Gadgetory

well guys here we are on what is probably one of the most important days for gamers in the last few years at least that's what Nvidia would like you to think after nearly two years since the launch of the GTX 980 Andy Pascal architecture we finally have some new round of gaming GPUs that offer a lot more performance but they cost a ton of money so this is the r-tx 2080 TI and the r-tx 2080 and unlike many previous generations of Nvidia graphics cards their new touring core architecture both changes and optimizes the way games are rendered there is a lot of exciting stuff going on below these awesome looking heat sinks and I'll be covering some of it here however if you're looking for some more technical details I'll leave a link to our review article posted on our official website down in the description now this video will be focusing on just one item that you all wanted to know about and that is how do these cards perform and are they really worth the huge amounts of money that Envy is asking for so sit back with your cup of coffee and let's dive in but first a message from our sponsor the new T Force Excalibur ddr4 memory is one unique RGB kit with awesome light spill from each module and the special edition has a cool totem design on the light bar you get lifetime warranty up to 4,000 megahertz speeds and full lighting control through the software check out the t46 caliber ram down below alright so I want to start off here with a closer look and what I think is one of the best-looking graphics cards ever created the r-tx 2080 TI and the RT x 2080 share an identical stream line exterior design that uses forged and milled aluminum pieces in either a brushed look or a flat black finish every single part of this card looks like it was put together with an eye towards detail unlike previous founders Edition cars that featured lower star coolers Nvidia decided to go with a different route with the RT X 2080 series they use a pretty low profile downdraft heatsink with a pair of axial fans to draw in cool air this does keep most the GPUs heat within your case but the design is supposed to lower noise and offer cooler core temperatures honestly the looks of these cards actually make me wonder why anyone would one of the triple slot monsters from bored partners some of those designs look way too bulky both of these cards require quite a bit of power so the r-tx 28 ET i needs to a pin connectors while the 20/80 needs a 6 pin and an 8 pin that means I would recommend at least a 600 watt PSU for the r-tx 20 atti and a 500 watt unit for the 2080 but if you're rocking a thread ripper or an overclocked Skylink X system add 200 wants to each of those numbers the back plates on the 2000 series are actually pretty similar to the ones on the 10 atti but the all silver colouring really makes them stand out there's also a little secret hidden here too this is the envy link connector which carries an SLI signal over a new interface at capable of transferring up to a hundred gigabytes per second of data between two cards now there's a few things that you need to know about envy link the most important is that in videos officially discontinuing three and four card setups in its current form the Arctic's series only supports dual card configurations the real benefit of envy link is that it can customize workloads towards the most suitable GPU and it can also share memory capacity and other key resources across multiple cards that means to eight gigabyte cards in sli would have access to a true sixteen gigabyte memory pool it also can scale in linear fashion when moving to high resolution displays and V to give a great example of nearly a hundred percent performance scaling when moving from 4k to 8k on current RT x-series GPUs that's pretty impressive guys and I can't wait to test it out the i/o area of these cards are actually quite unique there are three DisplayPort 1.4 eight outputs that can support 8k at 60 Hertz or 4k at 120 Hertz alongside a single HDMI 2.0 B port with hdcp 2.2 compatibility that can support 4k 60hz our content the cool addition here is the USB C port which is a first for any GPU this isn't actually your typical USB C data connector but rather a virtual link interface that's supposed to provide data and up to 27 watts of power to next generation VR headsets alright so let's end the tour right there and get on to what makes these cards tick this is the touring streaming multi processor which is basically a building block for the core the reason I wanted to show is to explain how NVIDIA has changed things up to benefit today's games and future technologies don't worry I'm not gonna get too technical here the first thing that you will see here is there are now dedicated integer and floating-point units instead of just CUDA cores this was done because Nvidia realized the CUDA cores were often running integer operations while the rest of the course sat idle so they broke the two pipelines apart so both types of operations can now be processed in parallel Nvidia says that this could lead to a 25 to 50 percent speed-up in standard shading tasks the biggest thing added into the main core blocks are the tensor course and each SM has eight of them these are units specifically designed