Jhact007 The name itself is so sexy that when you say it out loud you grow abs and sexy mustache and all the ladies will be at your feet. Amiriteorwhat.
"they got more money than you or me" linus you have more money than me. I'm cleaning restaurant fryers to get the money for a 400 dollar pc and you put like 4 titan x's in a rig
IKR, this happens to me ALL THE TIME! In other news, my $9000 a plate fundraiser failed. I don't get why, the plates were made of unobtanium! It was a BARGAIN!
I worked in a super complex military optical company, and I tell ya, they bought like 100 Quadros for all of their engineers so they can calculate some super colplex agile of glass and mirrors inside a gigantic replica of Hubble Lens just for practice before they actually make a new one for a future space telescope for the USA ... Just looking at the lens structure would make you tripping at the technology level of it!, looks like a gigantic micro hive of super accurate cutted lens (that shi't is actually semi hollow from the inside, full of tunnels and such that you can't see from the outside), the actual view of perfection in your eyes. So I guess that in those type of things, you can't except less than 100% accuracity, therefore, Quadro is the best choice, even when its 10 times the price.
I am a CAD CAM designer I used SolidWorks, Keyshot, Fusion 360, PowerMill type software I have a choice between GTX 1050 4Gb in gaming Pavilion and Quadro m2000 4Gb in Zbook workstation. please recommend me I am really confuse between these two machines because I need performance
@@MegaFaisalpk well, you probably would be better off with the quadro because of the high amount of vram and the other features mentioned in the video. But if it is much more expensive just go for the 1050 it has a similar performance.
Quadro cards are 10-bit whereas the GTX are 8-bit. So for Color Correction Grade Monitors in Cinema, colorists use 10-bit RGB Monitors. so you get the 10-bit signal from the cards to the monitors.
It's pretty much spot on. But here's an extra tidbit: in CAD applications, FireGL/Quadro performs A LOT better! (up to 10x or more compared to the consumer version, even though it's built using the same silicon) The difference all comes down to the drivers. Traditionally (although this trend is slowly starting to shift) CAD applications draw stuff in immediate mode which pretty much meant that you had to make a draw call for each element you see on screen, and things were kept in system memory. This is in stark contrast to video games which grouped elements together to minimize draw calls and leverage GPU memory as much as possible. So GPU makers intentionally crippled immediate mode performance on consumer card drivers without having any detrimental effect on gaming performance (which consumers care about), but dramatically slowed down CAD performance (which consumers don't care about) - forcing us poor CAD users to pay through the nose. Google for some choice keywords and you'll find some interesting investigations and findings, and you might wonder how GPU makers are able to get away with it at all.
Son of a Zombie I'd bet there isn't a dime to be made there. CAD professionals aren't going to spend any money on a hack that isn't supported by the hardware vendor or certified by the software vendors.
Folopolis CAD workers hate money? Maybe if the developer sold it to Autodesk, charged you $2000 for a license and rubber stamp "certification" you would like the idea better? it is clear now why NVidia is bending you guys over! lol
Son of a Zombie If you are a i.e. car manufacturing company, that makes yearly profit of X billion dollars, you ain't gonna use a hack to save a few thousand or even tens of thousands and risk stopping your production if something goes wrong. It's one of reasons why big companies rarely switch to FOSS. Not because FOSS is bad, but level of support they expect is usually lacking.
In my experience Xeon processors and Quadro perform a lot better than consumer grade "equivalent". I don't know if it's because of larger cache, ECC memory or better building quality in general. But when you work with big assemblies or complex shapes with something like Solidworks, Inventor or Creo consumer grade hardware inevitably fails at a certain point. I speak from many years of experience.
But Linus! Surely a Red Rocket-X is great for gaming! It's $8k! Srs tho. deadmau5 has a Quadro 5000 and has nothing but problems running games on it when he streams from his studio. So yeah...
werethless12LP video cards priced above the titan z and black are OBVIOUSLY not meant for gaming. They are programmed or BUILT differently. They are obviously made for professional video content creation. Just because a video card can help CREATE a pc game DOESN'T mean it's MEANT to be played on that game.
Raymond Tyson Did you even read the second half of my comment? Professional GPUs can play games, but they suck at doing so. Although, I doubt a Red GPU would be able to run any game..
This quickie videos are great. I only stumbled upon them 2 weeks ago, even though I've been a technician for years. They come in handy for easy explanations.
Im studying 3D design in University. They told me to pick up a Lenovo W540 workstation with a quadro. I also game and i knew that the card wouldnt push modern games, so i went with a Lenovo Y50 which rocks a 2014 maxwell based GTX 860m... Prior to university i would render high poly 3D scenes on an AMD APU based platform.. A super detailed scene at 720p, 2000 samples could take 4hours... With the CUDA function of my GTX 860m in the laptop, the same exact scene at 1080p, 2000 samples is about 40mins... Its a world of difference.. Glad i didnt get the workstation.. It was bulky, ugly, and hardly could push 40fps on games like Warthunder medium settings 1080p.. I push a full 90fps with Warthunder on movie and its really enjoyable. So note to anyone getting a laptop for 3D design or CAD work... The CUDA cores of a 2014 GTX seriously gets the job done + plays games where as a workstation with a quadro workstation laptop at the same price basically is not worth your while. Not to mention workstations are a lot heavier, thinker and bulkier..
And considering that Adobe's applications and drivers support gaming grade GTX cards and a few Radeon ones, forking out more money for a quadro or firepro just seems like a waste of money to the average design student/beginner.
see louder When you render single scenes in software like blender, use the GTX 860m gpu. It'll finish the job much faster. If you're doing a long video that includes multiple frames, I would suggest using the core i7 4700HQ CPU because it's more stable and is less likely to crash for larger scale projects. :D Hope that helped mate! The GPU will always be faster and infact it can usually do large scale projects too. If you dont want to take the risk of crashing it, still go with CPU for big video/animation renders otherwise for games, editing and saving video, photoshop, AI, etc.. - ALWAYS use the GTX 860m as the default render.
A Tesla is more for computations rather than physics or 3D rendering but it can be used for both of those as well, earlier models didn't even have any kind of display output iirc.
also I don't know about now, but originally I think Tesla wasn't a line of cards, but a peripheral cabinets that did calculations. (correct me if I'm wrong)
as an aero/automotive engineer, I 3d model near daily, as do almost all of my coworkers. we use regular gaming nvidia cards (combined with pretty hard core processors, heaps of memory and decent ssd's), though we do use different drivers more optimized for openGL, since most of us don't do that much with directX. it saves the company quite a bit of money so we can upgrade the workstations more often. rendering is still slow as balls if you want decent quality, but i only do it twice a year to generate shiny shareholder pictures. during rendering or simulation (FEM and CFD are even nastier on resources and i use them a whole lot more), i'll run my workstation at night, take it home during the weekend (allthough my electricity bill is not happy with that), bugger off to the workshop or coffee maker, or my absolute favorite; commandeer a second workstation and let that one do the hard work via remote access and/or mail notifications. you don't need 10x more expensive hardware to do decent work (not in my job anyway, might be different for the people at pixar e.d.), what you do need is decent hardware that you're allowed to pick yourself. tough CFD's and the likes should be run on blade farms or a linux box with a big AMD processor, fast memory and very good cooling
One of the other benefits on the precision of the quadros. Often we, in animation, deal with both very small numbers and very large numbers in a single scene. ie a mountain top with a character on it may be 2500000 units away from origin, and camera truck from far away may mean that we exceed the floating point precision of the scene. The double precision just helps to keep things from jittering as the viewport renderer tries to figure out what polygon is on top of another (called z-fighting), it also helps to draw curves because the tangency of a curve can be very small, especially if you zoom in on it, like in a grapheditor.
+aseredy Nice! Getting a pair of Quadro M5000 cards to play with, can't wait to see how well 3Ds MAX behaves with them as opposed to my GTX cards (which still kick butt in pure compute). Too bad Maxwell kinda curbed DP performance though...
