For the price of a Titan RTX you have a complete pc with a 2080 Ti. My choice is quickly made : I do not intend to buy anything at the moment.

You can get 2 apple stands for the rtx Titan

I need a rtx2080ti to run Microsoft PowerPoint at its full potential

when a gpu costs more than my car... *FEELS BAD MAN*

I'm going to buy a RTX Titan to fully enjoy Minesweeper.

Dang it would get me 3 more FPS in far cry than the 2080ti! Better take out a loan

Rtx 2080ti for gaming Rtx titan for hanging up on the wall

Woohoo, 5 percent faster and twice as much money.

The graphics card alone is equivalent to the price to build a full fledged High end Gaming PC

Turns down two or three settings down from ultra to high* Me: yeah I just saved $1,300

Oh wow i needed this video! Thanks! I wasn’t sure which one to get for my 600$ build

2500 for a graphics card lol that's more than my computer itself

AKA the RTX Definitely Can't Afford It VS RTX Can't Afford It

Dont be dumb jay if you really need more vram just download more of it. Just like refular ram.

skip to 4:40 for benchmarks guys!

You can get two apple Imac pro stand for the price of one Titan rtx *Worth it*

deep learning testing?? = MS Paint

Last time I got here this early Jay was air-cooling

when Unreal Engine 4.22 comes out I will be testing these RTX cards for reals 🔥

Couple this 2 RTX Titan with 8 Gold Trident Z Royal. Butler: *Here your gold stock sir*

AMD, please do something to force a change in the market.

Who wants to see RTX Titan vs Titan V..?? I do :-)

So lets answer my question. Can it even run Crysis ?

"When you look at it, it gets finger prints" - I feel with my eyes!

Nvidia : we need to release new card Employee: lets just make the same thing just change the name and add $1200 Nvidia: YOU ARE PROMOTED!

Hey Jay, I'm a graduate student doing research in ML and AI. What kind of tests are you looking for? In general, the reason GPUs are so heavily used in deep learning applications is because training a neural network with a huge dataset requires a large number of matrix multiplications, much of which can easily be parallelized via a GPU. There are several bottlenecks when it comes to training a neural network on a GPU. The amount of data you load into memory could exceed the total memory of the card. This happens quite often for huge datasets, which is why most people do batch learning where we train a model on a small subset of the data at a time. Another bottleneck could be the size and complexity of the neural network. A small, simple neural network can be handled on a CPU. More complicated architectures such as ResNet, LSTMs, DenseNet, etc. are more demanding. If you want to see the true benefits of the RTX Titan, you will probably want to try two types of tests. One, train a neural network that can be handled by both cards and show that the Titan finishes training much faster. Two, try to find problems where training the model on the RTX 2080 Ti causes "out of memory" errors or is extremely slow even with batch learning. Maybe some deep reinforcement learning tasks would be good to show as well.

4:24 Little did Jay know that Minecraft was going to get ray-tracing.

So hilarious diffrence One is silver and one is gold hahahaa Man that was funny🤣🤣🤣

The fact that you were wrong about the naming scheme makes the prices of each card even more ridiculous. Somewhat understandable if the 2080Ti was a TITAN (because that's how much they usually cost) but now they've released a TITAN that's double the price? Absolutely insane.

An extra 1300 dollars for 13 Gigs of Ram.

Titan RTX is Packed Card for Work Efficiency with Gaming and Workstation. Its Worth it Dudes

I simply don't understand why you don't try utilizing the cards in applications that can take advantage of their heavy lifting capabilities like Unreal Engine 4 live rendering, or performance testing the card in Houdini, Blender, 3DsMax, Maya etc. or contact Redshift and get a sample for rendering tests to see what the card can really do instead of meaningless synthetic benchmarks. Will the card boost video render times in NLE's?

Wow, the 2080ti is so cheap!!!

I think memory is the key why the 2080 ti is considered a gaming card and the Titan RTX isn't. Someone who does 3D-graphics and uses a GPU renderer like Redshift or Octane will run in to problems with rendering times if the scene uses more memory than is available on the card. Gaming cards with 8 or 12 gigs are usually enough for freelance work, but there might be some specific cases where a scene will use a ton of memory. Big Hollywood studios are also slowly moving towards GPU rendering solutions so someone who does lighting and rendering could use one of these cards. Texture painting software like Mari also benefits a lot from high memory. Titan RTX is actually relatively cheap compared to the quadro alternatives so I could see it being useful for many people.

A cool test would be Otoy’s octane benchmark software. I think they even have a beta that takes advantage of the RT cores. I don’t know if it’s available yet, but they do have a screen shot of the final score on their website and the difference in using the RT cores is something like 40% faster!

As a soon to be mechanical engineer I would love to see you test these 2 cards for CAD applications such as Solidworks, Catia and Siemens NX

For the AI/Deep learning benchmarks: PLZ include as many GPU generations as possible! i.e I am training deep learning models on both a GTX 970 and a GTX 1070 Ti. I don't want to buy a 2080Ti, if I don't know how faster it will be...

how many tabs on chrome can it run at 60fps?