to efficiently process deep learning algorithms for machine learning and artificial intelligence now I know that's a lot to wrap your mind around but NVIDIA is actually putting them to use with their deep learning super sampling anti-aliasing instead if the shader is being put to use here the tensor cores analyzed the scene and applying anti-aliasing where needed in real time the end result is much higher frame rates when a a is enabled within each streaming multiprocessor there's also for texture units and 96 kilobytes of shared memory finally there's a single RT core which is meant to be hyper efficient at processing ray tracing tasks something that previous architecture just couldn't do in videos hoping to put these course to use in future games for real-time lighting and shadow effects so now that you know what each streaming multiprocessor is let's take a bigger look at the tu 102 core that's in the RT X 20 atti this ship is made of up to 16 point eight billion transistors and 72 streaming multiprocessors clustered into six graphics processing clusters or Jie pcs separate from those sections are 96 ROP s that are broken into groups of eight six megabytes of shared l2 cache and twelve 32-bit GED r6 memory controllers but this is the fullest configuration and in order to minimize power consumption and increase yields Nvidia shaved off for SMS and eliminating their associated shader course texture units RT course and tensor course meanwhile one block of eight ro peas a single memory controller and some l2 cache were also removed and all of this leads to the RT X xx atti founders having some extremely high inspects which are way better than what Nvidia offered with a gtx 980ti it also happens to consume a lot of power but that was pretty much expected given the performance in videos advertising another big addition is the new GDD are six memory which allows the RT x-series to have impressive bandwidth which is something required for 4k the big deal here is the price with the founders Edition costing a crazy $1200 which is a thousand dollars more than the base card and five hundred dollars more than the gtx 980ti sure it comes with an awesome looking cooler upgraded components and a slight overclock to the bou speeds but men it's still expensive the r-tx 2080 on the other hand uses a different core than the ti called the tu-104 which means its specs are lower but they are still supposed to be more than enough to overcome the gtx 980ti it should also be noted that it's much more efficient and comes with eight gigabytes of gddr3 rather than eleven gigabytes the r-tx 2080 founders edition cost around $800 which is still pretty expensive especially when it compared to the gtx 280s original price of $550 however if it can outperform the gtx 980ti while offering the advanced r-tx features it may end up being competitive speaking of r-tx features the RT and tensor cores make these new cards part of invidious new r-tx platform this is an ecosystem made up of hardware software and api's to add advanced AI deep learning and ray-tracing into more traditional graphics pipeline honestly guys Nvidia is betting big on great racing and deep learning methods that will become a huge part of next generation gaming and it also seems to be working too since there will be at least 26 games that will come with TLS s support in the next few months and when you combine that with the ones that support rate racing there will be more games with our text features than the total number of DX tall titles that are available today but right now other than rise of the Tomb Raider there aren't any other games that support RT X technologies I'll be covering those in the future but what we care about is how well these two cards perform in games that can be bought today for this review will be using ask Alec X system with an i-9 7900 X or clock 24.8 giga thirty two gigabytes of memory at 3,600 megahertz and all the other components listed here one thing to note is that I didn't enable the new game mode within Windows 10 because it ended up messing up with some key benchmark results you should also know that every one of the numbers you'll see in the charts was an average of three separate benchmark runs we have more information about the benchmarks in the website article if you're interested ok so let's start off with battlefield one dx12 the first taste of what these RT x cards has to offer is pretty impressive the RT x 28 ETI is way way out in front and even the standard RT x 2080 is able to beat the gtx 980ti by about 13% the 4k results continue that trend too but the gtx 28 DTI stretches its legs even more offering at nearly 50% improvement over the gtx 980ti call of duty continues the trend we saw with battlefield 1 both RT x cards are well ahead at 1440p and the TI offers some impressive numbers however turn on the stress at 4k and the RT x 2080 TI's lead widens even more while the gap between the RT x 2080 and the gtx 980ti narrows a bit this is likely due to the RT X 2080 smaller memory footprint destiny is next and the performance seen here is just blazing fast from every one of the cards something I do want to mention right now is the RT X 2080s improvement over the original gtx 980 it typically hovers around 