring out how the partners or retailers would do that is a bit murky to me.

I think the partner making 10 series cards attractive and clearing them is exactly what Nvidia wants. For the partners the issue what price is necessary to push a large volume of cards out, and who takes the hit on that. Even if perf/$ in traditional games is flat and current 10-series owners aren't pushed to upgrade, it's a no brainer to buy RTX if price/$ is the same as the 10 series and you're in the market for a card. Not only that, system builders always want the newest SKU to refresh their own product lineup.

Clearing out any significant volume of old cards will require them to be priced decently lower than a comparable 20-series, so at least some discounts are probably incoming. If nvidia is discounting the price per die that they're selling to the AIB card manufacturers, they might not have a problem with it. If nvidia is selling them a 1080Ti die for the same price as they did at launch but the AIB is now having to sell the card at $500 instead of $700, that's obviously more of an issue for them.

 

[LTC8K6]

I think the partner making 10 series cards attractive and clearing them is exactly what Nvidia wants. For the partners the issue what price is necessary to push a large volume of cards out, and who takes the hit on that. Even if perf/$ in traditional games is flat and current 10-series owners aren't pushed to upgrade, it's a no brainer to buy RTX if price/$ is the same as the 10 series and you're in the market for a card. Not only that, system builders always want the newest SKU to refresh their own product lineup.

Clearing out any significant volume of old cards will require them to be priced decently lower than a comparable 20-series, so at least some discounts are probably incoming. If nvidia is discounting the price per die that they're selling to the AIB card manufacturers, they might not have a problem with it. If nvidia is selling them a 1080Ti die for the same price as they did at launch but the AIB is now having to sell the card at $500 instead of $700, that's obviously more of an issue for them.

If NV wants a lot of cheap 1080ti cards on the market, then the 2080 must actually be quite a bit faster in existing games. Otherwise, cheap 1080ti cards would certainly reduce the sales of 2080 cards.
If the 2080 and the 1080ti are close in performance, and one is $200 cheaper...lots of people will take the cheaper card.
Same for the 2070 and cheap 1080 cards.

There's something wrong here with high priced 20 series cards and cheap 10 series cards competing if some of these cards fall into the same performance levels.

The only thing that fits is if the 20 series cards are actually fast enough on release to make people pay more for them over the 10 series cards.
If they aren't, then I'm having trouble making any sense of this marketing.

 

[sze5003]

From what I saw In one of those time spy or 3d mark graphs was that the overclocked 2080 was equal in performance to the 1080ti. If that was indeed a 2080 they were testing not a 2070.

I also agree people that don't want the Ti and are looking for a 2080 will just save more money buying a 1080ti instead.

 

[MrTeal]

If NV wants a lot of cheap 1080ti cards on the market, then the 2080 must actually be quite a bit faster in existing games. Otherwise, cheap 1080ti cards would certainly reduce the sales of 2080 cards.
If the 2080 and the 1080ti are close in performance, and one is $200 cheaper...lots of people will take the cheaper card.
Same for the 2070 and cheap 1080 cards.

There's something wrong here with high priced 20 series cards and cheap 10 series cards competing if some of these cards fall into the same performance levels.

The only thing that fits is if the 20 series cards are actually fast enough on release to make people pay more for them over the 10 series cards.
If they aren't, then I'm having trouble making any sense of this marketing.

If the cheap 10-series cards are only cheap because AIBs are taking the margin hit and nvidia is still clearing out existing stock at full price, that's not really a bad thing for them though. Especially if initial volume on RTX is lower, it lets them clear all their existing inventory while charging a premium for the new parts. If people choose a cheap 1080 Ti over a launch 2080 that's not really a problem if they're initially capacity constrained on the 2080s.

 

[ioni]

In my opinion these tech reviewers should go out and buy the cards and then no nda is required and they can freely review.

If every reviewer did that, then how would the get reviews out for cards before release so people could decide to buy right away or not? They'd have to compete in the F5 derby, and even if they won, wait a few days for the card to ship, then a few more days to do the reviews, so the reviews would be out 1-2 weeks after the card was available to the general populous. Sounds like the only winner there would be the card manufacturer.

 

[DeathReborn]

I assume you factored in anyone with the ability to use a screwdriver could put an aftermarket cooler on the TV and go way beyond this performance level.

In the UK the Titan V is £2,800 and the 2080 Ti is £1,099 so they can instead buy a pair of 2080 Ti's and £600 to buy some water coolers for them.

