Crushing RTX 50 Benchmarks Leak | RTX 50 & RDNA 4 Release Date
HTML-код
- Опубликовано: 15 июл 2024
- Whokeys 25% off code：RGT
Windows 11 Pro $23,2:biitt.ly/581eD
Windows 10 Pro $17,6: biitt.ly/9f0ie
Windows 10 Home $15:biitt.ly/XYm9o
Office 2016 $28,8:biitt.ly/cy7Vb
Office 2019 $48.6:biitt.ly/YInvw
www.whokeys.com/
RTX 50 Wrecks In Benchmarks | RTX 50 & RDNA 4 Release Update
Some very interesting RTX 50 performance benchmarks have leaked, as well as some confusing rumours on the release date for both Nvidia RTX 50 Blackwell GPUs, and AMD's Radeon RDNA 4. Join us as we examine the latest RTX 50 leaks, which includes the interesting performance results which have surfaced, as well as what is going on with the conflicting reports on the launch plans for RTX 50.
Plus, we have a small update for AMD RDNA 4. As expected, RDNA 4 wasn't shown off at Computex 2024, but that didn't stop Hassan of WCCFTech trying to find out when RX 8000 will be launching - but how do these launch date rumours line up with what previous leaks and rumours, and our sources, have been saying? Even though RDNA 4 will not have a high end flagship graphic card, it's still highly anticipated as gamers hope for a affordable mid to budget range GPU.
Subscribe - ruclips.net/user/RedGamin...
Join our Discord - / discord
LINKS
x.com/3DCenter_org/status/179...
x.com/hms1193/status/17983384...
x.com/hms1193/status/17980627...
Tip on PayPal - paypal.me/RedGamingTech
AMAZON AFFILIATE LINKS
US - amzn.to/2CMrqG6
Canada - amzn.to/2roF4XO
UK - amzn.to/2Qi6LMS
GreenManGaming Affiliate link -
www.greenmangaming.com/?tap_a=...
SOCIAL MEDIA
www.redgamingtech.com for more gaming news, reviews & tech
/ redgamingtech - if you want to help us out!
/ redgamingtech - Follow us on Facebook!
/ rgtcrimsonrayne - Paul's Twitter
/ redgamingtech - Official Twitter
0:00 Start
0:02 Today's topics
1:20 RTX 50 performance preview
7:39 RTX 50 release date confusion
11:26 RDNA 4 release date
Royalty-Free Music - www.audiomicro.com/royalty-free-music 
Игры
![Sean Rii, Karyon, Sharzkii - Taungule [My Love] (Official Music Video)](http://i.ytimg.com/vi/eJoMuypbSzQ/mqdefault.jpg)
5:07 Jumpscare... Your face just melted!
Boo!
He has changed
One thing is for sure, the 5090 is going to be between 5% and 2000% than some other card from the past.
Do you have some secret reliable source? 😅😅😅
@@PjotrII my uncle who works at nintendo
And that's just the price. lol
@@ShaneMcGrath. the more you buy the more you save!
@@Healthy_Toki 🤣🤣
and here I am living in a 3rd world with 56kb internet watching this video in 240p
do you also have goat-powered internet
If it makes you feel better; I'm in the U.S (a "first-world" country?) watching it in 144p(to conserve my data).
I am sorry bro
so what, you can come here and take for free everything that you didnt work out in your place, you dont have to be angry your goverment for being religion oriented instead science and citizens growh
and here you are speaking fluent english not trying to get a green card and move to america so you can get paid better.
a lot of ppl were hopeful we were back to a "performance" generation after a "features" generation
Honestly, I'd rather wait 4 years for both a major performance increase and an abundance of new features rather than having the two be staggered every 2 years. Would make each generation more worthwhile.
Lmfao? 50% increase in rt and raster over an already monstrous card that bottlenecks most cpus is nuts. Plus nvidia is the leader on features. It will probably cost a kidney but they can charge it bc no one can step up to the plate and that 7900xtx performance proves it, lmfao stonked, and I am aware rdna4 will still used gddr6 so they will be absolutely shellacked, nothing will beat 5090 for a long time unless it comes from nvidia.