Sweet! Maybe it was nerfed because Max is a mostly Games suit, and you don't want to be building scenes for cinematics that most game engines will artifact on. I'm a Maya guy so I'm not too familiar with Max's tech (used it in limited capacity years ago). Though, I'm not sure having DP impacts artist productivity all that much, but all those annoying little artifacts that happen with pure gaming cards definitely go away. So, you may just notice better over all UI performance instead. Though any investment in productivity is a good one, and Maya's overall user experience is greatly improved with Quatros. Max will probably be the same.
aseredy Yeah, I've seen some kickass viewport improvements in Maya when switching from a more powerful GTX to even a lower-grade Quadro, so I'm expecting great gains here too. Can't wait :D
another thing quadros have is much slower memory buses. i've had to talk about this to someone i know who was confused about GeForce and Quadro before. i had to explain it as "a GeForce will do things as quickly as possible and get you that 60FPS, a frame didn't turn out right? drop it whatever, a Quadro will do it correctly and keep rendering until it's done right" another thing Quadros and Firepros have is the fact that they are much more hardy, i've heard tell of a computer science professor having two CUDA farms, one made of Quadros and one made of GeForce 780's, the 780's only lasted a couple of days before burning out from 24/7 use. the quadros were fine.
This video has been a great help in understanding the basic difference of two types of GPUs. But this is 5 years old and many things have changed. I mean new versions of quadro gpu are available and also Threadripper has made a great impact since then. Can you make a revised version this video
"Well jeeze, if only the position of my character's head could have been calculated to 15 decimal places..." This Linus is the best!! I can't stop laughing :D
***** Ah, thank you. I think I had tried looking there before but I only started watching Linus' videos moderately recently and with the amount of content he puts out I may have just given up before I got to it.
Why just GeForce vs. Quadro? Don't forget AMD's Radeon vs. FirePro ... Also, about NVidia artificially lowering the compute capabilities of GeForce cards ... I really wish they didn't. It started with the 600 series, so the 590 is better at double precision calculations than the 780 Ti. There's no reason the latest high-end card should be artificially crippled to perform worse than a card that's two generations old.
The difference is mostly the same between FirePro and Radeon (per the video description), except that AMD cards are MUCH better compute cards than NVidia's. I too dislike how Nvidia forces you to get their overpriced, super expensive "pro-grade" workstation cards if you want to get any decent compute performance out of your card.
The reason they do it is so they can sell GPUs that wouldn't pass QA as Quadro (Because they would need all 8 floating-point units working for double-precision math and one of them is faulty in some way), but do pass QA as GeForce (Because then they only need 4 floating point units for single precision math). It's the same reason AMD sells 6-core CPUs and (Intel sells Dual-core CPUs) - They're actually 8-core (quad core, in Intel's case) CPUs, but 1 or 2 of the cores on the silicon is faulty, so the faulty pair is disabled with a hardware fuse and the CPU is sold on as a cheaper part, to save them losing money by not being able to sell the part at all.
Muhammad Alwan That's a shame ... I use NVidia cards because CUDA is so well supported and easy to use compared to OpenCL. I guess I just need to learn to code better ...
It is called market segmentation. In cars, you have the same chassis and/or the same engine equipping models of completely different categories. In digital cameras, you had the same sensor equipping from a Sony Nex3 to a Nikon D7000.
not just Nvidia, it's logic, you always go to the deepest pocket with a stiff price, because you sell less and you have to justified the cost of development to this particular product. and this go the other way, when you have a wide average customers who counts the pennies you lower the price to sell more.
Linus left out the GeForce cards are made in way bigger numbers and stuff made in large numbers is always cheaper, so it's that + they have way more expensive memory and such + it's aimed at companies. So you probably don't have much to respect Nvidia for, hehe... ;) (though I'm a long time Nvidia user, have been since FirePro's ran 3DLabs Glint GPU's and were made by Diamond. I switched to Nvidia when they were bought by ATI who messed everything up - because ATI)
To an extent. For example, starting with the 600-series cards, NVidia purposely crippled their GPUs to perform slower with double-precision calculations. The 590 is twice as fast as the 690 for computing.
asdfrozen And that's why I run 2x EVGA 580 Classified's for CUDA (and OpenCL) crunching. But man are they HOT! This box alone runs at $100 for electricity a month, hehe... :D
Hey Linus, I'd love to see a video on what kind of a difference a high end mobo would make. Motherboards always seem to be almost like an afterthought in my builds, is there something Im really missing by getting a $130 mobo vs a $200+ one?
This is why the Mac Pro is so expensive. You can get similar performance from a custom build PC for less but you don't get the goodies from professional (that's why its called a Mac "Pro" grade hardware as seen from this video.
AlbertoScrubington You're probably calling someone autistic while 1) you don't know what autism realy is 2) he probably isn't . And by the way, I wouldn't be surprised that a mac does indeed have higher grade hardware. And no I am not an apple fan, far from it.
KingClaus of Yllin The Mac Pro specs' might justify it price the best compared to any other Apple product, but all their products are still heavily overpriced.
The reason the cost is higher from my experience, is that certain programmes require specific quadro cards purely for features within the programs themselves. I use Solidworks you need a quadro or FirePro to use graphics features in models.
Have you accepted lord Gaben, king of the gamers and the nerds and the computer men, lord protector of steam and all it's sale events, first of his name as your lord and savior?
Okay Linus but I want to see a Quadro m6000 vs a Titan X Pascal fps comparison in multiples games. Does the Quadro actually get more fps than the Titan X or what? I know it's not meant for gaming but how well does it perfom in games?
A quadro is made for companies, 1 card can power upto 6 monitors. I bought a nvs 290 for €7,50 and it works very good for games like gmod, minecraft, cs:go and it can power 2 monitors.
he did that on his main channel, and the drivers do for games, he played Crisis 3 on it and a titan, they were the same, but the Quadro has twice the RAM at 24GB.
Quadro, Tesla, Maxwell, Pascal, GeForce, Volta, Kepler etc. - I think you guys are perfect to explain and put each one in the right box to make sense of it all. PLEASE!!
Hey now I was a painter for many years painting all sorts of houses and Interiors of those houses and outside of the houses and I will tell you that even that 5% difference is huge overtime.purdy is one of the best brushes on the market and worth the money
Techquickie hey guys. i always have knowledge on the videos i watch that you post, but its great that your videos help clear up confusion or misinformation that I may have!
ChRoNicBuRrItOs it's from chaos theory it says that it's possible for a butterfly's wings to cause storms. He's joking saying that a gpu fan would cause even more chaos.
+Mario P. Cardenas Europe ftw fam. Looks like it could possibly be Norway. It looks very Norwayish if you know what I mean with the mountains and low population.
Thanks for resuming all the differences. Much of the internet say only they have more memory, but this is incomplete. from what i know now: For image/matrix transformations they are good, and they have all the software ready to do it. FPGA can do fast parallel processing, but they need to be programmed and cost more than GPU if doing image/matrix transformations. CISC CPU (i core) are good to do random sequencial tasks and complex maths with large floating numbers (they have a co-processor). RISC CPU are cheap but they need more steps to do the same task, they are in all tablets, PDA, digital mobile phones, GPS, and can be in some low end computers, cars, washing machines, routers, TV, and more.
If I recall correctly there was a regular GeForce which could be turned into a Quadro card by simply loading another BIOS. Can't recall which one though. It's a relatively recent card. But really all of these features are for powerusers.
not fully ... like no ecc and less reliable in obtaining more accurate results but u are 90 percent correct . same for some amd products . they increase rendering speed by a lot
hackad ay.com/2013/03/18/hack-removes-firmware-crippling-from-nvidia-graphics-card/ without the spaces ... this is just one example . there are older models also
Yeah, its quite hard to justify why Quadro/FireGL has better performance over GTX especially in recent years. I have work in this two "world" with the same CAD software (office Quadro FX 4600 and home gaming rig GTX660 x2), it does make a different. First, if you are using CAD software such Solidworks, you will see the different instantly (highlight and real time rendering). While in other CAD software its hard to spot the difference, especially if you are a beginner. But ill give you a hint where to spot, try to rotate a part. The part edges on GTX/HD are blurred just like in games (Motion Blur), while on Quadro/FireGL they are precisely placed (less eye fatigue). And yes basically they just changed the memory type to ECC, implement a special driver which there are plenty of hacked ones, and stamp a higher price. Comparing with games, GTX/HD has faster memory since it does not have to correct errors and can be overclocked easily. Meanwhile for working, as Linus explained the brush, this are just tools. Once you know when to use this proper tool then you will achieve a better product other wise its just an overpriced solution.
