Nvidia: Turing RTX Cards Up to 40 Percent Faster Than Pascal in Some Titles

[bit_user]

Its betetr that they show performance increases vs the previous generation and leave competitor comparisons to third party sites because otherwise they will just get called out.

Oops, sorry for the down-vote. I actually agree.

 

[bit_user]

What exactly is the point of this article? Just to repeat the same vague, unimpressive marketing numbers that were already covered by a couple previous articles?

I thought it had at least a couple details not previously disclosed.

I blame it on the trickle of information from Nvidia, where every new dribble becomes another story and only serves to stir the controversy.

 

[Gurg]

I'll accept that 2080~= to 1080ti in performance, so there is no reason to trade 1080ti for 2080. Seems like the real unanswered question is 2080ti for $1200 minus $300-400 from sale of used 1080ti vs buying another 1080ti for $650-$700 for sli.

Looking at Time Spy results with nice overclocked cpu (5820k) and 1080ti one vs two in sli resulted in a 61% increase over single card. 17126/10578 (Yes I understand benchmarks vs actual game results)

While I also understand the disadvantages of sli, I also question buying a GPU with RTX when there are very limited numbers of games for that eye candy in the next two year time frame until the next Nvidia card comes along. So roughly $700 for 61% increase in sli vs $800-900 for 2080ti with maybe 40%???? with limited eye candy games?????

 

[alchemy69]

I would hope that it was "up to" 40% faster in every title.

 

[Giroro]

You assume that nVidia is "paying" developers to include this when its the same as always. Its Gameworks where they work with developers. Probably don't need to pay them sicne they are AAA developers. Should I assume every time a new AMD GPU launches and some game shows off their new ideas or tech with it they were paid or that they just work with them like a lot of software developers do.

Well if that game was one of a very few number of games during a major press conference, then yes it's very safe to assume that a lot of money changed hands. Although you raise a good point about Gameworks. With how Nvidia operates there's a good chance they had the balls to charge the developers for early access to the tech in the first place.

Do you really think Intel/AMD/nVidia/etc just throw new hardware and ideas out there then the software developers utilize it? Its quite the opposite. They develope new hardware and ideas and work together on it quite heavily.

I agree Nvidia worked very closely with their partners to develop the new features of this card, but those partners were not the game developers - it was companies like Adobe, Disney, Google, and who are paying them $10k to buy nearly identical silicon when they call it a Quadro RTX 8000.
The Ray tracing and AI features were never designed for gaming. In this case I do very much think Nvidia developed the hardware first, and are selling it to gamers as an afterthought. I'm sure they hope that some developer out there does the work to find a real killer use for all that application-specific hardware because it helps their sales pitch... but gaming is only half of Nvidia's revenue, and they have the high-end market totally cornered whether or not any of the new RTX features actually catch on.

 

[TJ Hooker]

I was implying that should happen first. Before it's said the card is xx faster and before its recommended people by the cards both of which has happened here at TomsHardware. It's just frustrating this site is turning into a rumor / propaganda mill. TH would have never recommended people by a card before they test it until this week it happened. I miss the old reputable TomsHardware.

Tomshardware has always reported on press releases/announcements/interviews for new product generations from Nvidia etc. TH didn't say 'Turing is X% faster than Pascal', they said 'Nvidia claims Turing is X% faster than Pascal'. Which we all know to take with a grain of salt, given that press releases can usually be assumed to include some marketing fluff and non-independent benchmarks that are likely cherry-picked. I'm not defending the "just buy it" article, but I don't really see what's wrong with this one or how it's different than similar articles in the past.

 

[techgeek]

Here is what is not good for consumers: when nVidia doesn't compare their new card against any of the competitors cards and only against their own previous generation of cards.

I don't really see it as their job. I mean, how much do you really trust their numbers against their own cards? If they would release benchmarks against competitors, those would draw even more suspicion.

Ultimately, what we need is independent reviews. Was Pascal's announce this far in advance of its review embargo & ship date? I don't recall that. IMO, this lengthy speculation period is the real problem.