@@rodiculous9464 You almost seem happy that you're going to get fucked as a consumer. I guess it is pride month after all, enjoy.
I'd rather want an "affordable" generation.
@@nipa5961 Agreed
50% more perf.
100% more $$$.
Why would anyone upgrade?
Happened with the 4080 already.
people are still buying nvidia because they see value on it, that's reality, the internet idoelogy does not matter
there's no value on it
becasue its the fastest? O.o ferraris are still selling well althou the perfromance is barly 50% higher and the price over a "nomral" car is easly 10-15x
@@gamingtemplar9893 Quite the opposite, people are buying nVidia because they don't understand the value of it. In reality, nVidia has worst price/performance value. Almost every tech reviewer always recommend AMD in most budget range except the highest end. However, nVidia has the absolute mindshare and most consumers aren't tech savvy and are just looking at catalogues from OEMs like Dell, Asus etc. they will just buy what is offered.
The first graph is just looking at the effect of GDDR. Obviously performance will vary from game to game. Why would Micron know what the current performance of a 5090 was, or think they could leak it, if they just provide the memory for it? Micron says on their site ''GDDR7 is expected to achieve greater than 30% improvement in frames per second (FPS) for ray tracing and rasterization workloads''
because you know the exact spec of the memory so you can safely extrapolate backwards as to how much perf it takes to need gddr7, generally in gpus the limitation is taking all the frame data getting that data into the frame buffer faster so it can start the next frame.
@@smartmonkey777 It very much depends how fast you run the memory. And Micron is very unlikely to know what the final bus speed for the GPUs will be.
And GDDR speed is surely not the only factor in GPU performance?
You sir, are one of the few intelligent people here. That includes RTG. Those slides have zero information on the 5090, only the relative performance delta between memory types.
It’s interesting but I don’t think it’s a “benchmark” and I don’t feel like I can confirm any definitive proof from someone’s Estimates.
-But, it makes sense that GDDR7’s uplift in speed will probably provide a substantial performance increase over GDDR6X
Micron could have easily configured a RTX 4090 chip with the 3 different memory platforms. Why would Nvidia share their 5090 chip with a memory vendor / 3rd party?
@@louisfriend9323 There's a limit to the performance increase from increasing memory bandwidth. Not only that they would have to get a 4090 with a GDDR7 memory controller.
Proximate quarter difference in BW "release date" could be the difference between 'supplied for production' and 'released for end market sale'. mb
So, now that 16% IPC for Zen 5 is confirmed, we're just pretending like your Zen 5 leaks didn't happen?
"Zen5 IPC is ABSURD"
He makes a living posting click-bait titles with "anonymous rumors" claming X and Y which of course no one can argue against. Of course he will keep bs:ing for easy money.
Inters of cpu IPC increase anything above 10% is actually insane unless it's usually a new architecture. The 40% was never realistic if it was still based on Zen architecture.
@@kennygeheim4230 "My sources say 15 - 30% IPC but some sources go as far as 40%, one even 50%, but guys lets be realistic, 40 to 50% even if true, could only be in specific workloads. Realistically we can expect 20 to 25% IPC"
Usual verbiage.
You've got the best low pitched voice. Can hear you rambling for hours till I fall asleep. 😴
You need to watch "Moore's Law is Dead" and he too has a deep voice and he drones on and on for hours. This is my go to when I am having an insomnia night. And I do enjoy much of his content, very informative.
*Let us never forget the 50 series fiasco. Never let that disaster repeat itself again. Prices were jacked because the cards were sold in the wrong tiers.*
*4070 was a 4080*
*4080 was a 4090*
*4090 was a 4090 ti or TITAN*
*Nvidia messed it all up. And the prices along with it.*
4060 was a 4050.
5090 8GB will start at $5090.
