AMD vs Nvidia for Video Rendering - Adobe Premiere and Media Encoder

Does CUDA or OpenCL acceleration have any meaningful advantage over the other when it comes to exporting video using current GPUs?
Video sponsored by Blue Apron! First 250 people get two free meals on their first order using this link: cook.ba/1TTYduk
Cooler Master link: linustechtips.com/main/topic/4...
Pricing & discussion: linustechtips.com/main/topic/...
Support us: linustechtips.com/main/topic/7...
Join our community forum: bit.ly/ZkLvE7
/ linustech
/ linustech
Intro Screen Music Credit:
Title: Laszlo - Supernova
Video Link: • [Electro] - Laszlo - S...
iTunes Download Link: itunes.apple.com/us/album/sup...
Artist Link: / laszlomusic
Outro Screen Music Credit: Approaching Nirvana - Sugar High / approachingnirvana
Sound effects provided by www.freesfx.co.uk/sfx/
Blue Apron Integration completed in partnership with Mediakix (www.mediakix.com)

[@Ayalaskin]

Am nvidia fan. But this is desgusting. When AMD won by a bit no comentery was made. When Nvidia won they pointed it out loudly.

[@KaNoMikoProductions]

+Ayalas - Gameplays de Hearthstone That's Linus for ya.

[@pillsburyrampage]

+Ayalas - Gameplays de Hearthstone I agree, a reviewer should be a bit more unbiased

[@ABaumstumpf]

+Ayalas - Gameplays de Hearthstone Gotta agree. But at the same time i can kinda understand - the r9 290 has beaten the R9 390X - something went horribly wrong there.

[@peanutjam]

Noticed it too, less strong AMDs crushed stronger Nvidias and it's "all the same".

[@avi_mukesh]

Exactly what I was about to comment. I just don't get what Linus finds so bad about AMD

[@HardwareCanucks]

Cool test and interesting results. Would have loved a mention and testing on OFFICIALLY supported GPUs by Adobe. Rumor has it, GTX 780 (Officially supported by Adobe CC) is faster to encode/render files than a 980 Ti (not on a supported list) by almost 10%...

[@Bluxxa]

K

[@Yantonio7]

+HardwareCanucks so, in english means.. some older powertanks videocards like the gtx 780 show more performance than the newish... 980ti .. . :v

[@ChipMalfunction]

oh, hi! didn't notice you till now

[@luckyqualmi]

+Xin CPU is better quality, but slower. GPU is enough for youtube & small image films, etc. but if you want to do projects for TV, cinema, ... you should go for CPU encoding if possible.

[@ChipMalfunction]

***** He's in Canada currently

[@timmilgram]

Would love to see an updated version of this, with newer cards compared to the 980ti as the baseline

[@midyanshani]

Same

[@deebonash4487]

Same

[@jstuyfzand5467]

5:26, where is the fury?

[@qwerty-fc9ix]

+Joep Stuyfzand check bottom right corner.Fury shit the bed for some reason.

[@jstuyfzand5467]

koko power Lolz

[@marcorodriguez7748]

+Joep Stuyfzand it says on the bottom right corner, the r9 fury didn't make it. Lol

[@MazeFrame]

+Marco Rodriguez That does not make any sense. They should have repeated the test instead.

[@bLuDrGn2011]

+MatzeGamer you do know they usually run the test multiple times and then if any fails, they try to fix and get it to clear the tests.

[@smyalygames]

What about Intel Integrated 4600 HD? Must work better than a titan x.

[@IkHeetMich]

agreed

[@francooliver7546]

+Smyalygames It works great with TETRIS.

[@Warcheiftan]

+Smyalygames affirmative

[@daKingDoom]

+Smyalygames oooooooo ur smart never thought of that before, get this man a taco and a medal!

[@smyalygames]

+YasperOn “YasperOnTheKeyboard” TheKeyboard Imagine you are building your pc, you bought a i7 4790k, and a Titan X. When you finished and you are installing Windows, your mum screams at you for buying a titan x. THIS CPU HAS A GPU INSIDE IT WHY DID YOU BUY THAT TITAN X?!?!?!?

[@azlanbukhari926]

Linus being an Nvidia fangay as always

[@theintensegamer178]

ikr.

[@easymac9100]

he said he doesn't recommend one brand over the other for this usage. you clearly didn't watch the video, so why comment?

[@azlanbukhari926]

+Easy Mac Well if one product is clearly above the other one in the graphs (the product being AMD cards) and the oddity of the fact that only one pure graphics intensive test was highlighted in which Nvidia had a lead and Linus saying he doesn't think one brand has an advantage over another while showing a zoomed in graph of on specific Nvidia trumping AMD graph all the while the Fury X which was standing heads and tails over the competition was suddenly missing in the last graph. JayzTwoCents made a video not too long ago about cloud computing where he highlighted the fact that AMD is the preferred brand when it comes to video editing none of which being mentioned clearly highlights the Nvidia biased attitude of LTT, all the while in a recent WAN show Linus constantly had a "bitchy" attitude towards Polaris and said the only reason AMD has HBM cards is because there is so less HBM that Nvidia can't possibly incorporate it in a major release while AMD with their low sales can.

[@shottybeatssword]

+Azlan Bukhari Of course the Fury X beats a 980 Ti. If they stuck a Titan X in there, it would crush the fury.

[@DanielRichards644]

+Easy Mac he also said the super high end cards weren't worth the added costs and that the 970 and R9 290 where basically the SAME PERFORMANCE, and a little googling shows basically the same price (which is nearly half what a 980ti goes for FYI)

[@Flying--T]

I would like to see a Quadro and FirePro in these Tests

[@oudomsambatchorn851]

same

[@stewie_]

+FlyingT yup, same, those are the real deal!