While I agree that independent reviews are needed, that wasn't my point. When nVidia doesn't compare performance to AMD's cards, they are telling us that they don't see AMD as competition (not that many of us didn't already know that) and thus it points to why these cards are being priced higher than the previous generation (before the cryptocurrency boom). My point is this is bad for anyone wanting one of these cards. I would love to see AMD drop a bomb and release specs / paper launch / whatever to spoil nVidia's launch to convince them to price more competitively. It never works in our favor when there is one clear winner in a duopoly whether we are talking Intel/AMD or nVidia/AMD.

 

[mortemas]

Speaking of NVidia vs AMD. Interesting thing I read recently was that NVidia’s CEO, Jensen Huang, is the uncle of AMD’s CEO, Lisa Su!

 

[bit_user]

Speaking of NVidia vs AMD. Interesting thing I read recently was that NVidia’s CEO, Jensen Huang, is the uncle of AMD’s CEO, Lisa Su!

Well, it seems to have been discovered in June.

https://en.wikipedia.org/wiki/Jensen_Huang#Personal_life
https://en.wikipedia.org/wiki/Lisa_Su#Personal_life

I will have to check back on Wikipedia, to see if it ever gets debunked.

 

[kyotokid]

...well seems the memory stacking matter has been laid to rest. (as I expected). From the standpoint of a 3D CG artist, looks like the RTX Quadro line will continue to have the edge. For my purposes, VRAM is part of the performance equation.

Somewhat vague on that statement about "building software" around NVLink. What I find interesting is the NVLink bridge for the Quadro RTX line is only 39$ (as opposed to 79$ for the GeForce RTX cards) yet it is supposed to enable full memory pooling/stacking. Makes me wonder if that is not also a function exclusive to the Quadro series cards and/or proprietary Quadro drivers.

Yeah as I said on another forum, "time to make a big bowl of popcorn and get a beer" while waiting.

 

[JamesSneed]

I was implying that should happen first. Before it's said the card is xx faster and before its recommended people by the cards both of which has happened here at TomsHardware. It's just frustrating this site is turning into a rumor / propaganda mill. TH would have never recommended people by a card before they test it until this week it happened. I miss the old reputable TomsHardware.

Tomshardware has always reported on press releases/announcements/interviews for new product generations from Nvidia etc. TH didn't say 'Turing is X% faster than Pascal', they said 'Nvidia claims Turing is X% faster than Pascal'. Which we all know to take with a grain of salt, given that press releases can usually be assumed to include some marketing fluff and non-independent benchmarks that are likely cherry-picked. I'm not defending the "just buy it" article, but I don't really see what's wrong with this one or how it's different than similar articles in the past.

You do make a good point and I may have overreacted. However this following the just buy it we haven't reviewed it article is likely what made me see this in a different light. Anandtech skipped publishing this press release and I can only assume for the same reasons I took issue with it as its a marketing piece making unverifiable claims.

I do worry people are accepting these unverified claims on face value and pre-ordering cards to later find out that 40% increase was in very limited cases.

 

[darkomaledictus]

So doubling the frame buffer was a lie. Great, this gets from bad to worst! Was counting on a higher ram count and now that seems to go out the window...

 

[darkomaledictus]

@MGALLO848 to get that sweet 20 fps Ray Tracing experience at 1080p lol

 

[epdm2be]

"...Normally though its the benchmarks that matter...."

That still applies though.
These articles are merely teasers to hype interest into these new cards (and associated hardware).
Especially since sales of desktop machines is declining. Soon only high-end gaming rigs or the ultra-cheap office crap will be available. To that end I hope decent desktop machines that are "good enough" for (current gen) VR and games become increasingly more popular soon and drive people away from stupid consoles back into computers. Perhaps this is just my pipedream but any new tech, whether it is overhyped or not , helps in this matter.

 

[jimmysmitty]

You assume that nVidia is "paying" developers to include this when its the same as always. Its Gameworks where they work with developers. Probably don't need to pay them sicne they are AAA developers. Should I assume every time a new AMD GPU launches and some game shows off their new ideas or tech with it they were paid or that they just work with them like a lot of software developers do.