 

[sze5003]

If every reviewer did that, then how would the get reviews out for cards before release so people could decide to buy right away or not? They'd have to compete in the F5 derby, and even if they won, wait a few days for the card to ship, then a few more days to do the reviews, so the reviews would be out 1-2 weeks after the card was available to the general populous. Sounds like the only winner there would be the card manufacturer.

That's why manufactures would allow reviewers to buy the cards early. They give samples out ahead of time, let them buy ahead of time too, no difference.

 

[TheF34RChannel]

From what I saw In one of those time spy or 3d mark graphs was that the overclocked 2080 was equal in performance to the 1080ti. If that was indeed a 2080 they were testing not a 2070.

I also agree people that don't want the Ti and are looking for a 2080 will just save more money buying a 1080ti instead.

The 2080 would be much more appetizing, to me at least, if it came with more than its 8GB. I always like getting more VRAM (it's better to have and not need, then to need and not have). I'd be satisfied with 10GB.

 

[maddie]

If every reviewer did that, then how would the get reviews out for cards before release so people could decide to buy right away or not? They'd have to compete in the F5 derby, and even if they won, wait a few days for the card to ship, then a few more days to do the reviews, so the reviews would be out 1-2 weeks after the card was available to the general populous. Sounds like the only winner there would be the card manufacturer.

Mr Pavlov would be proud.

 

[RichUK]

Didn't realise EVGA offered a 3 slot design for the 2080 (deeper heatsink / fans, etc): https://www.evga.com/articles/01249/evga-geforce-rtx-20-series/

 

[coercitiv]

The only thing that fits is if the 20 series cards are actually fast enough on release to make people pay more for them over the 10 series cards.
If they aren't, then I'm having trouble making any sense of this marketing.

There's another scenario that fits, one in which 10 series cards don't drop much in price and 20 series cards are not much faster, creating a overlapped offering of both new and old cards. This would explain why AIBs bought the extra 10 series chips so close to new gen launch.

 

[maddie]

Didn't realise EVGA offered a 3 slot design for the 2080 (deeper heatsink / fans, etc): https://www.evga.com/articles/01249/evga-geforce-rtx-20-series/

Quick, quick, who knows the answer?

 

[LTC8K6]

There's another scenario that fits, one in which 10 series cards don't drop much in price and 20 series cards are not much faster, creating a overlapped offering of both new and old cards. This would explain why AIBs bought the extra 10 series chips so close to new gen launch.

It would quickly become apparent that the 20 series cards only offered RT over the 10 series cards, though. At that point, unless RT is a fantastic advance, 20 series cards are going to have a bad reputation.

 

[Qwertilot]

Wouldn’t be such a big surprise - the xx80 never gets *that* far past the previous generations xx80ti. The xx80ti is way ahead but crazily priced for the moment of course.

 

[dlerious]

Wouldn’t be such a big surprise - the xx80 never gets *that* far past the previous generations xx80ti. The xx80ti is way ahead but crazily priced for the moment of course.

Except for the 1070 which beat a 980Ti.

 

[coercitiv]

It would quickly become apparent that the 20 series cards only offered RT over the 10 series cards, though. At that point, unless RT is a fantastic advance, 20 series cards are going to have a bad reputation.

Not if perf/price points are chosen right. Take another look at the only performance figures Nvidia gave us.

Let's say this means RTX 2080 is 50% faster than 1080, excluding DLSS for now. If we normalize for price we need to compare 2080 to 1080Ti though, that means shaving 30% off that performance advantage. That leaves us with a 20% advantage, which can be managed by ensuring prices are lower, but not much lower as to make 2070 unattractive. DLSS and RT will make sure Turing looks more appealing for early adopters while raw performance numbers will likely favor Pascal in the eyes of perf/dollar aficionados.

Win win right? /s

 

[RichUK]

BF5 running at 4k with ray tracing enabled runs at sub 30FPS: https://wccftech.com/ray-traced-battlefield-v-runs-sub-30-4k/

Though tuning could increase this.

 

[Hitman928]

BF5 running at 4k with ray tracing enabled runs at sub 30FPS: https://wccftech.com/ray-traced-battlefield-v-runs-sub-30-4k/

Though tuning could increase this.

So, according to Digital Foundry (the youtube video WCCF links to as their source), the RT in BF5 is limited to reflections only. So The rest of the lighting and shadowing is being done "traditionally". Makes sense now why they only focused on the reflections and nothing else during the BF5 demo.

Also, Dice isn't using AI for denoising, they're using their own custom post processing algorithm that doesn't use the Tensor cores.