50% or so which is why i've been comparing it to the gtx 1080 TI up to this point these really are some fast cards guys but remember they also cost a small fortune my results in far cry 5 at 1440p are interesting since according to the system's resource monitor two cores on the oricon 7900 x were maxed out during the benchmark run with the RT x 2080 TI this could mean that its performance was slightly limited by the cpu and again at 4k the ti surges further ahead while the RT x 2080 and gtx 980ti are neck to neck forza was another weird one and it seemed to be CPU limited again at 1440p but that didn't stop the RT X cards from dominating they did however pull ahead at 4k and here even the RT X 2080 maintained a pretty big lead over the gtx 980ti that's pretty impressive you can actually tell that envious seems to have their optimizations done right before the Unreal Engine 4 and that's pretty important since there are dozens of games which use it here the r-tx cards absolutely dominate everything else especially at 4k when the 2889 jumps out to a nearly 60 percent lead what you will likely see as these benchmarks go by is that Nvidia score optimizations have allowed our TX cards to excel in TX 12 much like battlefield hitman proves there have been some major improvements in this field when compared to Pascal cards particularly at 4k when the new shading horsepower really comes into play moving on to a much more popular game like overwatch there is nothing to complain about it either at 1440p or 4k but there's something I wanted to mention here while the performance gains are 50% for the Arctic's 20 atti and about the same for the 2080 versus 1080 battle we can't forget how much more these new graphics cards cost they aren't inexpensive but they do give you some amazing performance in today's games shadow of war is an extremely taxing game on the entire system most of all the GPUs texture pipeline neither our TX cards had any issue pushing completely playable framerate at 1440p and 4k while improving upon their predecessors by about 50% once again now by this time I know I'm starting to repeat myself so let's just roll Rainbow six results since they're pretty much in line with the other games tested in this review alright so this is going to be a tough one since while one hammer to total war seems to be an extremely challenging game to run it actually isn't provided you use the DX 11 mode creative Assembly's dx12 implementation has been in beta for years now and it is still far from optimized it actually runs worse than the x11 and we've included here to show that even with a lack of in engine optimizations the RT X series has enough horsepower to deliver playable in battle frame rates believe it or not the latest Wolfenstein game actually provides one of the most surprising yet not completely unexpected results of this entire review at 1440p things seems to be going really well for the RT X xx atti and the Arctic's 2080 since they're able to chew through frame rates despite every detail being maxed out but when the resolution is increased to 4k the RT X 28 YZ performance simply Falls to the floor while the gtx 980ti and the r-tx 28 ET i surge ahead the reason for this it's actually pretty simple the 8 gigabyte frame buffer and lower memory bandwidth caused a severe bottleneck on the Manhattan mission we chose even vega 64 s higher bandwidth number allows it to finally become sort of competitive the final game here in this 13 title test is which are three and the RTS cards go back to their normal leadership positions at both 1440p and eventually 4k as well so there you have the performance results of the RT x 2080 TI and the Arctic's 2080 but testing can stop at this point since there were a few more additional things that we wanted to check out first and foremost let's actually see how these new cards you know compare or how they perform when you put them under pressure for an extended period of time because remember modern GPU start off with higher clock speeds as soon as the benchmarks begin but when temperature rises and videos boost algorithm kicks in and it started trice balancing the performance the temperatures and the power consumption which usually leads to lower frequencies and that's why before benchmarking each game we typically warm up all the GPUs with at least two minutes of load to better simulate how they will actually perform during real gaming sessions starting off with the actual temperatures there are actually some interesting things going on as you can see the RT x 2080 doesn't go above 72 degrees Celsius which seems to be what MVS default peak is for this card the boost algorithm won't allow it to go beyond that this means you could probably get some additional clock speed Headroom by simply increasing the temperature limit in overclocking software the RT X xx atti on the other hand is something interesting too since its temperatures are allowed to peak at a point that's 5 degrees higher and that's where it remains basically both of these are pretty cool running cards but something did strike me as weird though Nvidia could have easily squeeze some more performance out of the arctic's 2080 by simply allowing it to hit a higher temperature like the Arctic's 2080 Ti