Well if that game was one of a very few number of games during a major press conference, then yes it's very safe to assume that a lot of money changed hands. Although you raise a good point about Gameworks. With how Nvidia operates there's a good chance they had the balls to charge the developers for early access to the tech in the first place.

Do you really think Intel/AMD/nVidia/etc just throw new hardware and ideas out there then the software developers utilize it? Its quite the opposite. They develope new hardware and ideas and work together on it quite heavily.

I agree Nvidia worked very closely with their partners to develop the new features of this card, but those partners were not the game developers - it was companies like Adobe, Disney, Google, and who are paying them $10k to buy nearly identical silicon when they call it a Quadro RTX 8000.
The Ray tracing and AI features were never designed for gaming. In this case I do very much think Nvidia developed the hardware first, and are selling it to gamers as an afterthought. I'm sure they hope that some developer out there does the work to find a real killer use for all that application-specific hardware because it helps their sales pitch... but gaming is only half of Nvidia's revenue, and they have the high-end market totally cornered whether or not any of the new RTX features actually catch on.

I still doubt money exchanged hands especially when some were already Gameworks partners.

And the RTX 8000/6000 are not the exact same. Even the 2080 ti is still a slightly cut down version (4350s vs 4608 CUDA cores) and of course way less VRAM, the 8000 has 48GB of VRAM. Its impossible to compare the two completely.

And yes for the main design they work with the bigger companies but do you think they didn't go to major game companies and ask what they think of this? Do you think Microsoft just decided DXR is coming out thats it? No. They all collaborate together to work on the features. Shadow of the Tomb Raider is out is a few weeks. It supports RTX. The odds that it was a last minute addition are slim as those are major game engine reworks.

The market is cornered. May stay that way with AMD having to move everything 7nm to TSMC only and even then Navi is not rumoring to be much competition to Turing.

In fact it might be until 2020 if Intels dGPU turns out to be anything further than Larrabee.

 

[techgeek]

"...Normally though its the benchmarks that matter...."

That still applies though.
These articles are merely teasers to hype interest into these new cards (and associated hardware).
Especially since sales of desktop machines is declining. Soon only high-end gaming rigs or the ultra-cheap office crap will be available. To that end I hope decent desktop machines that are "good enough" for (current gen) VR and games become increasingly more popular soon and drive people away from stupid consoles back into computers. Perhaps this is just my pipedream but any new tech, whether it is overhyped or not , helps in this matter.

I'm not a console gamer, but the reality is if new tech isn't adopted by the consoles it's usually doomed to failure. So if we want raytracing and VR to succeed, then they need to be implemented at some level in the next generation of consoles. New tech adoption requires software developers to use it and the reality is most games are cross platform (not PC exclusive) and the developers target the largest audience and put their resources to use to satisfy that audience. Developers rarely go the extra mile to utilize new PC only features in these titles.

To make matters worse for VR is the high cost of entry. First you need a beefy GPU, but then on top of that you need the VR equipment. This market is razor thin, so you won't see much servicing here until they have the numbers to justify it. I really hope that the consoles get onboard with it and put out a decent VR experience. If they half-ass it then they could sour the whole thing by turning people off of it. I would bet most people who have a bad experience on a VR platform wouldn't be tempted to spend more money on it with the promise of a better experience. The whole adage of first impressions holds true here too.

 

[littleleo]

100% BS, Nvidia can't be trusted and Tom's has sold us all out can can't be trusted anymore. Tom's used to be a good source now I just can't trust it. They need to go back to what made them great. Firing everyone involved in that sham "Just Buy It" article would be a start. But they won't they have sold out and can no longer be trusted.

 

[jimmysmitty]

100% BS, Nvidia can't be trusted and Tom's has sold us all out can can't be trusted anymore. Tom's used to be a good source now I just can't trust it. They need to go back to what made them great. Firing everyone involved in that sham "Just Buy It" article would be a start. But they won't they have sold out and can no longer be trusted.