@@nipa5961 8GB is a bit too generous, 7.5GB... Now that sounds much better :D
you think Nvidia raised prices on accident?
nVidia didn't mess up, they planned it to be that way. Because if their 4080 was supposed to be a 4090, then 7900XTX would have beaten a 4090 in terms of rasterization performance. nVidia never like to lose, 6900XT came very close to 3090 while undercutting nVidia, same for 6950XT and 3090Ti that they won't let it happen again. So they pushed the power envelope so hard on their supposedly 4090Ti and call it a 4090, and there is nothing AMD can do. Even if they would to come up with a 7950XTX at the same power draw, it still won't beat it and it'd look bad losing to a "4090". nVidia also knows that they have the mindshare and the money to gobble up the manufacturing capacity that they will profit from selling the cards at an increased price and nVidiots will still buy them even if AMD cards have the better price/performance ratio.
Zen 6 higher tier + RTX 5080 ti Super will be my next upgrade
Am I imagining things or is the audio slightly out of sync with the video.
There is no reason for Nvidia to push 50 series to next year, it makes sense they would release the mobile versions in q1 '25 .. that's a better explanation, a leak confirming if indeed they will nerf 512-448 it would be much appreciated
they can take it easy if AMD doesn't release anything high-end
I pray to god that we see a 512 bus on the 5090, I know it’s a vain hope, but a man can dream
so you can run what games or are you actually going to use it in a supercomputer like it's designed to?
@@edwardpark247 I train AI models for a living, so yea, i’d like 32gb of vram without have to pay $3000 for a prof card lol
@@bigbo1764 That is a very cool job. I'm at the very beginning of learning the ins and outs of generative AI. It's fascinating stuff.
One thought, NV might try to increase bandwidth by using faster RAM rather than make an increase in the bus width. I trust NV's engineers to do amazing work, and also, I trust their marketing department to piss me off about something. Like weird specs that are borderline trolling (8GB VRAM wtf)
Like telling reviewers they should give good reviews or they're not gonna send them samples... happened to Hardware Unboxed during COVID. It ended up with NV's community manager apologizing and blaming it all on cabin fever.
AMD managed to parley a foothold in the market and turn it into gold, I admire them for their pluckiness and success. NV I admire, but I cheer for AMD, while holding NV stock. *shrug
Thank you. I was worried they were going to nerf its potential.
Nvidia will release RTX 50-series cards in mid-late November to take advantage of holidays sales. Whether it's a 5080 or 5090 first will depend on the amount of high criteria chips left from AI projects although the 5080 would be a better seller during the holiday than the 5090 with its higher price tag being out of reach for more people. So which one comes out will be undecided until last minute but one, at least, will release mid-November.
Main reasons being touted for a 5090 deelay is the ban by the U.S. government and the 5090 cannot be imported or sold in China.
Claims are the Nvidia wants to do a global wide launch so it has to be the 5080 and hopefully the later 5090 launch will not sting as bad to the very large Chinese market.
No way the 5090 has less than 160 SM’s. The total of SM’s of Blackwell is 192. It would make more sense if it was 176 SM for a 5090.
The may be setting up for another SUPER lineup to correct all the nerfs they make. Absolutely ridiculous.
It's like they sell us their botched cards first. Then give the perfect cards a rebadge and sell them later on.
@@asmongoldsmouth9839Sadly seems to be how it goes... Save up the best ones to release Super 1 year later
@@asmongoldsmouth9839 I'm thinking 5090ti with 32-30 gb, while 5090 will be massively cut down and have 28gb ram
They could use the total 192 of Blackwell SM's.
@@ahs9674 they have to find a way to get around the RTX 4090 ban in China for Blackwell
I do not know, I think it is hard that 5090 makes the same jump a 4090 did from 3090. In that case they were jumping from an 8nm Samsung Node to a 4nm (5nm) TSMC node. That was one of the main reason the 6000 series of AMD were more competitive than the 7000. The 6000 was using a 7nm that was better than Samsung 8nm. Nowadays both AMD and nVidia are using the same 5nm node (a bit tweaked for nVidida). However, next gen will probably use the 3nm that is not a huge step as that from S8 to T4.