RTX Titan vs RTX 2080Ti Winner? GTX 1080Ti...

Can we just appreciate how sexy this card is?

I would call it the Titan R and it's definitely a card ment for game developers not gaming

3:04 "And when u look at it it gets finger prints" jay2cents 2018

Hey Jay, some people say it's not worth upgrading to the 2080 Ti for gaming at 1440p max settings over 100fps. I think these cards are more aimed towards 4K monitor game players and not just 1440p. Especially that now we're getting 4K 144hz monitors well over $1,000 in price like the Acer Predator X27. My MSI Gaming X Trio 2080 Ti runs games over 100fps at 4K... Technology is getting so superb now that we're literally achieving frames above 100fps at max settings in 4K resolution and this is why we're getting monitors that can support those high frames at that resolution. I have nothing overclocked on my computer at all and I reach up to 120 fps at times at 4K resolution WITH max settings on my 2080 Ti. I play Black Ops 4 (not as graphically intense as say...GTA5) all the time, but I achieve an average of 100fps at 4K easily in an entire online match in that game alone without breaking a sweat. The 2080 Ti is a monster. Black Ops 4 is probably not the best example, but at least it's there for evidence and something a lot of you can think about. If you don't game in 4K, don't buy the 2080 Ti or the RTX Titan. It's really that simple. The 1080 Ti can even do 4K (barely past 60 to 70fps tho) and can achieve really high fps at 1440p in the high 100-150fps too. Why spend over $1,000 on a GPU if you're not gonna game in the highest resolution other than CAD, content creation, modeling, etc.? That's just my opinion... Also, you can actually take advantage of the extra VRAM of that RTX Titan by increasing the render resolution of games being played. Example, I know if I adjust the render resolution under Black Ops 4's settings when I'm already playing at 4K, that my VRAM increases HEAVILY and can actually max out my 2080 Ti VRAM. Especially by turning the render resolution to at least 140 when it's already at 4K. lol. I keep at 100 tho cuz there's no real use for it when high frames is the goal. With over 20+ VRAM on the RTX Titan, I'm sure most could potentially increase that render resolution to 200. That would literally double the graphical resolution to 8K but being played at 4K I'd imagine. Although I agree that another $1,000 wouldn't justify a couple settings being altered to achieve something like that...

I noticed the shortcut to Crysis. LOL, that game is 12 yrs old and still being used to test high end pc's.

10:00 look at the temps and clock speed, the titan probably doesn't have enough cooling

Suggestions for Deep Learning benchmarks: - Mostly benchmark model TRAINING speeds. - Compare 2 model types: 1. Convolutional Neural Networks (CNN). 2. Recurrent Neural Networks (RNN). - Compare PERFORMANCE per $. - Test different model sizes: From 1 layer to 20+ stacked layered (or more).

3:04 when you look at it get finger prints.... 😂😂😂😂

Like a i9 over a i7, Getting Titan over 2080ti do not justify the price if it only for gaming, but it does as a combo Gaming/Workstation machine.

When a graphics card costs more than your car.

Meh gt1030 still better tho

How well does it run Crysis?

I do 3D modeling/rendering and gaming, so the Titan cards are perfect for me.

Jay, Thanks again for a great video. Also send a shout out to Phil. Great team and great content :)

It is good for GPU rendering (iRay, Octane) where you need your scene to be stored in VRAM to actually work. Considering multiple 4096x4096 (and higher res) textures on each object - it takes a lot of memory and 11 GB is not enough. So this is professional card that is good for semi-professional rendering jobs (big studios go for Quadro rendering farms, I guess).

Titan RTX isn't a gaming-directed GPU, it's was designed for Artificial Intelligence. Train a huge ML model on the Titan vs the 2080ti, then you'll see the true potential of the Titan.

+$1000 for -10 fps! Take my money already!

*EVERYTHING JUST WORKS*

But can it run Minecraft at 720p 30fps ?

Say what you will but the titan rtx being 10% faster is HUGE usually with titan cards we pay 2x for a 1 fps loss and a 1gb gain, here we got 12gb more AND Noticble performance uplift

As a metalhead I really enjoyed the harder music during the results montage! Wouldn’t mind more of that and happy new year!

Microtransaction should NOT be somthing acceptable..... like cmon

Thank you, Jay, for considering AI/Deep Learning workflow benchmarks (along with gaming ones)! FINALLY! We don't only play games, many of us try to do some research too...

I'm still keeping my 2x780gtx OC to 1034ghz. I will keep it for next.. 5 years 😁. Easily. No problems with most games.

24GB of VRAM isn't enough for my (neural network) model optimization (feed forward training)... I was hoping for an option with ~128GB of dedicated VRAM.

So what about Titan V what's the purpose of that card if the RTX titan is for Al and deep learning?