While I was not a fan of that article myself it is still the opinion of one man. There are actually plenty of the ones on the review team that are actually trustworthy. Personally if Chris Angelini posts anything I trust it instantly. Igor is also very respectable.

I know the article can say a lot for the site but it should be sins of the father type situation. Most of the people doing the reviews are in the same boat as us and a lot are also just community members who have been picked to do reviews and write articles for the site.

 

[cryoburner]

Especially since sales of desktop machines is declining. Soon only high-end gaming rigs or the ultra-cheap office crap will be available. To that end I hope decent desktop machines that are "good enough" for (current gen) VR and games become increasingly more popular soon and drive people away from stupid consoles back into computers. Perhaps this is just my pipedream but any new tech, whether it is overhyped or not , helps in this matter.

I really don't think the overall decline in PC sales is from a lack of interest in PC gaming. People are not dumping PCs en masse to move to consoles, and in fact the PC gaming market has been growing faster than the console gaming market for years.

The decline in PC sales is more due to other parts of the PC market, particularly those lower-end systems used for office work or web browsing not getting replaced as often, which makes sense, since the rate in advancement of new hardware is slowing, and existing hardware already meets the demands for most desktop software. And of course, lots of people do more of their general computing tasks on smartphones now as well. So these declines are happening more in the casual and office use markets more than anything.

I'm not a console gamer, but the reality is if new tech isn't adopted by the consoles it's usually doomed to failure. So if we want raytracing and VR to succeed, then they need to be implemented at some level in the next generation of consoles.

VR already is implemented at some level on at least the PS4, and with over 3 million units sold so far, the PSVR outnumbers sales of the Vive and Rift combined. 3 million is still only a small portion of the 80+ million PS4 consoles out there though. I do suspect that VR will likely play a larger role in the next generation of consoles though.

As for raytracing, that's arguably a bit different. You don't need raytracing-capable hardware to run a game that features raytraced effects, so it should be treated as an advanced lighting effect. It may require some additional development resources to implement properly, but it's not like VR where the game needs to be built around it from the ground up. I suspect popular game engines like Unreal will make it fairly easy to apply raytraced effects, which can simply fall back on less-advanced effects should the appropriate hardware not be available.

To make matters worse for VR is the high cost of entry. First you need a beefy GPU, but then on top of that you need the VR equipment.

This is still somewhat true, but it has been getting a bit better. Windows Mixed Reality headsets have been available for a while for as little as $200-$250 with controllers, which I would say is very reasonable, even if their ergonomics and controller tracking might not be quite as nice as something like a Vive or Rift for a couple hundred dollars more. Eventually, we will likely see further price drops for the original Vive and Rift as well, and perhaps even updated budget models with as good or better specs to complement their next generation of higher-end hardware.

The graphics card market seems to be recovering from the mining craze as well, and an RX 580 can be had for a little over $200 now, which will likely do a reasonable job running most existing VR games. Depending on how the 2060 performs and is priced, we will likely see even better performance in the sub-$300 price range within the next few months, and I imagine AMD will be offering some good competition in that performance range not too long thereafter.

 

[stdragon]

While I also understand the disadvantages of sli, I also question buying a GPU with RTX when there are very limited numbers of games for that eye candy in the next two year time frame until the next Nvidia card comes along. So roughly $700 for 61% increase in sli vs $800-900 for 2080ti with maybe 40%???? with limited eye candy games?????

There's also the the aspect of depreciation in value the older the cards get. So might be a break-even choice after you factor in selling your 10 series card now vs selling two 10 series cards later.

 

[jimmysmitty]

Especially since sales of desktop machines is declining. Soon only high-end gaming rigs or the ultra-cheap office crap will be available. To that end I hope decent desktop machines that are "good enough" for (current gen) VR and games become increasingly more popular soon and drive people away from stupid consoles back into computers. Perhaps this is just my pipedream but any new tech, whether it is overhyped or not , helps in this matter.

I really don't think the overall decline in PC sales is from a lack of interest in PC gaming. People are not dumping PCs en masse to move to consoles, and in fact the PC gaming market has been growing faster than the console gaming market for years.