The paint brush example is not very correct, what you pay for with a top quality brush is longevity more than quality. You can pick up a $6 brush and it'll work almost good as a $80 bush on the first day, but a month down the track the $80 brush wont have you wanting to paint walls with your socks.
The double-precision floating-point operation on consumer GTX Titan is fast (if not faster than Quadro and Tesla product line) like it never been crippled at all, they are comparable. Maybe the only difference between GTX Titan and Quadro-Tesla are ECC memory support and product reliability for continuous usage.
+Casper sanderson It's much of the same. * Their target market is businesses with lodsemone, so prices are high. * They support ECC (error-correcting code) RAM * More L3 cache * Higher-binned * Options with moar cores
I used to use an EVGA SR-2 system with two Xeon X5650s for gaming (OC'ed via BLCK to 3.5 ghz) they did fine, in fact you can overclock current xeons with haxx0r XMP profiles. And there are engineering sample Xeon E5-2690 v4s on Ebay atm for $550ish, which are 20 core/40 thread. Go nuts if you have the money =P
I have been watching this video since 2014 hoping to find something new that justify the cost of Quadro cards. Am i going to watch again in 2017 or coming years?
The reason for the cost being so high is quite simple. If you develop a product for consumers, that is built - let's say 1 million times - you can split the cost over those 1 million built products. But if you can only sell 10.000 professional cards, the cost of developing the product is split only over this 10.000 pieces. So you got 100x as much for developing for the card. The price for a product isn't just the cost for the resources you need and the production. Why does everyone forget that? It's also the cost for development, better support for professionals, etc. Why do you think professional operating systems are way more expensive than those for home users?
another difference: consumer products have roughly the specs u expect from them, they are always some veriations. a quadro gpu has the exact output as any other quadro so u can compare the results. 5000 dollars issnt much. the screens at my eeg-workstations are in that pricerange.
In addition to double-point computation, other gotchas are that you need a Quadro to provide a 10-bit/channel color signal to a wide-gamut monitor. Also need a Quadro to use NVIDIA's NVAPI. The NVAPI provides warp/blending capabilities at the OS level allowing you to create seamless projected desktops on both flat and curved surfaces.
i would just like to say that lynda is awesome, just starting my level 5 computing to get my diploma and this was a great since i havent really done anything in computing in a good 3 years except play games
Alexander Smith I was thinking that there wasn't really a point in doing a video about Tesla's on Linus' channel since he focuses more on gaming oriented products and misconceptions. Most people haven't even heard of a Tesla GPUs and very little people even use one unless your gonna power a super computer or something. On top of that, they are much much more expensive and no average consumer can afford one at M.S.R.P Don't get me wrong, I'd be interested in such a video, but most of the gaming community and consumers aren't interested and this channel isn't towards professionals, just gamers and enthusiasts.
I imagine one key difference here is that there's a much more in-depth development for the drivers and much better if not very customized support for the specific programs these graphics cards will be running. So they're more reliable for their specific use cases and they probably finish their workload faster. Imagine you're a movie studio head honcho and you know that by using these cards, your renders will get done considerably faster thereby allowing you to do more work. Or say the company is an oil or minerals exploration/extraction company that relies on a certain set of programs to map out where and how to dig next. The faster and more reliable your processes, the faster you get to digging or the better you get to timing your new projects in response to market conditions, etc etc etc. The 10x price tag can easily be paid for by that saved time not too far down the road.
Actually, gamers are starting to need double precision graphics rendering. I know a lot of games that would benefit from it. Some of you may have have seen a phenomenon called z-fighting, where two objects close together or very far away look like they're clipping through each other at random points across their surfaces, or which one is in front is flickering constantly. This is because the view transformation is becoming numerically unstable because your view distance is too long and/or you can see things too close to you at the same time. Going from 32-bit floats to 64-bit floats eliminates this issue. Also, calculating at 64-bit isn't dormant in GeForce cards. Rather, it's deliberately crippled. GeForce cards can do it, but it's about 8 times slower than it is on a Quadro - on purpose - making it unusable for scientific calculation or games.However, to be fair to NVIDIA, AMD is doing exactly the same thing with their FireGL line. Yes, that's why the Mac Pro is so expensive, in case anyone is wondering.
IshayuG No, Z-fighting happens because two or more faces of geometry are co-planar. Nothing to do with view distance. And if the vertices of two models are in precisely the same place, changing the precision of their world co-ordinates will not change their location. Launch your favourite game engine on a PC running a Titan. Put two planes in exactly the same place with their normals facing the same direction. Notice the Z-fighting?
redchant Two problems here. First, putting in a Titan and launching your favorite game won't do anything. The game is programmed to use single-precision floating point numbers, and even though your Titan supports double-precision does NOT mean the game will use it, and so you're almost certainly not going to notice any difference whatsoever. Secondly, I already mentioned Z-fighting from co-planar surfaces, but this is not the only type of Z-fighting that can happen. If the view transform is numerically unstable, you can get some points that were supposed to be in front of other points suddenly be behind them or right on top of them when they're displayed on the screen, and this is, by far, the most common source of Z-fighting. It's generally caused by too long view distance.
I use a pretty bad Work PC, with some pretty bad GeForce Card. Then a man came to me and aaid "Its dangerous to play with 5 FPS, take this" and he gave me a the Quadro FX 1500. I was so happy.
In games, it doesn't matter if a bullet shotted hit correctly. But it does matter in industries if calculations go wrong slightly. which justifies the cost.
Can you please stop saying FireGL, it's FirePro and has been for several years now. And on the same note, can we please stop sidelining FirePro, they outperform their Quadro counterparts on every level, except for the W8000 which is improperly marketed.
"Nvidia would be crazy not to take advantage of that". No, they'd be good moral people to not take advantage of that HAHAAA! :D ...but then again, there actually IS justification for the ridiculous cost, and they are not likely making more money by selling QUADRO Graphics cards as they do selling GeForce cards. BUT JOHN!, you're probably saying, The cost is MUCH greater! So of course they are making more money by selling those!! Weellll, I'd bet money against that and here's why. The consumer market is massive and consists of every single person in existence...just about...a large percentage of it anyway, but the market for 3D production that requires the quality and precision of a Quadro is MUCH much smaller. So what does that mean? It means that they have a big factory - it's not that simple, I know, just for example sake though - and all those costs to run that big factory including insurance, power consumption, employee pay and benefits, etc. have to be divided up between every card they manufacture to get an average cost per unit. And as is per usual, when you make a smaller batch of anything, it costs more per item. This is all an over simplification, but I think it's probably the best reason they are charging so much. Of course, there is some greed in there too, but the ridiculous price is not COMPLETELY unjustifiable I'd say. Only somewhat :P
+John Meise Yeah, but they probably make a killing on each of these cards. That said, most businesses don't care as the time wasted waiting on rendering or w.e justifies the cost of paying 10x the money. They may lose more money just in time wasted if they are a multi-million dollar company.
Actually the Purdy paintbrush has a market leading, "leading edge" cut, without getting to paintbrush nerdish, the individual bristles don't splay apart
actually the pratice of disabling some feature and selling as a lower end model has to do with defect and process variation. The chips might be design equaly, but after manufacturing and testing, some might present defects that disabling features or setting a lower clock frequency masks
+Mr.Iron I think he meant to ask if he can SLI a Quadro and a Titan X together. No. You can with either 2 Titan Xs together or 2 Quadros together, but not with one each.