That said, this is Micron test, we do not even know the game they are refering to, it could be AWE2 so a jump from 27/30 fps to 40/45 in 4K native or it could be a usless game where ray tracing is just minor. We do not know the platform, it could be using better memory and cpu for the GDDR7 and worse for GDR6x and non x. The impac may be low, but it can change some points. That said also AMD will use GDDR7 and also Samsung will make GDDR7 (they already do) so Samsung memory could be also better. Too much unknow variable to predict a non launched product yet.
as long as the 5090 is 25% better than my 4090 AND has DP 2.1 UHBR20 Ports. sign me the fuck up for day 1!!
Dumb@ss.
You speak like an Aussie of which I am one of them.....fuck yeah.
Ken oath ! 👍
@@johnvandeven2188he’s fake. He’s not gonna buy the xx90 cards.
5060 rtx for only 599$. Almost double performance of 4090😮
5060 will be almost 50% faster than 3060.
@@anttikangasvieri1361 man, the 4xxx series was really a joke.
@@Raikos100I call it the 40 series fiasco.
with a whopping 8gb of vram, trust us you dont need any more.
will perform like the 4070 at best, potentially way worse if as bad as rtx 4000 gen, 4060 is worse than the 3060ti... wich is hilarious so dont expect much and 8gb ofc.
Those graphs are meaningless. They list them as memory platform comparisons without specifying what platform they are referring to. So selling that as any kind of memory comparison without using the same base platform is a scam at the least. Or dishonest marketing ... however you want to look at it. Those can be 2080/3080/5080 and it would still be true from the perspective of what the graph shows ...
EDIT: RDNA4 and Linux drivers, they have to start early on those to get a release kernel ready when they want to release the cards, so basing the release date on Linux kernel/mesa changes is not really reliable.
uh, that sounds kind of paranoid and reactionary... we'll see what comes out and when, no point in debating it now
I don't understand how anyone could be excited to pay more for less this gen after last gen.
Some folks have a lot of money. I'm sure you have seen people driving Mercedes Benz's and Porsche's and Land Rovers and what-not, these folks can afford it without having to stress about it and there are a LOT of these people out there.
I hope with the 5090 path tracing and mesh shaders becomes the default in every game. Those two thing combined are really next generation and make for amazing looking games.
I never realized that you're all jacked. It's good to have another fit nerd around. Oh, and you know, a window aircon doesn't really cost that much. Use it before you shoot and the studio time will be a lot more enjoyable.
Bro you should be a voice actor. You'd play a perfect villain with that voice
Since Jensen promised to release Rubin (the next gen architecture after Blackwell) in 2025 during Computex, shouldn't we be getting Blackwell this year or are we getting two different families of GPUs next year?
Rubin is for IA not for gaming... The main focus for Nvidia is datacenter and generative IA
Let's see when amd is going to fix gpu memory leaks from vega to rdna... These were supposed to be ready by March. Will rdna4 have a h/w fix for this?
5:08 RGTech is hosted by a shape shifter. Subscribed!
We are the people who have been waiting for news of the 5090 graphics card for a long time
Nvidia made us very excited, especially with the leaks, with a 70% increase in performance, so I saw it as if it was a major, controversial global event
And now, after all this time, things have changed and made it worse, which is today’s leaks, which were disappointing to me.
I changed Nvidia and made the 5090 just try to be a little better than the 4090 and we probably wouldn't notice the difference.
Worst of all is that the real reason behind this play is only profit
seems like just a heavy Ray tracing card release.
"Heavy RT release"? Raster shows 42% in 4k and RT 50%. Performance jump in lower resolutions in raster is kinda a standard for some time already and it will be standard in the future. Because as GPUs get wider you need to have growing capability to feed it all. And that part is not progressing nearly as fast as raw performance.
@@korinogaro yeah it is what i said.
New selling point will be the ai shader cache system for games.
@@jasonvors1922 Yeah or AI cooling cooler.