The decline in PC sales is more due to other parts of the PC market, particularly those lower-end systems used for office work or web browsing not getting replaced as often, which makes sense, since the rate in advancement of new hardware is slowing, and existing hardware already meets the demands for most desktop software. And of course, lots of people do more of their general computing tasks on smartphones now as well. So these declines are happening more in the casual and office use markets more than anything.

I'm not a console gamer, but the reality is if new tech isn't adopted by the consoles it's usually doomed to failure. So if we want raytracing and VR to succeed, then they need to be implemented at some level in the next generation of consoles.

VR already is implemented at some level on at least the PS4, and with over 3 million units sold so far, the PSVR outnumbers sales of the Vive and Rift combined. 3 million is still only a small portion of the 80+ million PS4 consoles out there though. I do suspect that VR will likely play a larger role in the next generation of consoles though.

As for raytracing, that's arguably a bit different. You don't need raytracing-capable hardware to run a game that features raytraced effects, so it should be treated as an advanced lighting effect. It may require some additional development resources to implement properly, but it's not like VR where the game needs to be built around it from the ground up. I suspect popular game engines like Unreal will make it fairly easy to apply raytraced effects, which can simply fall back on less-advanced effects should the appropriate hardware not be available.

To make matters worse for VR is the high cost of entry. First you need a beefy GPU, but then on top of that you need the VR equipment.

This is still somewhat true, but it has been getting a bit better. Windows Mixed Reality headsets have been available for a while for as little as $200-$250 with controllers, which I would say is very reasonable, even if their ergonomics and controller tracking might not be quite as nice as something like a Vive or Rift for a couple hundred dollars more. Eventually, we will likely see further price drops for the original Vive and Rift as well, and perhaps even updated budget models with as good or better specs to complement their next generation of higher-end hardware.

The graphics card market seems to be recovering from the mining craze as well, and an RX 580 can be had for a little over $200 now, which will likely do a reasonable job running most existing VR games. Depending on how the 2060 performs and is priced, we will likely see even better performance in the sub-$300 price range within the next few months, and I imagine AMD will be offering some good competition in that performance range not too long thereafter.

The PC gaming market always has ups and downs though. However it is never dependent on the console market. During the console market crash in the 80s, pre-Nintendo, PC gaming was still going on its own. Not as large as now but still it still had new games made and released.

Raytracing can be done via software but is of course like anything superior when you hace dedicated hardware backing it. It has been capable since DX11 in software.

CR is like anything else. Its still a new entry market. It is adopting farther and will get more advanced and more accessible. Hell I can't wait till Tiger Electronics makes their own VR headset for all their great games.

 

[englandr753]

@JimmySmitty

"GSYNC 2 will launch with HDR support so not sure how they are not supporting it."

Could this be the reason my 1080ti "flickers" while running my 8 series Samsung 55" tv on occasion? Im on my laptop and don't see the normal menus to reply to previous messages.

@Jamessneed

"I do worry people are accepting these unverified claims on face value and pre-ordering cards to later find out that 40% increase was in very limited cases."

Was my exact thoughts as well and until the benchmarks come out to prove otherwise, I'll clinch fist my $$$ and wait for Volta.

 

[bit_user]

Was my exact thoughts as well and until the benchmarks come out to prove otherwise, I'll clinch fist my $$$ and wait for Volta.

You mean Turing, or is there something I'm missing?

Because the only Volta GPU was the V100 (sold as Tesla V100, Titan V, and Quadro GV100). Turing is basically its successor, inheriting its CUDA core redesign and improving on its Tensor cores. So, unless they're going to release low-end GPUs without the RT Cores and call them Volta, I think Volta is already in the rearview mirror.

 

[bit_user]

Speaking of GPU family names, I wonder when they'll name one after Stephen Hawking.

He's been such a pop culture figure that many are probably unaware of one of his earlier discoveries: a form of radiation that can escape from black holes, eventually causing them to evaporate.

https://en.wikipedia.org/wiki/Hawking_radiation#Black_hole_evaporation