[@oudomsambatchorn851]

yeah

[@mathiasronnholm5010]

+LinusTechTips yeah

[@andreaswbj]

I agree only pro cards makes sense in a work reated task.

[@ScuzzyAmv]

"Showed us what we were finally looking for.." Nvidia finally wins a test. Pathetic Linus, seriously..

[@MarioDragon]

AMD fanboys fish hard

[@whitewolf1477]

lol linus gets magical hair cut at 6:27

[@mikailn3892]

+whitewolf147 lol XD

[@kquote03]

:)😆

[@Neumah]

+whitewolf147 Helloo... it's an Adobe effect. GPU assisted.

[@WetaMantis]

+whitewolf147 doesn't work for me...

[@xpjohn]

+whitewolf147 neat

[@BloodravenD]

Between all the server crap, and the last couple of videos, I'm unsubbing. Linus is a joke.

[@svenvandam5251]

ok bye

[@kdoe1305]

Bye. ¯\_(ツ)_/¯

[@bigpod]

+BloodravenD server stuff is cool at least to me bc normaly we cant aford such stuff so is fun to watch someone so smart tinkering with it and at the same time finding ideas how to get this sort of stuff on cheap

[@PlanetCoolMinecraft]

These comments make me feel like Linus should just quit. "Linus we need something simpler" is accomplished and he just gets "Linus this is shit", then he goes on to do high end machines and he gets "Linus we need something simpler" again.

[@djdelarosa25]

Cool. Good riddance.

[@MiDnYTe25]

As an owner of a 980 Ti, I would like to see less Nvidia bias on LTT.

[@arnoldderuiter886]

+MiDnYTe25 Thank you. Amen to this, Nvidia owner here too, will be switching to AMD GPU besides my FX8350. But seriously Linus, what a fanboyism for the green team. Not cool.

[@MiDnYTe25]

Arnold de Ruiter You might want to upgrade your CPU to at least an i5, in case you are getting a beefier GPU (R9 390/970 category and up mostly) not to get bottlenecked in games though.

[@tommor5519]

+MiDnYTe25 linus has always loved nvidia, ever since the launch of the 680, it seems that he can say no wrong about them.

[@chasetherightenergy]

Why is there such a thing as "nvidia" and "AMD"-fanboys. they're fucking components. just buy the best for your needs and budget

[@ian1064]

agree

[@soulshot96]

+bAWSM Buuut, Nvidia is EVIL, and we have to help the underdog Ayymd or they will fail and evil nvidia will take over da werld.../s

[@correzzana1992]

Best comment ever. Owned both ATI (in the past) and Nvidia, always been satisfied. Don't get why too much people get this rival thing soo hyped

[@AnimagicMissile]

+bAWSM While I agree that being a fanboy is silly (I'm personally using nVidia at the moment because it better suited my needs even though I like AMD more as a company) video cards can be very expensive. People don't want to feel like they've wasted such a large sum of money on something that isn't the best at what it does. So they'll defend it blindly to maintain the warm comfort of their delusion. It's understandable, really. I just wish they didn't have to act like such douchebags about it...

[@Folopolis]

+bAWSM There is something to be said about corporate citizenship. Other than that, it is the nerds' equivalent of a favorite sports team.

[@deXXXXter2]

For linus when AMD is faster it doesn't matter, but when Nvidia is 'only here we can see the difference and of course it is for the company that sponsors us, because they make beefier cards'. And before you answer - it doesn't matter if card is faster thanks to better specification or for example better driver optimazation like it may be in first two tests. Faster is faster.

[@jarnee1]

+deXter2 I completely agree, he seemed to go through a series of workloads until he found a scenario where Nvidia was quicker, I guess you have to justify your sponsors product placement somehow, how then, is he going to explain the descrepency when AMD's DX12 performance shits all over Nvidia's cards...wont be able to hide then.

[@ZeroSins]

+deXter2 Anyone else notice that he removed the R9 Fury from the list? I sure did. Why not have the R9 Fury X?

[@akp1983]

+deXter2 Be careful you'll be called a fan boy just for noticing the obvious. It's a joke how much it has been happening. Come on Linus let us see who sponsors you, and how much money they give you. bank statements will do for proof.

[@prankedpulsive620]

+Zero Sins AMD was just too good

[@SuperWasara]

+Zero Sins It didnt passed the test..

[@mephlesn5020]

So when AMD wins by almost 2 minutes its no big deal? But when Nvidia wins by just over a minute and a half its apparently a huge deal and has to be pointed out? No bias my ass.

[@ITTFami]

AMD wins by a small margin: "your video card doesn't matter" AMD wins by a slightly bigger margin, the CPU-only benchmark takes double the time: "you don't need a GPU" Nvidia wins by a small margin, Fury series isn't even mentioned: "AAAND BOOM, THE BEEFIEST VIDEO CARD HAS AN ADVANTAGE" Seriously, Linus, it's getting too noticeable by now. I was really skeptical about you being an Nvidia shill, but this just confirmed it.

[@LocalAutist]

Future calls in, they changed settings to benefit nvidia

[@hfzhoez46456]

First Test: Fury beats 980ti by 30 seconds. GPU doesn't matter. Second Test: Fury beats 980ti by 75 seconds, 290 beats 980 by 80 seconds. GPU doesn't matter Third Test: Conspicuously missing Fury, 980ti beats 290 by 57 seconds. Shell out your money for a 980ti!

[@yolsclassics6347]

watch the whole video. Linus specifically states he can not and will not recommend a brand.

[@KaNoMikoProductions]

5:03 "Showed us what we finally looking for, the scenario in which Nvidia beats AMD." There, fixed it for you.