Nvidia would be better off making a AI processing unit board specifically designed for game AI calculations and game image upscaling for one of the PCI-E slots and a GPU for the rasterization/physics and other tasks on the other PCI-E slot! It would practically work and realistically sells a lot better...
plot twist it's a 5080. Very unlikely as that would mean we're on the cusp of another Pascal-like leap in performance gains. Would be nice but with companies just throwing more and more watts at their chips to get gains, it's likely a beefed up 5090 variant that'll end up being cut down to keep MSRP down and profits up.
It doesn't really matter now for AMD if its 500% faster than 7900xtx.
Sources seem to be tight lipped right now. All the videos here are not referring Twitter etc.
There's no way the 5090 will have less than 160 SMs. That is just ridiculously cut down for a 5090, and if it has less than 160 SMs, then there should be no price increase over the 4090 to at least make up for the lackluster hardware upgrade. RTX 50 is getting nerfed way too much.
they want to put fully ( or very close to) enabled GPUs in their pro cards, something like successor to RTX 6000 Ada 48GB that is like more then 8 Tflops faster then 4090 ( in FP32)
@@evilsatorii It's not even that. Nvidia have zero competition at that level and can do what they want. People will still buy the 5090 if it is heavily cut down because there quite literally will be no alternative.
@@riven4121 exactly. AI is where money is made and there is no competition at high end, so we get modest improvement.
@@BridgeTROLL777 AI is where most of their profit is made. I won't deny that. Gaming still earns them billions, though, and there is a lot of money to be made in that market if they do it right. It would make no sense for Nvidia to give gaming very little priority, considering it's their second largest business. It just won't be quite as high of a priority as AI. If anything, it would make the most sense if they indicated a strategic aim to strengthen its market across its data centre and gaming segment.
@@Gamer-q7vWell the 4090 was more cut down than the 3090 was. Along with no competition in the GPU market it would not surprise me.
Battlemage ftw
So RDNA4 that already is a generation behind, comes even later than the 50 series? How on earth are AMD gonna catch up.
How on earth that matters ? People just want AMD to compete so that Nvidia can have decent prices .
@@potatorigs2155 and my question is how are AMD gonna catch up to compete when things look like this.
Intel is progressing faster now than Amd for discrete gpus. That is what you should be looking at. Amd is likely going to choke for the next three to five years. Too much copium getting inhaled by Amd crowd.
Why do you think AMD even wants that. People are blided to buy Nvidia only even when raster XTX was and is still good card. Let the Nvidia fans buy new gen card even higher prices. They keep buying and supporting Jensen AI
if nvidia keeps doign it like till now, i think amd will never catch up nvidia, better i am worng, but i think thats gonna happen.
prepare to be disappointed
This dude is a fraud
In other words, you/we still have no idea of what and when Blackwell is going to be released. So again, more meaningless rumours and speculation. He's just covering all options!
So, Jensen has a "famous leather jacket". I say bah, humbug. It was Steve Jobs that made famous the clothes he wore when doing a presentation. But I will take my hat off to him for his net worth of $7 billion and how successful Nvidia have become while under his charge.
Leather jacket? Lets talk about spatulas. BTW, Jensen's net worth is over 100B, dude.
@@edwardpark247 Not according to Google. But hey, hell never get to spend much of it in his lifetime.
its probably gonna be 28gb and 448bit bus, any higher then that and they eat into the profits of their workstation cards
I'm hoping we see this unbridled 5090, even if the 5080 is chopped way down for sanctions. Thumbs up!
I'm glad I don't play a drinking game for every time he says "it's gonna be very interesting" and the occasional very curious lol sheesh it's borderline annoying
Well to be honest it makes perfect sense to make gaming GPUs gens later (at least as long as AI craze lasts) longer. Gamers are dirt poor in comparison to corporations and you can use wafers that would go to gaming GPUs for AI GPUs.
Nvidia = the king of avoiding the 512 bit or greater busses while investing into cores until the outlet explodes :)))
5:08 OMG The truth has been discovered Red Gaming Tech is run by AI!
Those performance numbers are bad, it's 14% at 1080p, when the 4090 is compared to the 3090 performance is much higher, around 60%, so the jump is lower this time "only" 40% if these numbers are true.