I use CAD... And I have never understood the "This is for Professional CAD design" argument... CAD isn't even close to being as demanding as gaming is. I mean, gaming rings in so so so exponentially higher in demand, it's not even something that can be expressed.... it's Windows 3.1 vs 10... it's an abacus vs ... windows 10... It's a computer in a rudimentary software setting, crunching away after you click "cook", rendering out some basic textures on some basic low poly (in respect to games) shapes, with pre baked lighting, and zero effects (in respect to gaming), with no dynamic anything. Workstation laptops do it with ease.... 5 Grand in video cards is something you will not EVER find in any of the places I've worked... So I really don't even get who the target is for something that is the same as something else, but due to ... Drivers? does it shittier in nearly all ways. I mean, who even is that target audience? Assholes and idiots ?

But Jay, I don't even have money in my bank for Christmass gifts...if someone breaks into my account they'd probably feel so bad they'd make a charity donation lmao

If you put 2 TITAN RTX and i9-9900k and 3x32 RAM DDR4 SSD

Hey Jay, i am a Cloud Specialist and i can add a little info for your video. The Tensor core on nvidia cards are gaming changing since most of Tensor resources we have to run models in the cloud are software based, which adds a layer of overhead for most things we wanna do. For example to run any Deep Learning model on the AWS Cloud, we pay for underlying hardware that is not designed for AI applications, so the software do all the job. With hardware tensor cores, a new path opens for developers to use low level APIs to reach specific goals. I hope the big cloud companies start to acquire those cards to provide hardware oriented AI services!

Been waiting for this video

These results are actually surprisingly good by titan standards, usually a titan is identical(give or take a few frames) and typically double vram honestly having it in some cases make a 30 fps difference, does it justify the price? Ofc not but your value gamer isnt exactly the ones buying one of these are they

Man human eyes won't able to see more than 1080p

Please test performance in 8K. Does the extra memory bandwidth help?

love the metal breakdown during the benchmarks \m/

I'd love to see how it handles with Adobe After Effects!

WHY are people hitting dislike on your video? there was nothing controversal at all you simply provided the info we wanted and told us what we could see for ourselves in the video. SMH haters!

You should test 2080 Ti SLI vs Titan RTX SLI at 8K. The Titan's extra RAM may make a difference at that resolution.

rtx is probably one of the least exciting series i've ever seen in long ass time yeah ray tracing is cool but it's not cool enough to play at 1080@70-ish fps with 2400$ card .. especially NOT in a multiplayer games

RE: What Deep Learning tests you should do: Train the Tacotron2 text-to-speech model from Nvidia's github repo (Google it). One a 1080ti it takes about a week to train a voice with 13000 sentences in wav format. Love to see how it goes with two Titan RTXs NVLinked!

I'd like to thank you for presenting actual benchmarks. There are many KZfaqrs out there that just read the news which infuriates me because I just read it myself and wanted benchmarks which they failed to provide. hahaha

0:40 Dbrand GRIP!!! Never actually seen anyone but myself use that awesome case!

Is Nvidia trying to kill itself?

11:00 - Actually you're right the first time, the 2080 Ti is not a gaming card, it's a table scrap compute card, and the tech has been that way for a while now.

it’s for rendering not for games: Vray RT, Fstormrender, Redshift, etc. or for large scale master plans created with Lumion. RTX option is going to be implemented in Lumion, Unreal Engine and Arnold very soon.

Test it on threadripper with both cards and NV link.

The same day I received my RTX 2080Ti, this week (had been waiting for the MSI Gaming variant to become available in my country, since August last year, but it never came in stock so I gave up and got a different brand), I ran the FFXV benchmark, and saw there was 1 card faster than mine. That was the first time I heard about the Titan RTX. Kind of considered returning my new card and get a Titan instead, but it'd just be a waste of money, for gaming only. And it's not even that much faster apparently, for games. If it had been 50ish % faster, I would have bought one.

what if this privacy ad was actually a secret thing that stole credit cards

Was the mural networking and AI stuff to hard? Where's the follow up on that?

Every episode on this channel: "..I digress.." Maybe you should change the channel name to "JayzDiGress".. just a thought..

5:18 the 2080ti actually beat the titan

I miss this benchmark metal music and have been searching through dozens of videos to find it.

Those reflections in Metro are screen space reflections lol.. Metro Exodus only has raytraced global illumination

5:20 is the summary :)

"This is the gold-colored manifestation of my incorrectness." LOL

Now that Arnold GPU is coming out in a few days, I'm pretty sure those Titan RTX cards are going to disappear. The larger RAM pool is the biggest draw for 3D animation.

It's the same as the 2080ti when it comes to the speed but this one has 24gb instead of 11gb.It's kinda a hybrid card when it comes to the use. So it's nice for a person that wants to game and be productive. Saved you a few minutes

Really digging the new music for the benchmark segment of the video.

Dumb for games. 'Smart'... for Otoy Octane and large scenes. (And some 'prosumer' AI: again, moar ram).

Will this be the same story for the 3090 vs a potential 3080Ti (not announced) ?

All these new RTX cards are way ahead of their time and its going to take years for game developers to really show what they can do. I think new games built from the start with RTX tech will really push these cards and then people will start to see the difference between older cards. performance to cost is low for now with these RTX cards but they do future proof your pc.