[@beigebox1990]

+Queen Bitch KaNoMiko You should get all the likes. So true that they left Fury out of the results.

[@robertoliver8289]

+BeigeBox Why do people like you always think they have something against AMD? They aren't trying to ruin AMD. If anything AMD is doing that itself XD. In all seriousness. Where is the love?

[@BDHtP]

Has nobody else noticed that the Fury is suddenly missing in the last bench? Could it be that Nvidia would have lost?

[@beigebox1990]

Robert Oliver Linus has a history of that, and it's plainly obvious in this video.

[@robertoliver8289]

BeigeBox What has he done in the past? I never get this war between card manufacturers. It is just as bad as the Xbox One vs. PS4 argument. Who cares?

[@avi_mukesh]

Wtf man, at least give AMD some credit where it deserves it. This was so biased.

[@TheFPSPower]

Export to cineform - 1080p -- AMD wins by 33 seconds, no mention Cineform to 4K H.264 --AMD wins by 80 seconds, no mention Chroma Keyed footage 1080p -- Nvidia wins by 57 seconds, Nvidia is apparently clearly better... (I know the Fury pooped out) Linus, its not the first time you mention absolutely nothing about AMD and go on to talk about how good Nvidia is, I'm starting to look at you as someone with bias information about graphics cards, and I don't even care if its about $$ or not, I just don't like it.

[@OGPatriot03]

+TheFPSPower I've never had an issue with my 2 FuryXs. Seems odd that linus leaves it out, then screams about a win for Nvidia.

[@TheFPSPower]

Patriot 03 I think the Fury X he has gave him problems before cause he dropped it, not sure if it was replaced, it was long ago.

[@94fleetwood49]

I don't mind Linus videos, but for gpu comparisons; I try to totally avoid his biased fanboyism.

[@noobasdfjkl]

First Test: Fury beats 980ti by 30 ish seconds. GPU doesn't matter. Second Test: Fury beats 980ti by 75 seconds, 290 beats 980 by 80 seconds. GPU doesn't matter Third Test: Conspicuously missing Fury, 980ti beats 290 by 57 seconds. Shell out your money for a 980ti! Final Conclusion: AMD vs. Nvidia doesn't matter.

[@Krisztian5HUN]

NvidiaTechTips

[@TeemoForLife]

Nvidia>AMD

[@TeemoForLife]

+Stunt Growtopia Gtx 480 > R9 390X

[@VivaRonaldo19]

+Teemo4Life g41 express Chip set>>>>> gtx 980ti 😎😎

[@chirpphixel6251]

+Teemo4Life GeForce 256 >>>>>>>>>>>R9 390X

[@Krisztian5HUN]

+Samuel Lim Savage 4 16MB PCI >>>>>>>>>>>>>>>>>>> GTX 980TI hexa SLI!!!

[@Hon2838]

5:26, r9 fury suddenly disappeared lol

[@paul1978g]

+AMD and Nvidia GPU vs Games Yes, because if you read the text bottom right, it failed to complete this particular test.

[@MrBoggy101]

+AMD and Nvidia GPU vs Games In the bottom right it says "R9 fury was behaving strnagely and did not complete this benchmark." Also, that's a direct quote. They spelled strangely wrong.

[@edwardhobart-smith]

+AMD and Nvidia GPU vs Games very disappointing! even if the fury was behaving strangely.. they should take the time to fix it and definitely should not shout out nvidia as the clear winner in that case!

[@paul1978g]

Edward Hobart-Smith Fix the dodgy hardware? That's not their job, that's AMDs. If the sample AMD sent is faulty, it's not down to the reviewer to fix it. If it won't complete the test, there's not much Linus can do about it.

[@LinusTechTips]

+Edward Hobart-Smith it's like people don't even listen. I said in that test more powerful GPU hardware showed it's muscle more. The winner is the person who recognizes the difference in performance between a $200 and $600 video card in that test is not a big deal for a task that their system will do once per day AT MOST even if they produce daily content.. I didn't draw a specific conclusion about and vs nvidia because neither had a clear advantage

[@remiltonsilveira7314]

I want to thank you guys so much for this video!! And if you have more tests like that, i would be glad to watch, thumbs up and favorite it!

[@rjodesigns2782]

we need another updated version of this. once the AMD release the 6000 series vs NVIDIA 3000 series

[@saveraptor2209]

LOL WHERE DID THE R0 285 COMPAT/FURY GO AT 5:15, the foot note is a joke, right?

[@No1TypeC]

Why would it be a joke?

[@mattmellor4597]

Strnagely Kek

[@digfidge]

+SaveRaptor Operator error.

[@kekkonenprkl]

Apparently there is some kind of bug with GCN 1.2 cards (Fury, Nano, 285) in that use scenario.

[@darkzsniperz3914]

look at the bottom right corner, the card was behaving weird.

[@vanguardas9927]

AMD cards won more benchmarks than Nvidia, but Linus says it DOESN'T MATTER. If Nvidia wins he'll ask you to buy 10 Titan Xs. Fucking biased sponsored channel.

[@yosyp5905]

The difference was like... 20 secs?

[@Txx00xic]

Benchmarks with multiple inconsistencies. Please check your statement before looking like a rabid fanboy.

[@vanguardas9927]

Yosyp The point is... when AMD wins it's "just the same"... if Nvidia wins "they win hands down", even it's by the same margin as AMD

[@Txx00xic]

+Vanguardas Vucava The margin between these cards in these tests are very small at best winning by a few seconds on either side to call Linus an Nvidia shill is pointless when he simply favors the cards that have had more horsepower for the last decade in benchmarks. Not saying AMD GPU's are bad as I rather like the 2xx series (the 3xx series is complete garbage tho) and their Fury line up the cards simply cannot keep up with their Nvidia counterparts. Stock clocks? Yeah they hold their own and do better in some cases but when the cards are overclocked? Boom, the gap widens far and AMD's cards can't catch up.