Oh, and I forgot to mention, you can put them together on the same board. You just can't connect them together in SLI, they have to work separately, switch when you need to do something.
+charles chan Just like Chad said - you can't SLI them, but they can be on the same board. This is great if you drive your display and viewports on the Quadro - you get the acceleration there, which is what Quadros do best - while the GTX card will kick in when rendering - which is what it does faster than the Quadro cards. Best of both worlds, IMO!
+PIXELFLUX Not unless they both have the same amount of VRAM and are running on same firmware. You theoretically CAN turn an equivalent Quadro to a Titan X and vice versa (and it has been done with older cards).
Guys! I Have a question :D Im making a PC and here are the specs: Intel i5 Skylake 6400K (OR) AMD FX8320 Black Kingston HyperX Fury 8GB SandDisk Plus 120GB SSD WD 1TB HDD (Here is what im wondering about) Zotac GTX 950 AMP! NVIDIA Quadro K420 Cane theese two pair?
+YellowPeaches Nope :( sorry to say, you can only do SLi with two of the same cards, however I do belive you are able to pair 2 different AMD cards. :)
you forgot the most important factor completely, quadros are much much much faster in certain applications thnx to optimized drivers. So the quadro cards pay for themselves in no time to the right consumer.
The two driving factors are quality and purpose. If a game hangs in the middle there is no mayor harm done, restart and that's it. If you need to be working extended hours, 10, 12 or more and for a professional task that has to be delivered on time, The value of the card dilutes on the amount of jobs that are done with the card. When the difference between getting the job done and not getting it done is a couple of thousand dollars that value is spread across the amount of job that you do, if you are a gamer or casual designer, it is more efficient to buy the best dollar/value ration. That the reason even at mechanical task professional tools are several times more expensive that consumer ones, same applies to cooking, farming, painting and and task that has the be done. Possibly a consumer card has 95% of the capabilities of the pro one, but is that 5% the one that robs you the most margin on the job. I cannot comment on the price differences because that is market driven, but normally professional people pay premium just to make sure that the product is as close as 100% perfection.
+Skukkix23 Agreed. I still have my Purdy brushes. I will not borrow them to anyone. You never knew what you are missing until you try one and compare it to cheap and midrange brushes. It took use to convince me that I'm never buying a cheap brush again except for using it as a tiny broom on the shop toys.
I don't think he is saying professional grade products like the brush or card aren't better suited for their tasks, just that how much better they are doesn't equal the major price increase.
Just some food for thought, the university i go to (I study aerospace engineering) and the laptops they provide in their laptop scheme are using quadros, thats why i searched it actually
I go to a tech school for computer programming and web development, and recently a small group we work with got some computers donated (all computers they use are donated, so you never know what you're getting). We went over set them up, and two didn't work so we took em back to the shop to work on em. I opened them up, and almost had a heart attack when i saw Quadros just chillin' in there.
NOAA: "Dude, Call of Duty looks frickin awesome on this 5 grand card we got for weather simulations."
OMG this deserves a like XD
Massimo O'Kissed Underrated
National Oceanic and Atmospheric Administration?
Jonathan Macbeth , Yep :)
We used to play StarTrek on multi-miilion dollar mainframes.
with a little more practice he might defeat eminem
LOL... I was thinking the same thing.
ooooooooh
Omg am dying 😹xd
he looks like eminem
Not funny
Quadro is a cool sounding name though.
Uh. Not really Oo
Quadro means painting in Italian.
The name in itself is worth the money. Who cares about the card if whenever someone asks you which card you have, you can now answer with "Quadro"?
In portuguese quadro means frame.
Jhact007
The name itself is so sexy that when you say it out loud you grow abs and sexy mustache and all the ladies will be at your feet.
Amiriteorwhat.
"they got more money than you or me"
linus you have more money than me. I'm cleaning restaurant fryers to get the money for a 400 dollar pc and you put like 4 titan x's in a rig
+Oscar Arenas Pretty sure he just borrows the cards from NCIX
nope
+Oscar Arenas Not even just for the videos?
+233kosta for the 4 titan X rig, it was a rendering rig. he still has it if I'm correct
+Oscar Arenas Ah, well if it's his own he'd buy them, yeh.
lol "well geez, I wish my character's head could be calculated to 15 decimal places"
+Nathan Weisser I wish this every game I play :)
+Nathan Weisser honestly, that would be nice lol
Boom Headshot
What about sniping in arma 3? I need about 18 decimal places.
IKR, this happens to me ALL THE TIME!
In other news, my $9000 a plate fundraiser failed.
I don't get why, the plates were made of unobtanium! It was a BARGAIN!
I worked in a super complex military optical company, and I tell ya, they bought like 100 Quadros for all of their engineers so they can calculate some super colplex agile of glass and mirrors inside a gigantic replica of Hubble Lens just for practice before they actually make a new one for a future space telescope for the USA ...
Just looking at the lens structure would make you tripping at the technology level of it!, looks like a gigantic micro hive of super accurate cutted lens (that shi't is actually semi hollow from the inside, full of tunnels and such that you can't see from the outside), the actual view of perfection in your eyes.
So I guess that in those type of things, you can't except less than 100% accuracity, therefore, Quadro is the best choice, even when its 10 times the price.
Well..There go your taxes..
And we never heard of this guy again :O
I am a CAD CAM designer I used SolidWorks, Keyshot, Fusion 360, PowerMill type software I have a choice between GTX 1050 4Gb in gaming Pavilion and Quadro m2000 4Gb in Zbook workstation. please recommend me I am really confuse between these two machines because I need performance
@@MegaFaisalpk well, you probably would be better off with the quadro because of the high amount of vram and the other features mentioned in the video. But if it is much more expensive just go for the 1050 it has a similar performance.
@@MegaFaisalpk Quadro 100%
Turn on youtube CC and go to 1:10. So that's what Quadros are for...
Ej Ignacio i like large condoms
Ej Ignacio ok I admit it, I burst out laughing as soon as I saw it.
Ej Ignacio I use Trojan Magnum, anything else is just too damm small.
Ej Ignacio Omfg lmfao
Ej Ignacio So much useful information with one click thank you so much ^^
Well jeez, if only the position of my character's head could have been calculated to 15 decimal places...
Or that of the enemy.. If it was calculated to 15 decimal places I would have gotten that headshot....
KurtJMac's probably saying that.
Wow a stolen comment
Quadro cards are 10-bit whereas the GTX are 8-bit. So for Color Correction Grade Monitors in Cinema, colorists use 10-bit RGB Monitors. so you get the 10-bit signal from the cards to the monitors.
Voodoo 5 VSA-100 is also a 8-bits, then I dont need to have a GTX !
You should call this series "A Quick F.A.P."
lol
lolz XD hehehehe
A.F.A.P.
LOL
Se yall again at 2064
This was 7 years ago, and that was 2015
Where'd all the time go?
9now
It's pretty much spot on.
But here's an extra tidbit: in CAD applications, FireGL/Quadro performs A LOT better! (up to 10x or more compared to the consumer version, even though it's built using the same silicon) The difference all comes down to the drivers.
Traditionally (although this trend is slowly starting to shift) CAD applications draw stuff in immediate mode which pretty much meant that you had to make a draw call for each element you see on screen, and things were kept in system memory. This is in stark contrast to video games which grouped elements together to minimize draw calls and leverage GPU memory as much as possible. So GPU makers intentionally crippled immediate mode performance on consumer card drivers without having any detrimental effect on gaming performance (which consumers care about), but dramatically slowed down CAD performance (which consumers don't care about) - forcing us poor CAD users to pay through the nose.
Google for some choice keywords and you'll find some interesting investigations and findings, and you might wonder how GPU makers are able to get away with it at all.
sounds like there is a fortune to be made for some hacker/software developer to make custom drivers and sell it to CAD folk for a few hundred bucks
Son of a Zombie
I'd bet there isn't a dime to be made there. CAD professionals aren't going to spend any money on a hack that isn't supported by the hardware vendor or certified by the software vendors.