maybe this was done to ensure a bit more performance a separation between the two cards or maybe for another reason altogether different anyways onto clock speeds now and here you can see what kind of effect in videos GPU boost 4.0 technology has on clock speeds as temperature increases rather than pushing fan speeds up so the RT x-series gets overly the fluctuate core frequencies to balance temperature and power consumption this also goes to show why it is very important to either perform longer benchmark runs to ensure the cards have a warmup period before testing each game while the r-tx 2080 remains around 1900 megahertz for the first 30 seconds or so by the end of the 5-minute test that gets reduced to 1845 megahertz this is literally right in line with the envious boot specification the r-tx when atti gets a bit more shaved off by going from 1785 megahertz to 1680 measures after 5 minutes it too remains right near these 1635 megahertz of founders Edition spec so it's not bad at all but what about power consumption well let's have a quick look I want to start with the r-tx 20 atti because in pre overclocked founders edition form it consumes quite a bit of electricity under load the r-tx 2080 isn't that bad since it manages to outperform the gtx 980ti yet it requires less juice however this card doesn't tell the whole story since as a graphics card is able to process more information CPU answers to memory load increases as well so other components are making both of these cards look more power-hungry than they really are and related to slower older GPUs from a performance per watt standpoint the r-tx 2080 and r-tx 20 atti are far superior to anything that's just released up to now I mean that mega 64 is just embarrassing even though there are two of the most powerful GPUs ever made the r-tx foundation cards are also some of the quietest I've experienced at least that's what it happens when the r-tx 2080 TI isn't operating at extremely high frame rates this is what the r-tx 20 80 TI sounds like an idle and now this is what it sounds like when playing games at well over a hundred and sixty frames per second there is a little bit of coil whine that's for sure and it becomes really noticeable especially if you literally have your ears close up to the card as you can see I am running it on a test bench on an open test benchmark to be more specific but if you're wearing headphones or if it's inside a case it might not be as noticeable however check this out this is what the RT X xx DTI sounds like when it's running at a more reasonable 80 frames per second and under it's almost completely quiet meanwhile our r-tx 180 sample didn't exhibit any of these issues and it operated as quiet as a mouse through testing so if heat sinks that seemed to be more than ready to tackle the heat produced by the tio 104 and the t1 of course let's actually go into some overclocking but before we can do that I want to mention that we haven't explored all the options when it comes to squeezing every bit of performance of these r-tx cards for example and videos for partners are actually including a scanning tool that Nvidia developed alongside the GPUs 4.0 technology this is what the EVGA implementation looks like basically what it does is make overclocking easy for newcomers by launching a simulated load on the GPU while gradually increasing voltage and clock speeds if you move the power and temperature limit sliders to higher values the scanner will automatically shift its targets as well it will eventually settle on a final overclock that should be stable for longer periods of gaming what I did is use invidious counter tool to set a baseline overclock on both cards and then went ahead to using manual inputs to dial the clock speeds I should also mention that the automatic scanner doesn't touch memory frequencies so I needed to work on those too basically my goal was to hit the highest speeds possible that were also stable for longer gaming sessions and that means the results will show frequencies each GPU settled on after 30 minutes of gaming I should also mention that the fans were said between 41% and 52% for all tests and even then neither card came close to the assigned temperature targets if you recall the r-tx 20 atti founders Edition settled at 1680 megahertz in stock form but with a bit of tuning it ended up continually running at 19 80 megahertz the memory didn't get all that far though going from 14 gigahertz to just over 15 point 5 gigahertz one thing to make note of is GDD are six memory has error correction routines which will cause it to throttle rather than show rendering errors most of the time so detecting it's true stable overclock is pretty challenging the amount of headroom for the r-tx 20/80 is pretty reasonable with its going for a boost speed of 1845 megahertz to a relatively constant 2085 megahertz the memory on this particular card actually came close to hitting 16 gigahertz which is super impressive so guys I think after spending a few hours working on overclocking the r-tx cards they do have some gas left in the tank if you actually need some more perform however much like with Maxwell and Pascal they are strictly limited by the amount of additional voltage you can apply in