[@vanguardas9927]

Txx00xic We're not talking about gaming performance here. Last decade? Tell me more about Fermi? lol Maxwell doesn't win everything either. R9 380 R9 390 R9 Fury and R9 Nano are better deals than their Nvidia counterparts.

[@lilturk726]

In tbe first video I jokingly said I was surprised Linus didnt fill the server with 980ti Strixs. Looks like I wasnt too far off lol.

[@lilturk726]

The*

[@vHypz]

Why Strix series cards from Asus?

[@lilturk726]

Nigel Linus likes Asus products, and they're really solid cards.

[@factsverse9957]

+lilturk726 Yes. Linus likes ASUS products and is biased by NVIDIA.

[@chickentomato9584]

If money is no object nvidia is better

[@StoneMountain64]

i just got a zotac 980ti, it is heaven so far :D

[@echox000]

StoneMountain64 *I just saw your comment on Ninja's news interview*

[@bluehabs]

what a time to be alive

[@xplodnow]

Alright guys, lets look at this objectively shall we. Lets just compare the most commonly used two cards at similiar price range at release. The R9 290 and GTX 970. 1st Result: Export to cineform - 1080p (amd faster by 22s) R9 290 : 418s GTX 970 : 440s Linus," The only real conclusion... you are going to be getting a very similiar experience" Fine. That assessment is agreeable. 22s is not much. 2nd Result: Cineform to 4K H.264 (amd faster by 117s/2mins) R9 290 : 1062s GTX 970 : 1179s Linus, "..which might lead to the conclusion which graphics card you choose doesn't matter.." Really? 2mins is not significant? Even when comparing with other cards, all amd cards (excluding r9 390x) are significantly, at least 1 min, faster than Nvidia. But this doesnt mean anything you say. Ok sure. Linus, "...except that our third test using...showed us what we were finally looking for, the scenario in which a POWERFUL VIDEO CARD demonstrates.." 3rd result: Chroma keyed footage 1080p (Nvidia faster by 5s) R9 290 : 279s GTX 970 : 315s Linus, "..so the beefiest gpu, in this case, a GTX 980 Ti in this system COMES OUT ON TOP, with the LESSER CARDS performing about AS MUCH WORSE as you would expect given their overall horse power." So only this result matters? Screw the first 2? Lol. Even when comparing the 980 ti to 290 the difference is only about 1min. He said it himself, this is what they were "Finally looking for". The R9 Fury also suddenly goes missing from this result. Pure coincidence surely. ;) Granted in conclusion linus said it doesnt matter which card you get in the end cos they all perform similiarly. The conclusion is sound and no card performed significantly bettter than the other. Its just that the way the information is presented and shown laughably shows the underlying bias LMG has. Draw your own conclusions from what you think. This is just my opinion.

[@Mrmhmd1993]

thumbs up amazing comment

[@Mattia_98]

+Snaiper Suji Mayne well I belive him. the only things I would have liked to have seen were quadros (but maybe they hadn't any) and Intels iGPUs (but maybe the CPU hadn't one).

[@LinusTechTips]

+xplodnow let's look at this objectively. Everything I said was true. More powerful cards performed better and from an opencl vs cuda standpoint (in the specific workload we are investigating,which is premiere) it doesn't really matter which team you choose. I'd have loved to run the fury, but I wasn't going to waste more time on it when we already had the conclusion we needed.

[@xplodnow]

+LinusTechTips I agree. Everything you said was indeed true. And i have no problems with that. Its just that the way it was presented/scripted makes it sound a lil bit biased. That's just was comes across to me as a viewer watching the video. Maybe you didn't realise it, maybe it wasn't intentional or maybe I am reading into too much. Small things can give big impressions is all I'm saying. Maybe be more aware next time? Haha. But hey, it was a interesting video, big fan of the channel.

[@A.C.E.]

I like how at 5:15 when he said "we were finally given the results we were looking for," nvidia cards held the top 3 places. However when AMD won in the other tests by fairly significant numbers, "the graphics card you choose doesn't really matter."

[@rhstovall6386]

came to see the nvidia bias and got the nvidia bias.... not sure why I'm surprised

[@revup67]

excellent review, wish there was a 2019 version

[@Azivegu]

Thank you for doing this video! I have been wanting to see this for some time. I was wondering if you guys could also try the same test with Vegas/Movie Studio as well as I have heard that OpenCL works much better on it compared to CUDA. Also, I know it might be difficult to obtain them, but could you also give a rendering video where you compare the Quadro and FirePro cards?

[@AKShamu28]

Not hating on the ad, but it seems rather weird that this channel is sponsored by Blue Apron. The sponsorship people were probably like: "I really like this computer channel, but how do we get them into cooking..." And then the pizza warming computer video came out and the sponsor people were probably like, " GOGOGOGOGOGOGOGO! Now's the time!"

[@magzire]

what ever pays the rent

[@alishershermatov6052]

think about it. This channel gets a lot of views and what does everyone need to survive? food so why not put an ad for a food company in this video.

[@postalsamGAMING]

Linus, a message from England, you no longer have permission to say "bloody"

[@TheRealSarcasmO]

So need to revisit this with the major video cards and different editing programs

[@xSmym]

5:11 No R9 Fury for some strnage reason

[@Kyance]

+x Smym -It was probably better than the 970 anyways- nvm just read the bottom right corner lol

[@begobogo2583]

Bottom right of the screen, they put a foot note

[@benew4555]

+bego bogo These idiotic unboxers should include gtx 780/780ti if 290 are there.