Folopolis CAD workers hate money? Maybe if the developer sold it to Autodesk, charged you $2000 for a license and rubber stamp "certification" you would like the idea better? it is clear now why NVidia is bending you guys over! lol
Son of a Zombie If you are a i.e. car manufacturing company, that makes yearly profit of X billion dollars, you ain't gonna use a hack to save a few thousand or even tens of thousands and risk stopping your production if something goes wrong. It's one of reasons why big companies rarely switch to FOSS. Not because FOSS is bad, but level of support they expect is usually lacking.
In my experience Xeon processors and Quadro perform a lot better than consumer grade "equivalent". I don't know if it's because of larger cache, ECC memory or better building quality in general. But when you work with big assemblies or complex shapes with something like Solidworks, Inventor or Creo consumer grade hardware inevitably fails at a certain point. I speak from many years of experience.
But Linus! Surely a Red Rocket-X is great for gaming! It's $8k!
Srs tho. deadmau5 has a Quadro 5000 and has nothing but problems running games on it when he streams from his studio. So yeah...
I mean come on! Best setup for gaming evar!
instagram.com/p/qFb4XfOx6X/
werethless12LP video cards priced above the titan z and black are OBVIOUSLY not meant for gaming. They are programmed or BUILT differently. They are obviously made for professional video content creation. Just because a video card can help CREATE a pc game DOESN'T mean it's MEANT to be played on that game.
Raymond Tyson Did you even read the second half of my comment?
Professional GPUs can play games, but they suck at doing so. Although, I doubt a Red GPU would be able to run any game..
That quadro gets less fps than my 780.
Why? Because drivers.
***** Yeah, iirc, certain 690s could be ran as a Quadro 4000..
This quickie videos are great. I only stumbled upon them 2 weeks ago, even though I've been a technician for years. They come in handy for easy explanations.
Changing the speed to 0.5 Makes linus sound like He's drunk AND STONED! XD
Thanks bro! I set it to .75 and now finally I can understand what he says, as I´m brazilian.
this made my day thanks you! lol
Hilarious 😂
😂😂😂😂😂😂
Techdrunkie
Im studying 3D design in University. They told me to pick up a Lenovo W540 workstation with a quadro. I also game and i knew that the card wouldnt push modern games, so i went with a Lenovo Y50 which rocks a 2014 maxwell based GTX 860m... Prior to university i would render high poly 3D scenes on an AMD APU based platform.. A super detailed scene at 720p, 2000 samples could take 4hours... With the CUDA function of my GTX 860m in the laptop, the same exact scene at 1080p, 2000 samples is about 40mins...
Its a world of difference.. Glad i didnt get the workstation.. It was bulky, ugly, and hardly could push 40fps on games like Warthunder medium settings 1080p.. I push a full 90fps with Warthunder on movie and its really enjoyable.
So note to anyone getting a laptop for 3D design or CAD work... The CUDA cores of a 2014 GTX seriously gets the job done + plays games where as a workstation with a quadro workstation laptop at the same price basically is not worth your while.
Not to mention workstations are a lot heavier, thinker and bulkier..
And considering that Adobe's applications and drivers support gaming grade GTX cards and a few Radeon ones, forking out more money for a quadro or firepro just seems like a waste of money to the average design student/beginner.
Hey Scotty I have the Y50 also and its great! I am curious which gpu renderer are you using? (you mention you went from 4hours to 40mins)
Thx!
see louder When you render single scenes in software like blender, use the GTX 860m gpu. It'll finish the job much faster. If you're doing a long video that includes multiple frames, I would suggest using the core i7 4700HQ CPU because it's more stable and is less likely to crash for larger scale projects. :D
Hope that helped mate!
The GPU will always be faster and infact it can usually do large scale projects too. If you dont want to take the risk of crashing it, still go with CPU for big video/animation renders otherwise for games, editing and saving video, photoshop, AI, etc.. - ALWAYS use the GTX 860m as the default render.
thx dude!!
***** i dont know how this works... how to let Cpu deal with certain works without Gpu involved?
"It's the cat's pyjamas" gonna start using that one
Dog's Bollocks pal
And you can add bee's knees.
Is nvidia Tesla too the same thing. Also make a video about random IOPS and similar terms used in ssd and hdd spec sheet
A Tesla is more for computations rather than physics or 3D rendering but it can be used for both of those as well, earlier models didn't even have any kind of display output iirc.
Thanks.....
Tesla cards are purely for scientific computation. They don't even have display outputs.
also I don't know about now, but originally I think Tesla wasn't a line of cards, but a peripheral cabinets that did calculations. (correct me if I'm wrong)
Thomas Alex They're good for things where you just need math, and not a video or anything like that.
as an aero/automotive engineer, I 3d model near daily, as do almost all of my coworkers. we use regular gaming nvidia cards (combined with pretty hard core processors, heaps of memory and decent ssd's), though we do use different drivers more optimized for openGL, since most of us don't do that much with directX. it saves the company quite a bit of money so we can upgrade the workstations more often.
rendering is still slow as balls if you want decent quality, but i only do it twice a year to generate shiny shareholder pictures. during rendering or simulation (FEM and CFD are even nastier on resources and i use them a whole lot more), i'll run my workstation at night, take it home during the weekend (allthough my electricity bill is not happy with that), bugger off to the workshop or coffee maker, or my absolute favorite; commandeer a second workstation and let that one do the hard work via remote access and/or mail notifications. you don't need 10x more expensive hardware to do decent work (not in my job anyway, might be different for the people at pixar e.d.), what you do need is decent hardware that you're allowed to pick yourself. tough CFD's and the likes should be run on blade farms or a linux box with a big AMD processor, fast memory and very good cooling
Fascinating
One of the other benefits on the precision of the quadros. Often we, in animation, deal with both very small numbers and very large numbers in a single scene. ie a mountain top with a character on it may be 2500000 units away from origin, and camera truck from far away may mean that we exceed the floating point precision of the scene. The double precision just helps to keep things from jittering as the viewport renderer tries to figure out what polygon is on top of another (called z-fighting), it also helps to draw curves because the tangency of a curve can be very small, especially if you zoom in on it, like in a grapheditor.
+aseredy Nice! Getting a pair of Quadro M5000 cards to play with, can't wait to see how well 3Ds MAX behaves with them as opposed to my GTX cards (which still kick butt in pure compute). Too bad Maxwell kinda curbed DP performance though...
Sweet! Maybe it was nerfed because Max is a mostly Games suit, and you don't want to be building scenes for cinematics that most game engines will artifact on. I'm a Maya guy so I'm not too familiar with Max's tech (used it in limited capacity years ago). Though, I'm not sure having DP impacts artist productivity all that much, but all those annoying little artifacts that happen with pure gaming cards definitely go away. So, you may just notice better over all UI performance instead. Though any investment in productivity is a good one, and Maya's overall user experience is greatly improved with Quatros. Max will probably be the same.
aseredy Yeah, I've seen some kickass viewport improvements in Maya when switching from a more powerful GTX to even a lower-grade Quadro, so I'm expecting great gains here too. Can't wait :D
PS2 has z-fighting issues unfortunately. :(
and they're too visible, even at a such a low resolution games run at.
Magic blue smoke?
Good thing I have blue LED fans in my case, makes it soo much faster, specially when it starts smoking.
another thing quadros have is much slower memory buses. i've had to talk about this to someone i know who was confused about GeForce and Quadro before. i had to explain it as "a GeForce will do things as quickly as possible and get you that 60FPS, a frame didn't turn out right? drop it whatever, a Quadro will do it correctly and keep rendering until it's done right"
another thing Quadros and Firepros have is the fact that they are much more hardy, i've heard tell of a computer science professor having two CUDA farms, one made of Quadros and one made of GeForce 780's, the 780's only lasted a couple of days before burning out from 24/7 use. the quadros were fine.