overclocking tools the course rarely reached the higher power limit and never even came close to the temperature limits so it's more than obvious than higher voltage setting would allow you to go even further but how do these overclocked r-tx cards perform well let's check it out in battlefield one performance 40 r-tx 20 DTI and the art TX 2080 gets a good increase of about 12% in both cases but honestly I would likely just keep the settings at stock speeds unless you absolutely need those extra few frames per second far cry 5 provides a bit more interesting results the r-tx 20 atti gets a big bump but the our TX 2080 sees a small increase this could likely be due to a frame buffer or core architecture limitation rather than clock speeds okay so I'm going to try to wrap this up as quickly as possible but there is a lot to cover there's no denying the fact the our TX 20 atti and the our TX 2080 are crazy fast but having looked at the individual charts what do they actually mean and the bigger overall picture the first thing I wanted to take a look at is the our TX 20 atti boundaries additions performance across all games being tested it is extremely good at 1440p but this is obviously a card meant for 4k and high resolutions in some cases it was obviously bottled act by other system components at lower resolutions but when gaming in UHD it's almost twice as fast as the GTX 1080 and over 40% faster than a gtx 980ti those are massive numbers guys switching to the RT x 2080 and i'll say that it offers a huge performance at 1440p by cleanly beating the gtx 980ti at very good frame rates at 4k when compared against the original GTX 1080 or even Vega 64 it doesn't even come close however there will be some rare ultra high resolution situations where it's eight gigabyte frame buffer and memory bandwidth will become a limitation alike in Wolfenstein just be aware of that and modify a setting to reduce a game's memory footprint the be quiet dark bass pro 900 revision 2 is here with the modern IO for your type C accessories and Qi charging gadgets the new hub is good for 8 fans and RGB strips plus steam terior is incredibly modular for airflow water cooling invert systems and now with the power supply shroud check out the V to down the alright guys so I think it's time for some real talk the archaic series is really really exciting and you know the frame rates being put out there are huge but don't forget the fact that both these cards are really expensive I mean the r-tx 2080 costs $800 and the 2080 ti goes for around $1,200 that's expensive in my opinion people with a GTX 1080 or gtx 980ti should really look into upgrading just yet because those cards still offer great performance across the board even at 4k I just think that binding to the know the whole r-tx series is just a little bit too high especially for people who have just recently purchased a Pascal GPU now I probably wait until the new year to see how things could shape up with the RT X features and HDR displays but again until then the 1000 series cards are just more than capable enough to a game or pretty much play any title that's currently available in the market today but if you're looking for even more performance and bragging rights and if you aren't thinking about picking up a 4k HD unmonitored or high refresh rate then yeah the RT x 2089 the 2080 are the only options you have available but what about gamers have skipped Pascal altogether well now might be the great time to upgrade right before Christmas seasons huge lineup of new games if you had the money back in the day to buy the original type next then you should be looking into grabbing the RT X to an 80 TI the same goes for the 980ti users it might be time to order the RT x 2080 the performance delta between high-end maxwell cards and their touring equivalents is just massive now NVIDIA is obviously asking gamers to pay a premium for features that may never become widespread in some games honestly RTX technologies like ray tracing and TLS anti-aliasing are pretty cool to see in tech demos and stuff but you know it may take months or even years till we start to see them in more than a few titles in the end only you can decide if this is a good buy for your needs but if I have to make a choice I actually choose the RT X 2080 founders Edition it costs about hundred dollars more than the custom GTX to me DTI cards but it's reasonable frame rates in most games Nvidia also made it a pretty efficient quiet and overclockable if you need a bit performance plus it doesn't require you to put a huge $1,200 bent on upcoming tech so there you have it we finally have some new graphics cards expensive graphics cards to be more precise but we can soon expect less expensive alternatives to follow the RT x 20 ATT imer takes 2080 now for those of you who are who do have the money to purchase the RT x 2080 and the Arctic's 20 ATI you can be confident that these GPUs will be a fastest on the planet for the foreseeable future I need your with hardware connects thank you so much for watching make sure to subscribe to our new boot sequence channel for the latest tech news and rumors and of course you can check out some relevant content over here I'm signing off and I'll see you guys in the next one