[@Coaster_Contrarian]

+x Smym Yeh, I noticed that. Weird.

[@Maindric]

I was thinking the same thing.

[@TheFireNeox]

Amd wins: "There all close, doesn't matter what you take" Nvidia wins: "My god its incredible how much of a difference there is, how amazing"

[@Ephremjlm1]

I literally just bought the full Adobe suite like 2 days ago so this was super helpful.

[@EposVox]

Do you have any Quadro cards to compare this to, to bust the myth?

[@ElsK81]

+EposVox or some firepro

[@LinusTechTips]

+EposVox Unfortunately, not :(

[@EposVox]

LinusTechTips Damn. Well if you get ahold of any, Quadro vs. normal high-end cards might be cool to see for AME.

[@coopergates9680]

Well, FirePro S10000 is basically an HD 7990 with more memory (and ECC memory), lower clocks, and 1/8 of the stream processors disabled.

[@10100rsn]

Any reason the Fury was left out of the final graph?

[@TitanOne1337]

It behaved strangely and didn't completed the Benchmark

[@zJericho101z]

+10100rsn I can think of 8 reasons....

[@TitanOne1337]

+Opt1mystyq Which?

[@10100rsn]

Dennis Bondar Thank you Dennis. ;)

[@TheKratoKiller]

AMD are less efficient run hotter and less powerful why would you buy one

[@janski1218]

Last thing your Nvidia overlords want you doing is commenting on when AMD comes out on top 2/3 tests (even if it is only slightly), they might stop giving you free stuff if you did that :O

[@blade00362]

+janski lad CUDA proves it sucks ass once again

[@Poprotta]

+janski lad amen.

[@kylehues8465]

Been waiting for this!

[@ragingroosevelt]

Interesting video! So, this week's sponsor is pretty cool. I was lucky enough to be one of the first 250 to try it. Pretty interested to see how it goes.

[@Derively]

I would like to see some Quadro and FirePro cards too, i'm curious how much difference they make.

[@hellNo116]

+Derively you know that in terms of pure performances are basically the same with their equivalent consumer one, right??? it is the same chip under the hood. so it shouldnt be a meaningful difference.

[@casualtake1497]

+ilias 1161 except Quadro or Firepro usually have double vram

[@boxadmiral]

+ilias 1161 But you can do special things with them, or something. I'm certainly no expert

[@casualtake1497]

Mike Peabody yeah, Quadro and Firepro equipped with double precision that useful for rendering in 3d

[@hellNo116]

Said Alghifarix well the r9 290 was faster than the r9 390 and 390 has double the vram. i dont think that will affect the performance in any way. also no mater what i dont think they will be able to convince neither nvidia nor amd to send them those cards. they have no real use for them since they are not useful in video editing from what i understand but they are for 3d animation

[@mr_NAK]

So it only matters when nvidia is faster right?

[@TheHobohobbit]

he basically said that of you are doing any of the first benchmarks the card you choose doesn't really matter. he also said the last benchmark was the only one that showed expected results (more powerful = more better) and that it still doesn't really matter (still much faster than CPU alone)

[@wUFr]

+MrGalotsa well at least he said he don't recomment nvidia or amd at the end, just "get gpu" ... but seemed pretty weird, when he didn't said a word when AMD was faster just for all the AMD haters, i am using Nvidia card...

[@Krisztian5HUN]

+MrGalotsa as always on this c$hannel

[@imranakhtar4478]

+MrGalotsa hahaha, yes it seems linus is nvidia fan boy b/c he get them free of cost

[@TheHobohobbit]

+Imran Akhtar he also got the AMD cards free, that's how this whole business works.

[@TheRawSound007]

You should make another video rendering comparison video with the new Nvidia and AMD GPUs out there. Would love to know how things have changed over time.

[@wolfsingleton]

Thanks Linus, this was exactly the kind of testing I needed to see. Of course, seeing that case jammed full of cards made me drop my mouse.

[@skillzforMC]

what about sli or crossfire?

[@EricTanGH0ST]

+Thijmen van den Berg I don't think it can do that...I recall Linus mentioned it earlier while building it.

[@charlesjoyce3933]

+Thijmen van den Berg Doesn't work like that for rendering...

[@GhostsGraphics]

+Thijmen van den Berg It will still only use 1.

[@EricTanGH0ST]

But guys bigger question is, this Mobo doesn't even freaking support SLI / Crossfire.... If Im not terribly mistaken. Linus made a different video about this before, this is a Server board not a consumer mainstream one...

[@areskrieger5890]

+Eric Tan You are correct, it's a specialized server board, and servers don't care about sli or xfire. So it's almost never a feature on those boards.

[@Hankextreme]

No Fury X in the third benchmark ? Must've been an "accident". AMD owning nVidia cards in the previous 2? Let's not talk about it. The majority of people are still using nVidia cards for any kind of "professional" use and trash on AMD. lol

[@Txx00xic]

>implying AMD doesn't deserve to be trashed on

[@dkis8730]

5:11 in the bottom right corner it says "the and r9 fury preformed strangely, and was not in this test"

[@taylorcrouch1401]

I don't really agree that there was a bias. When AMD did better it was by like 30 seconds out of about 6 minutes or about 100 seconds out of about 16 minutes. However, when Nvidia did better, it was by what 50 seconds out of 4 minutes. That is proportionally a larger increase, and because that focused on a specific task, he highlighted it for the people who work in said task. Leading to the conclusion that, for the most part, the card maker doesn't matter, what matters is that you have a video card.