This video has been a great help in understanding the basic difference of two types of GPUs. But this is 5 years old and many things have changed. I mean new versions of quadro gpu are available and also Threadripper has made a great impact since then. Can you make a revised version this video
True
As always, love your quick-paced, no nonsense presentation style!
"Well jeeze, if only the position of my character's head could have been calculated to 15 decimal places..." This Linus is the best!! I can't stop laughing :D
For a future Fast as Possible, maybe a video all about RAM/Memory? Like what the speeds mean specifically?
They already did one I believe.
/watch?v=dWgzA2C61z4 On his main channel, but still relevant.
Seriously? Look at the older videos...
ALREADY DONE.
***** Ah, thank you. I think I had tried looking there before but I only started watching Linus' videos moderately recently and with the amount of content he puts out I may have just given up before I got to it.
RoximRox That's okay :) you can watch them all in just a few weeks! I've done it ^^
For 5 thousand dollars it better run crisis 3 at 400 fps and have 10k resolution support. Lol
+morraynumbah52 It won't, but if you're an engineering consultancy and work for the aerospace industry it'll pay for itself on the first job
+morraynumbah52 you just didn't watch the video at all
Alejandro Vargas
I did watch the enire vid, I was fucking joking.... Wait how else would I know the price of the fucking video card? Telepathy?
Alejandro Vargas
haha, nice to see that there's idiots that start shit even if it has nothing to do with them, esavages.
+Alejandro Vargas said no comedian ever
Why just GeForce vs. Quadro? Don't forget AMD's Radeon vs. FirePro ...
Also, about NVidia artificially lowering the compute capabilities of GeForce cards ... I really wish they didn't. It started with the 600 series, so the 590 is better at double precision calculations than the 780 Ti. There's no reason the latest high-end card should be artificially crippled to perform worse than a card that's two generations old.
do they really do that
The difference is mostly the same between FirePro and Radeon (per the video description), except that AMD cards are MUCH better compute cards than NVidia's.
I too dislike how Nvidia forces you to get their overpriced, super expensive "pro-grade" workstation cards if you want to get any decent compute performance out of your card.
Intel does this to CPUs two. :P A low end i5 is the same chip as the i7 with stuff turned off.
The reason they do it is so they can sell GPUs that wouldn't pass QA as Quadro (Because they would need all 8 floating-point units working for double-precision math and one of them is faulty in some way), but do pass QA as GeForce (Because then they only need 4 floating point units for single precision math). It's the same reason AMD sells 6-core CPUs and (Intel sells Dual-core CPUs) - They're actually 8-core (quad core, in Intel's case) CPUs, but 1 or 2 of the cores on the silicon is faulty, so the faulty pair is disabled with a hardware fuse and the CPU is sold on as a cheaper part, to save them losing money by not being able to sell the part at all.
Muhammad Alwan That's a shame ... I use NVidia cards because CUDA is so well supported and easy to use compared to OpenCL.
I guess I just need to learn to code better ...
1:31
This is why I love this channel...the production values
fuckin awesome!
It is called market segmentation. In cars, you have the same chassis and/or the same engine equipping models of completely different categories. In digital cameras, you had the same sensor equipping from a Sony Nex3 to a Nikon D7000.
So Nvidia is ripping off the rich and developing new tech at a reasonable price for us regulars? Now I respect Nvidia even more!
not just Nvidia, it's logic, you always go to the deepest pocket with a stiff price, because you sell less and you have to justified the cost of development to this particular product. and this go the other way, when you have a wide average customers who counts the pennies you lower the price to sell more.
Fucking robin hood man.
Linus left out the GeForce cards are made in way bigger numbers and stuff made in large numbers is always cheaper, so it's that + they have way more expensive memory and such + it's aimed at companies. So you probably don't have much to respect Nvidia for, hehe... ;)
(though I'm a long time Nvidia user, have been since FirePro's ran 3DLabs Glint GPU's and were made by Diamond. I switched to Nvidia when they were bought by ATI who messed everything up - because ATI)
To an extent. For example, starting with the 600-series cards, NVidia purposely crippled their GPUs to perform slower with double-precision calculations. The 590 is twice as fast as the 690 for computing.
asdfrozen And that's why I run 2x EVGA 580 Classified's for CUDA (and OpenCL) crunching. But man are they HOT! This box alone runs at $100 for electricity a month, hehe... :D
3:29 that background looks beautiful
Hey Linus, I'd love to see a video on what kind of a difference a high end mobo would make. Motherboards always seem to be almost like an afterthought in my builds, is there something Im really missing by getting a $130 mobo vs a $200+ one?
This is why the Mac Pro is so expensive.
You can get similar performance from a custom build PC for less but you don't get the goodies from professional (that's why its called a Mac "Pro" grade hardware as seen from this video.
LOL
AlbertoScrubington
You're probably calling someone autistic while 1) you don't know what autism realy is 2) he probably isn't . And by the way, I wouldn't be surprised that a mac does indeed have higher grade hardware. And no I am not an apple fan, far from it.
KingClaus of Yllin everybody has to make me feel bad about my self on the internet
KingClaus of Yllin The Mac Pro specs' might justify it price the best compared to any other Apple product, but all their products are still heavily overpriced.
Maxwaehrens Indeed they are
The reason the cost is higher from my experience, is that certain programmes require specific quadro cards purely for features within the programs themselves. I use Solidworks you need a quadro or FirePro to use graphics features in models.
I like to see a 6 year old video get recommended now bruh
I feel you
Haha, lol
I wonder if the topic us trending or something.
Man, when the best graphics cards came with DVI
Same
May GabeN bless your soul and your cats Linus
Have you accepted lord Gaben, king of the gamers and the nerds and the computer men, lord protector of steam and all it's sale events, first of his name as your lord and savior?
Marios5556 Yes I have.
Oh man I just fart as I read this comment comen on XD
First comment on this lmao
Okay Linus but I want to see a Quadro m6000 vs a Titan X Pascal fps comparison in multiples games. Does the Quadro actually get more fps than the Titan X or what? I know it's not meant for gaming but how well does it perfom in games?
there are no graphical drivers on the quadro, it actually can't be used for gaming.
sorry brah.
A quadro is made for companies, 1 card can power upto 6 monitors. I bought a nvs 290 for €7,50 and it works very good for games like gmod, minecraft, cs:go and it can power 2 monitors.
He made a video
Matthew Ihlan Can you give me the link?
he did that on his main channel, and the drivers do for games, he played Crisis 3 on it and a titan, they were the same, but the Quadro has twice the RAM at 24GB.
Quadro, Tesla, Maxwell, Pascal, GeForce, Volta, Kepler etc. - I think you guys are perfect to explain and put each one in the right box to make sense of it all. PLEASE!!
Hey now I was a painter for many years painting all sorts of houses and Interiors of those houses and outside of the houses and I will tell you that even that 5% difference is huge overtime.purdy is one of the best brushes on the market and worth the money
After watching this video, I've come to the realization that Linus would be totally screwed if he had no arms.
Guess we're all industry professionals now, with all the prices being where they're at...
$500 for an 80 class GeForce card. Gosh, I miss those days.
Techquickie hey guys. i always have knowledge on the videos i watch that you post, but its great that your videos help clear up confusion or misinformation that I may have!
Classic video title that tells me what I'm going to learn instead of using clickbait. Very nice.
Wait a minute, if butterfly can make hurricane on other side of world, what will do GPU with two or even three fans?
What?
ChRoNicBuRrItOs it's from chaos theory it says that it's possible for a butterfly's wings to cause storms. He's joking saying that a gpu fan would cause even more chaos.
Gary Merchant That is why we all need to keep our cases closed. Also, dust can become rather problematic.
@@GarrySkipPerkins but... the force from sliding the case side panel shut!
My dad a gamer. Bought a quadro card back in 2010’s for his video business but justified it that it could also play videogames
That road so beautiful... I wonder where it is...
+NenekTerbang I thought the same thing. Probably somewhere in Europe.