[@LinusTechTips]

+Taylor Crouch thanks so much for "getting it".. The comments on this video appear to come mostly from people who didn't even bother to watch the conclusion or weren't listening..

[@-yttrium-1187]

+LinusTechTips I guess you should accept that people can't interpret the numbes by themselves.

[@genestarwind928]

+LinusTechTips Ohh we 'get it'.... AMD clearly wins, so it doesn't matter which gpu you get.

[@shauryathakur5649]

+LinusTechTips agreed that the improvement was of about 25% and in conclusion you did say that Nvidia vs amd really didn't matter. But the fact that the r9 fury which was ahead of 980ti in all other tests (even by a narrow margin) was suddenly 'behaving strangely did not complete the test' in which going by the trends it should've been on top is seems fishy. Given the fact that all the gpus in all of your editing rigs are Nvidia ones and the fact that you are planning to put 8 980tis (which are most likely to be sponsored) doesn't help either.

[@NekoCoaster]

I don't see anything shady lol, linus did a great job and I don't see anything wrong with it.

[@PoetPlays]

Unfortunately, Premiere and Media Encoder aren't really optimized to render through GPU - in that they only render "effects" through GPU, and do the bulk of the work through the CPU. This is why the tests with chroma key or any transformation effect that change the bulk of the footage will see better results, and why having a beefy card isn't worth it for most of the workload done through Adobe products if you aren't using those effects regularly. I've heard, however, that products on Apple like Final Cut and Compressor rely much more heavily on GPU rendering, and I'm curious how much faster they preform in comparison to the exact same tests run on the Adobe products... but that's obviously a software comparison and is independent of the hardware at that point.

[@ikue]

Linus don't you mean the R9 Fury smashed the 980ti in the benchmark where it's "results went missing" or whatever? Good one, at least the sponsors are proud lmao.

[@Txx00xic]

Nice tinfoil hat you got there

[@samhewitt4940]

+Txx00xic Its funny because its actually true that linus is known to be incredibly bias towards nvidia.

[@Txx00xic]

+Sam Hewitt can't blame him, AMD has been flopping for over a decade.

[@bell___]

+Txx00xic only on the CPU side of things.

[@Txx00xic]

+Belle Rotmg and their other failed projects like mantle. Their GPU market is the only thing keeping them afloat atm and I understand why as their 2xx and 3xx cards are good for the price, They have to really hit it out of the park with Polaris and Zen or it's another 10 years of mockery and rabid fans biting at each other over which card is more powerful.

[@nanomage]

Where are the Quadro and FirePro cards? I'd be really interested to see a comparison of professional digital creation hardware vs consumer ditto (enthusiast or otherwise). Would for example the M6000 Quadro stomp a Titan X or would the differance in performance be too minute for small video creation and editing businesses to make financial sense?

[@TowDow3]

+nanomage thats the kind of tests they should be doing. workstation gpu v consumer gpu

[@Pertamax7-HD]

Nice sir linux

[@HeyDan1983]

Oh Linus you just make me hate nvidia more for your bias.....

[@noscopesallowed8128]

Who said Linus is out of ideas? Videos like this are why LinusTechTips is so great!

[@InitialDAmine]

+Max McTabish except when the results are biased, aside the fact that he didn't use a Fury X (since he is using a 980ti) You can see who wins what in each benchmark, but the only benchmark is the one where there is no Fury and only a 290 and 390x vs 980ti and 980.

[@techblade837]

+InitialDAmine he didn't use a Titan x eather

[@GriggsHw]

+InitialDAmine Bottom right corner, Fury behaved strangely and didn't complete results, but I guess for you guys you can just easily find a reason that Linus is Nvidia biased isn't it? It's a wonder why all these people called out Linus for being biased but still stay around since they think Linus is "Nvidia piece of shit". Linus wouldn't even be talking about AMD or let AMD wins at all in any benchmark if he is really biased.

[@Arudis4x]

+TechBlade 980ti would probably beat it since it wouldn't throttle as much. But there is a lot wrong with his methodology.

[@strajkz]

+TechBlade Titan X isn't the direct competitor to the Fury X, if they used a Titan X they would have to use the next gen dual GPU that AMD is going to launch.

[@IAmRyunis]

Now, without starting more drama about Linus being biased etc... Where the fuck did the R9 Fury go in the last bench?

[@iamoctonate]

lol, it went so fast it got negative seconds and popped off the chart

[@theintensegamer178]

+Nate Dimelfi lel

[@theintensegamer178]

nvidia pays people like Linus to make them look good. so when the fury destroyed the ti the had to 'forget' to put it on the chart.

[@rbspace454]

Thank you for this great information as usual!!

[@CAOHnutrition]

That is good to know!

[@TheActualTed]

r9 290 is better than r9 390X. Interesting...

[@humza22]

yeah I thought higher number is better performance but apparently not

[@Txx00xic]

The 3xx series is quite literally a rehash of the 2xx series but with more VRAM and slightly over clocked.

[@TaireTV]

390x has lower TDP, and performs slightly worse than its predecessor

[@Txx00xic]

+Norry Fox Show us a benchmark?

[@Txx00xic]

+Norry Fox Well that's unfortunate. If anyone could run proper tests on this it would be Guru3D

[@NoahThomas29]

I don't get why there is a bias against AMD on a lot of tech channels

[@seananderson7087]

+NoahThomas29 check out TekSyndicate, They are kinda AMD biased but not in a bad way, more of a bang for your buck kind of way, overall they tell it like it is though.

[@BeastinlosersHD]

+Sean Anderson amd makes great cards... But the drivers are not that good and a lot of devs use nvidia software, therefor nvidia gets better performance on some games. That's why all the Bitcoin / alt. Coin miners use amd.

[@the-bgrspot6997]

+BeastinlosersHD Whoa nice bullshit comment

[@HeyDan1983]

+NoahThomas29 Because all tech youtubers are bought by nvidia, nvidia seems to spent millions on marketing and it seems it is working...

[@thewok2291]

nvidia cards are generally more powerful for gaming and video editing

[@emptypocketsgarage9378]

Thank you Linus. Finally the answer I've been searching for! I knew LTT would come through with straight forward definitive answers for a novice like myself instead of "You have to have a Quadro or Fire Pro if your going to edit videos" .

[@CapturedProductions1]

Great video guys! Keep up the good work.

[@MrShott13]

What about Nvidia's Quadro, and AMD's Firepro GPU's, arent those specificly designed for rendering?

[@thewatcher_476]

yes they are and IMO they should've done this video with those cards

[@IzzyTheEditor]

+MrShott13 Nope. From my personal experience, the Quattro's are designed for number-crunching & dense polysets for higher-end 3D apps Like Maya, Max, Houdini and CAD applications etc. For editing and motion graphics, at least with AE, PR & C4D, The GTX/Titan cards and comparable Amd's are plenty:) (not to mention WAY cheaper)

[@Oscar4u69]

Seriously Linus, you are so biased that you didn't wanted to show the results from the fury X at 5:10 , that "strange" behavior was the fury X totally beating the 980ti? how much money nvidia paid you this time...

[@randomaccount53793]

Yea I was also wondering where the Fury X went in the last test

[@Anzac97]

+rodrigo maldonado Someone didn't read the bottom right hand corner.

[@dingdong2031]

Or it just wasn't working properly. Just because something doesn't show up, it doesn't mean that they didn't want to show it. If it had been running fine then they would have showed it. Not tying to battle for each side just trying to battle for LTT.

[@phil211293]

+rodrigo maldonado Don't you mean "strnagely" Seriously how can someone fuck a word up that badly?

[@noneofyourbusiness4294]

easy thing to repeat a teste but nah... It's AMD, not nvidia. Linus, this is ridiculous

[@kevinhochstatter4384]

Hey im curious what your encode time with all of the gpus working in tandem was compared to a single gpu , or are you making a separate wrap up video on the encoding rig?

[@geoffreybassett6741]

Thank you so much for this, not enough people use this when benchmarking a card and it is my #1 priority when buying a new card.

[@SpaceSaiilor]

the worst thing that happened to this channel was moving out of the house.

[@Andee...]

+Shirt Neck why? :P

[@SpaceSaiilor]

Andyholm content went downhill. Linus dosent let Luke sleep.

[@Andee...]

Shirt Neck oh, i see. damn poor luke :/

[@micThurrr]

how come the fury benchmarks mysteriously disappear?

[@thepwnagepk]

+Michael Theurer 'R9 Fury was behaving strangely and couldnt complete the benchmark' is at the bottom right of the graphs where its not shown

[@TheAssirra]

+Michael Theurer it saids so in the video down right "R9 Fury was behaving strangely and did not complete this benchmark"

[@BoltplayergamingChadked]

+Michael Theurer take a look at the right bottom corner at 5:15, you'll get to know

[@Sketch1994]

You just hit a vein of nostalgia with that 7800!

[@simeonkolev1231]

Great! I am waiting for the second video on this topic :). I would like to see how SLI/Crossfire affects performance and not only Premiere, please add After Effects in the tests.

[@spik330]

No testing a quadro for professional use?

[@bishrarar3015]

they should have put a Firepro W9100 in there :(

[@gaiustrollius9920]

I think the r9 fury should be on all tests you did. Also, could you add the fury x to the test? You know, to compare it with the gtx 980ti.

[@sharptenor]

Love this. I edit on a rig with 2 Fury X and it is worth mentioning that Fury X isn't currently supported in Premiere. While Fury does well in exporting footage, it seems to cause a number of crashes while actually editing (Adobe had me switch off acceleration). So it's important to note that exporting isn't the aspect worth considering in selecting a card. (Also Adobe: add Fury X support!!) +LinusTechTips I would LOVE to see a whole series of videos on the hardware side of video editing.

[@hitardo]

Loved the video! But I cannot stop thinking about the Quadro NVIDIA series.

[@Trivian]

1: AMD performs better by 33S : No noticeable difference. Very Similar experience. 2: AMD performs better again by 80S: Graphics card is optional. Which graphics card you chose doesn't mater. 3: Nvidia performs better by 57S: GTX 980Ti is the beefiest gaming GPU! It comes out on TOP! Lesser cards are performing as much worse as you would expect! Also, R9 Fury was behaving "strangely" and nobody knows where is R9 285! And a Nvidea 960 makes a guest appearance.

[@alexanderbolton3562]

+Trivian It extremely disappointing for him to be like that. You would expect an unbiased review but however many times you look at it. Its just pure biased and the NVidia fans even saw that

[@MalphasMedia]

"R9 Fury was behaving strnagely" Someone is gonna get fired.

[@igorvieira344]

awesome video. please, dig deeper! more of this

[@NathanaelNewton]

Need an updated version of this video..

[@tiagocpinto]

AMD cards win in 2 benchmarks no commentary.... nvidia wins 1 ohh this is what we looking for..., i like you guys but could you guys be not nvidia fans boys?

[@dandi2864]

+Tiago Pinto It's about who gives them the most money or products, nvidia obviously gave them a crap ton of free samples and sponsorships

[@TheBaldOne]

Why in the first slide we see a Fury X, but then when the 980Ti wons there's no Fury X to be seen?