+Mario P. Cardenas Europe ftw fam. Looks like it could possibly be Norway. It looks very Norwayish if you know what I mean with the mountains and low population.
+NenekTerbang It's Romania, Europe.
+Mario P. Cardenas europe
+Tristan King you are right
Thanks for resuming all the differences. Much of the internet say only they have more memory, but this is incomplete.
from what i know now:
For image/matrix transformations they are good, and they have all the software ready to do it.
FPGA can do fast parallel processing, but they need to be programmed and cost more than GPU if doing image/matrix transformations.
CISC CPU (i core) are good to do random sequencial tasks and complex maths with large floating numbers (they have a co-processor).
RISC CPU are cheap but they need more steps to do the same task, they are in all tablets, PDA, digital mobile phones, GPS, and can be in some low end computers, cars, washing machines, routers, TV, and more.
I see, yeah I came across these quadro cards and was wondering about them, thanks! Also, I lost it when you showed the bottled water fabric!
If I recall correctly there was a regular GeForce which could be turned into a Quadro card by simply loading another BIOS. Can't recall which one though. It's a relatively recent card.
But really all of these features are for powerusers.
I believe that I've heard about it in various Geforce models...
not fully ... like no ecc and less reliable in obtaining more accurate results but u are 90 percent correct . same for some amd products . they increase rendering speed by a lot
hackad ay.com/2013/03/18/hack-removes-firmware-crippling-from-nvidia-graphics-card/ without the spaces ... this is just one example . there are older models also
Corristo89 I think the GTX7xx series may have had some models that applied to.
GeForce GTX 690 could be modded into a Quadro K5000 and GeForce Titan into a Tesla K20.
Yeah, its quite hard to justify why Quadro/FireGL has better performance over GTX especially in recent years. I have work in this two "world" with the same CAD software (office Quadro FX 4600 and home gaming rig GTX660 x2), it does make a different. First, if you are using CAD software such Solidworks, you will see the different instantly (highlight and real time rendering). While in other CAD software its hard to spot the difference, especially if you are a beginner. But ill give you a hint where to spot, try to rotate a part. The part edges on GTX/HD are blurred just like in games (Motion Blur), while on Quadro/FireGL they are precisely placed (less eye fatigue). And yes basically they just changed the memory type to ECC, implement a special driver which there are plenty of hacked ones, and stamp a higher price. Comparing with games, GTX/HD has faster memory since it does not have to correct errors and can be overclocked easily. Meanwhile for working, as Linus explained the brush, this are just tools. Once you know when to use this proper tool then you will achieve a better product other wise its just an overpriced solution.
Intriguing
The paint brush example is not very correct, what you pay for with a top quality brush is longevity more than quality. You can pick up a $6 brush and it'll work almost good as a $80 bush on the first day, but a month down the track the $80 brush wont have you wanting to paint walls with your socks.
Finaly a normal, understandable explanation of differences between the two! Thx!
The double-precision floating-point operation on consumer GTX Titan is fast (if not faster than Quadro and Tesla product line) like it never been crippled at all, they are comparable. Maybe the only difference between GTX Titan and Quadro-Tesla are ECC memory support and product reliability for continuous usage.
can quadro run hogwarts though?
do a video in exactly the same way - but with intel xenon vs ordinary
xeon*
+Casper sanderson It's much of the same.
* Their target market is businesses with lodsemone, so prices are high.
* They support ECC (error-correcting code) RAM
* More L3 cache
* Higher-binned
* Options with moar cores
He did on his main channel
+CrusheR BeaR Well, you also forgot to mention that there is no 18-core i7 available on the market :D
I used to use an EVGA SR-2 system with two Xeon X5650s for gaming (OC'ed via BLCK to 3.5 ghz)
they did fine, in fact you can overclock current xeons with haxx0r XMP profiles.
And there are engineering sample Xeon E5-2690 v4s on Ebay atm for $550ish, which are 20 core/40 thread.
Go nuts if you have the money =P
I have been watching this video since 2014 hoping to find something new that justify the cost of Quadro cards. Am i going to watch again in 2017 or coming years?
Diadem Raman I was trying to figure out if I wanted a quadro, but I'll just buy a GTX Titan
The reason for the cost being so high is quite simple. If you develop a product for consumers, that is built - let's say 1 million times - you can split the cost over those 1 million built products. But if you can only sell 10.000 professional cards, the cost of developing the product is split only over this 10.000 pieces.
So you got 100x as much for developing for the card.
The price for a product isn't just the cost for the resources you need and the production. Why does everyone forget that?
It's also the cost for development, better support for professionals, etc.
Why do you think professional operating systems are way more expensive than those for home users?
another difference: consumer products have roughly the specs u expect from them, they are always some veriations. a quadro gpu has the exact output as any other quadro so u can compare the results.
5000 dollars issnt much. the screens at my eeg-workstations are in that pricerange.
In addition to double-point computation, other gotchas are that you need a Quadro to provide a 10-bit/channel color signal to a wide-gamut monitor. Also need a Quadro to use NVIDIA's NVAPI. The NVAPI provides warp/blending capabilities at the OS level allowing you to create seamless projected desktops on both flat and curved surfaces.
i would just like to say that lynda is awesome, just starting my level 5 computing to get my diploma and this was a great since i havent really done anything in computing in a good 3 years except play games
When you set the speed to 0.5, it just sounds like Linus is drunk. But when you set it to 1.25, he just sounds like he's had some coke.
why am I getting this recommanded while watching videos on rtx 3080?
Can you do a video about the Nvidia Tesla K models?
Alexander Smith ayyy lmao. that doesn't even have a display output, it just calculates raw data. essentially a very expensive and beefy calculator.
MeGusta GameStation Uh, Exactly..
If you need dedicated CUDA cores for GPU rendering, then the Tesla is useful. Where are you going with this?
Alexander Smith
I was thinking that there wasn't really a point in doing a video about Tesla's on Linus' channel since he focuses more on gaming oriented products and misconceptions.
Most people haven't even heard of a Tesla GPUs and very little people even use one unless your gonna power a super computer or something. On top of that, they are much much more expensive and no average consumer can afford one at M.S.R.P
Don't get me wrong, I'd be interested in such a video, but most of the gaming community and consumers aren't interested and this channel isn't towards professionals, just gamers and enthusiasts.
I imagine one key difference here is that there's a much more in-depth development for the drivers and much better if not very customized support for the specific programs these graphics cards will be running. So they're more reliable for their specific use cases and they probably finish their workload faster. Imagine you're a movie studio head honcho and you know that by using these cards, your renders will get done considerably faster thereby allowing you to do more work. Or say the company is an oil or minerals exploration/extraction company that relies on a certain set of programs to map out where and how to dig next. The faster and more reliable your processes, the faster you get to digging or the better you get to timing your new projects in response to market conditions, etc etc etc. The 10x price tag can easily be paid for by that saved time not too far down the road.
great reviews, fast paced video and great points!
the world needs more reviews like this! well done
3:01 i want that framed on my wall
Actually, gamers are starting to need double precision graphics rendering. I know a lot of games that would benefit from it.
Some of you may have have seen a phenomenon called z-fighting, where two objects close together or very far away look like they're clipping through each other at random points across their surfaces, or which one is in front is flickering constantly. This is because the view transformation is becoming numerically unstable because your view distance is too long and/or you can see things too close to you at the same time.
Going from 32-bit floats to 64-bit floats eliminates this issue.
Also, calculating at 64-bit isn't dormant in GeForce cards. Rather, it's deliberately crippled. GeForce cards can do it, but it's about 8 times slower than it is on a Quadro - on purpose - making it unusable for scientific calculation or games.However, to be fair to NVIDIA, AMD is doing exactly the same thing with their FireGL line. Yes, that's why the Mac Pro is so expensive, in case anyone is wondering.
IshayuG No, Z-fighting happens because two or more faces of geometry are co-planar. Nothing to do with view distance. And if the vertices of two models are in precisely the same place, changing the precision of their world co-ordinates will not change their location.