[@kulmajaba]

Sweet! This was exactly the video I was hoping for. Suggestion: test whether SLI or Crossfire affects render times.

[@Tron08]

I'd love to see these tests again but with low end and high end workstation cards (Quadros and/or Fire Pros). Would be interesting to see how cheaper workstation cards compare against more expensive gaming cards.

[@vjHeaven]

So in two tests AMD scores, one with over 100 seconds better and you say it doesn't matter what you choose, then in the third test NVIDIA scores better and ding!! we get the GPU's highlighted and stuff. Not a fanboy but these marketing gimmicks are making me sad.

[@johnnychang4233]

Does the workstation versions of FirePro and Quadro counterparts offer better performance with rendering software? I see you guys just used gaming enthusiast cards for all your setups.

[@wapted]

+johnny chang Firepro and Quadro versions do tent to provide better performance for these types of compute tasks.

[@hikaru-live]

+wapted Not really. As long as the graphics chip is the same and and memory is not posing as a bottleneck, there is no real benefit switching over to professional-grade cards. For example, GTX 750 Ti vs Quadro K1200 or GTX Titan X vs Quadro M4000

[@wapted]

陈北宗 You would have thought, but looking at benchmarks says different...

[@hikaru-live]

+wapted Benchmarks and real life performance are fairly different. You need to research with your own use case in mind, even including what version of the tool you are using (even the same tool, when compiled differently, makes a difference.)

[@wapted]

陈北宗 By benchmarks, I meant people using software (like solidworks) with gaming cards vs firepro vs quadro. Then posting the times they took to complete tasks.

[@zippityzbrake]

I love how the slogan says "Repeating these is pure laziness" and that slogan has been used before. Made my day

[@Modpil]

Hey Linus,love your hoodie.Where can I find one?(If possible please link it)

[@faylow6416]

in the third test there was no fury card like really linus ?

[@skipp8268]

It "behaved strnagely" how serious

[@ProtoMario]

I was going to leave a comment about how Intel and Nvidia pay off Linus buuuut I found this comment which sums up why this video is not unbias - "Well if one product is clearly above the other one in the graphs (the product being AMD cards) and the oddity of the fact that only one pure graphics intensive test was highlighted in which Nvidia had a lead and Linus saying he doesn't think one brand has an advantage over another while showing a zoomed in graph of on specific Nvidia trumping AMD graph all the while the Fury X which was standing heads and tails over the competition was suddenly missing in the last graph. JayzTwoCents made a video not too long ago about cloud computing where he highlighted the fact that AMD is the preferred brand when it comes to video editing none of which being mentioned clearly highlights the Nvidia biased attitude of LTT, all the while in a recent WAN show Linus constantly had a "bitchy" attitude towards Polaris and said the only reason AMD has HBM cards is because there is so less HBM that Nvidia can't possibly incorporate it in a major release while AMD with their low sales can". Damn Linus man, what is going on really?

[@ExtremePotato]

+ProtoMario Nvidia/Intel gave him a small loan of a million dollars

[@Lone_Rocket]

My intergrated intel graphics does just fine rendering my videos. Thank you very much.

[@MrLuc420]

I would love to see a performance test comparing professional vs gaming video cards.

[@itsthatguyfromthething]

"Midrange Xeon" lol

[@LinusTechTips]

+Julien Nogueras there are xeons for a couple hundred bucks..

[@LETHALF90]

There are xeons that are just i7s without igpus or oc. Stop being ignorant.

[@itsthatguyfromthething]

+LinusTechTips yeah, but hearing you say it like that was kind of funny

[@TheSpacecraftX]

+Medo Gabr On what planet is even an i7 mid range.

[@TheSpacecraftX]

Rowan Bink What?

[@thehinzee]

Did linues really call the gtx 980ti the "beefiest gaming gpu" Has he heard of the R9 295x2?

[@JohnDoe-uq2qd]

+KSK- hinzeePlaysDayZ He called it the beefiest gaming gpu in comparison to the cards he was using for the test.

[@jakebezzina6729]

+KSK- hinzeePlaysDayZ followed by the Fury Pro Duo. I can literally feel the "oddly working" Fury being tampered with by Linus while fucking with the r9 390x to make it worse than an r9 290 which was glorified by him back then.

[@LinusTechTips]

+KSK- hinzeePlaysDayZ a GPU is a graphics processing unit. The R9 Fury Duo is a video card. They are not the same thing. The Fury Duo uses TWO GPUs to achieve its performance - an approach that doesn't always scale in certain applications like this one. Hope that clears things up :)

[@thehinzee]

I...have been schooled....

[@jakebezzina6729]

LinusTechTips I know I was acting like a bitch before but can you please make a video explaining why the results were so unusual and also re-run the test after the Fury is back up and running. Although this is not the first time a fury has failed.

[@KSUBlueDragon]

This is nuts!!! that would be awesome to have a gaiming rig server like this...wonder what all it takes to do this kind of thing...

[@therealphee]

Hi! I love your channel. I recently bought a gtx 980 for my Mac Pro and I'm curious about benchmarks between Metal and Cuda on this card when using premiere pro. Im hoping to speed up my rendering when applying denoising.

[@coolbeanz147]

Just to confirm is this the 3rd time that "Repeating these is pure laziness" has been used?

[@SteveDice21]

Forget about the Fury X disappearing from the graphs. How the hell did the 290 beat the 390x?!

[@FroztiProductions]

for real wtf

[@TrekSupport]

Why did you stop using the R9 fury in the Chroma key footage 1080p test? It seemed to be doing pretty well against it's Nvidia counterparts in earlier tests but did not show up in the 1080p test.

[@benjones5799]

any chance of doing that with some quadro cards :) plus would be good to see timeline playback.