Launch your favourite game engine on a PC running a Titan. Put two planes in exactly the same place with their normals facing the same direction. Notice the Z-fighting?
redchant
Two problems here.
First, putting in a Titan and launching your favorite game won't do anything. The game is programmed to use single-precision floating point numbers, and even though your Titan supports double-precision does NOT mean the game will use it, and so you're almost certainly not going to notice any difference whatsoever.
Secondly, I already mentioned Z-fighting from co-planar surfaces, but this is not the only type of Z-fighting that can happen. If the view transform is numerically unstable, you can get some points that were supposed to be in front of other points suddenly be behind them or right on top of them when they're displayed on the screen, and this is, by far, the most common source of Z-fighting. It's generally caused by too long view distance.
I use a pretty bad Work PC, with some pretty bad GeForce Card. Then a man came to me and aaid "Its dangerous to play with 5 FPS, take this" and he gave me a the Quadro FX 1500. I was so happy.
CraftorD Quadro is bad for gaming what the hell are u talking about.
***** ?
***** Its good for design and stuff, quadro is not good for gaming.
Like I said my previous Graphics Card was the Standart GPU of a Working PC and compared to the Quadro FX 1500 the Quadro is about 5x better
Sell it to buy a gaming GPU lol
In games, it doesn't matter if a bullet shotted hit correctly. But it does matter in industries if calculations go wrong slightly. which justifies the cost.
Gee, this video is from 6 year ago and I just got my Quadro K5000 today. Still works pretty awesome though.
are you wearing a shirt with a faze clan logo on it?
The company is five four not faze
i thought that too!
I was joking
Could you do a 2019/2020 version
Same. It will be the same in 2021 too.
video card manufacturers have been screwing the gamer for years!
You know is old when they show Lynda
fast forward 7 years and every GPU is made from unobtanium and Linus has joined an Elf Cult
Only 500$? 😭
I hate today's economy
Can you please stop saying FireGL, it's FirePro and has been for several years now. And on the same note, can we please stop sidelining FirePro, they outperform their Quadro counterparts on every level, except for the W8000 which is improperly marketed.
"Nvidia would be crazy not to take advantage of that". No, they'd be good moral people to not take advantage of that HAHAAA! :D ...but then again, there actually IS justification for the ridiculous cost, and they are not likely making more money by selling QUADRO Graphics cards as they do selling GeForce cards. BUT JOHN!, you're probably saying, The cost is MUCH greater! So of course they are making more money by selling those!! Weellll, I'd bet money against that and here's why. The consumer market is massive and consists of every single person in existence...just about...a large percentage of it anyway, but the market for 3D production that requires the quality and precision of a Quadro is MUCH much smaller. So what does that mean? It means that they have a big factory - it's not that simple, I know, just for example sake though - and all those costs to run that big factory including insurance, power consumption, employee pay and benefits, etc. have to be divided up between every card they manufacture to get an average cost per unit. And as is per usual, when you make a smaller batch of anything, it costs more per item.
This is all an over simplification, but I think it's probably the best reason they are charging so much. Of course, there is some greed in there too, but the ridiculous price is not COMPLETELY unjustifiable I'd say. Only somewhat :P
+John Meise Yeah, but they probably make a killing on each of these cards. That said, most businesses don't care as the time wasted waiting on rendering or w.e justifies the cost of paying 10x the money. They may lose more money just in time wasted if they are a multi-million dollar company.
Actually the Purdy paintbrush has a market leading, "leading edge" cut, without getting to paintbrush nerdish, the individual bristles don't splay apart
actually the pratice of disabling some feature and selling as a lower end model has to do with defect and process variation. The chips might be design equaly, but after manufacturing and testing, some might present defects that disabling features or setting a lower clock frequency masks
Does Linus wear a FaZe clan t-shirt ?
that's not faze lol
+N0SMAN its fenchurch
Great it's getting recommend after 6 year
can i sli a titan x and a quadro?
Yap
+Mr.Iron I think he meant to ask if he can SLI a Quadro and a Titan X together.
No. You can with either 2 Titan Xs together or 2 Quadros together, but not with one each.
Oh, and I forgot to mention, you can put them together on the same board. You just can't connect them together in SLI, they have to work separately, switch when you need to do something.
+charles chan Just like Chad said - you can't SLI them, but they can be on the same board. This is great if you drive your display and viewports on the Quadro - you get the acceleration there, which is what Quadros do best - while the GTX card will kick in when rendering - which is what it does faster than the Quadro cards. Best of both worlds, IMO!
+PIXELFLUX Not unless they both have the same amount of VRAM and are running on same firmware. You theoretically CAN turn an equivalent Quadro to a Titan X and vice versa (and it has been done with older cards).
This video is old and you're still taking me to school! Thanks
Excellent thanks for properly sorting that out.
Guys! I Have a question :D
Im making a PC and here are the specs:
Intel i5 Skylake 6400K (OR) AMD FX8320 Black
Kingston HyperX Fury 8GB
SandDisk Plus 120GB SSD
WD 1TB HDD
(Here is what im wondering about)
Zotac GTX 950 AMP!
NVIDIA Quadro K420
Cane theese two pair?
+YellowPeaches Nope :( sorry to say, you can only do SLi with two of the same cards, however I do belive you are able to pair 2 different AMD cards. :)
+HoxRC23 Wait... what about if i have 2 monitors? :D also whats a AMD counterpart to Quadro?
YellowPeaches Nope, still wont work with 2 monitors :(. But the AMD counterpart is the FirePro series. :D
Hox Tech
Il check :D
OK :D
you forgot the most important factor completely, quadros are much much much faster in certain applications thnx to optimized drivers. So the quadro cards pay for themselves in no time to the right consumer.
Beyblade Gforce!
Let it reap
The two driving factors are quality and purpose. If a game hangs in the middle there is no mayor harm done, restart and that's it. If you need to be working extended hours, 10, 12 or more and for a professional task that has to be delivered on time, The value of the card dilutes on the amount of jobs that are done with the card. When the difference between getting the job done and not getting it done is a couple of thousand dollars that value is spread across the amount of job that you do, if you are a gamer or casual designer, it is more efficient to buy the best dollar/value ration. That the reason even at mechanical task professional tools are several times more expensive that consumer ones, same applies to cooking, farming, painting and and task that has the be done. Possibly a consumer card has 95% of the capabilities of the pro one, but is that 5% the one that robs you the most margin on the job. I cannot comment on the price differences because that is market driven, but normally professional people pay premium just to make sure that the product is as close as 100% perfection.
I didn't knew this channel was this old
okay, brushes are the worst comparison, you get so much more if you purchase a more expansive brush.
+Skukkix23 Agreed. I still have my Purdy brushes. I will not borrow them to anyone. You never knew what you are missing until you try one and compare it to cheap and midrange brushes. It took use to convince me that I'm never buying a cheap brush again except for using it as a tiny broom on the shop toys.
I don't think he is saying professional grade products like the brush or card aren't better suited for their tasks, just that how much better they are doesn't equal the major price increase.
I NEED SOME UNOBTANIUM FOR MY PC!!!
but you cant obtain it. its unobtainable
Graffiti Cat
Thats the problem with unobtainium :(
use da future
So basically, the law of diminishing returns is a real thing?
+Kyle E Exactly.
Great Video 1:00 - 1:25 AKA Supply and Demand. + 2:15 DP FLOP= Expensive
Just some food for thought, the university i go to (I study aerospace engineering) and the laptops they provide in their laptop scheme are using quadros, thats why i searched it actually
swag
Shows 500$ GPU in 2014:”it’s the cats pajamas!”
Anyone ReMEmbEr??!
This video should be flagged for inappropriate content because it shows porn in the first few seconds of the video...
Mate, you convinced me - I subscribed! Well done.
I go to a tech school for computer programming and web development, and recently a small group we work with got some computers donated (all computers they use are donated, so you never know what you're getting). We went over set them up, and two didn't work so we took em back to the shop to work on em. I opened them up, and almost had a heart attack when i saw Quadros just chillin' in there.