Lossless scaling software came out with a 3X frame generation which works fairly good with older hardware.
So forget the rtx 40s, even 1060s stays relevant.
Not doing RDNA 4 this year at least around Christmas is a mistake. Not interested in RDNA 3.
I have to agree. RDNA 4 sounds like something this market needs. Affordable mid range GPU's that bring 30% or more performance to the $500 price category. There's also a good chance that it will include AI infused computing because they're adding it to their upcoming laptops. A midrange GPU like this would demolish the value of Navi 31.
Yes. In fact, to have hopes for any kind of marked share gain, they should release it ASAP and make it a bargain. Regardless of RDNA 3 stock situation. But it doesn't seem they have any fight left in them.
@@jaxonswain3408 except the timing is really bad if leaks are true. launching rdna4 next year is a huge mistake. thos cards will be dead on arrival. simply cuz nvidia would have already launched the budget class of 50 series and even intel would have released battlemage gpus. the current time frame makes most sense for amd cuz there aren't any great options under 500 usd rn.
There could be a damper on the production of these cards. Micron just lost a patent trial to Netlist for stealing and implementing their technology. There's more patent trials coming up for infringement also Netlist vs Google and Netlist vs Samsung
Its def a massive upgrade but wiating until 6000 series will def be smart. Also i know nvidia will hold back if rnda3 is still happening. So I feel the 5000 seriees will be great but may not be as high as shown here if they dont have to or they may wait even longer lol
What's the drop on the price tag an explanation of why we gotta pay at that price point? 😅
Nah man, just wait on RTX 7000
@AwesomeBlackDude agreed lol if you own a 4000 I can't imagine buying a 5000.
@raeferwilson2599 why? General rule of thumb is 2 generations not 3 why should anybody go and get a 5000 right after? 6000 is smart. Get the most out of what you paid for and then get something that will be even more competative next generation for games. Because these cards are good enough even for the games coming. Going 6000 will then allow you to keep up with the more advanced games that will be coming out 3 to 4 years from now.
I disagree I can sell my 4090 for more if I get the 5090 at launch.
Why the performance increase in 1080p is so low? Ah, as someone that preffers to play games at 240fps this is annoying, all these companies care are about rAyTrAciNg and " fOuR kEy".
5:08 Mask off
Giving the people what they want is a poor business model. Expect disappointment.
Hyped for Alien Romulus?
Yes, I am itching for joy. lol 2:10
I do agree that RDNA 4 is not coming this year, to me it does not make sense for AMD to release new GPU's considering the fact they only started exporting RX 7900 GRE outside of Asia market in 2024 and the sale of that card are very good. So the question is: Why would AMD want to cut the price of RX 7900 GRE by releasing new GPU's on RDNA 4 when the sale are high? (don't have to answer this) - For most people talking about RDNA 4 this question never cross there mind. Cheers mate!
Please keep the nvidia info first bc i do not watch amd gpu news like most lol
As a potential 5080 buyer I am only concerned with "REAL" frames produced by the new card. Don't care at all about how many fake frames it can generate!
AMD is using Nvidia's strategy of market dominance against intel while baiting Nvidia to overperform to the point where they are on the edge of their designs (see the new power connectors). At that point, AMD can counter exactly where Nvidia is lacking. I am guessing Nvidia will outgrow this in favor of AI in not very long. I think they sensed that, too, by not mentioning GPU's at their last conference.
Amd is not using some kind of clever strategy here. They don’t put enough resources into their Gpu department and they are falling farther and farther behind while Intel is gaining ground. There may be some potential for Amd frame generation working on many titles one day, but that is the only bright spot for now. Maybe they can turn around FSR with machine learning
@@xlr555usa I too think that is a lame strategy, nVidia is in such lead that AMD won't be able to provoke a reaction from them. If anything, i think their GPUs are designed for compute first rather than gaming, and they probably had to push their cards harder, sacrificing efficiency for gaming performance.
As for AMD, i don't think they didn't put enough resources into their GPU design. RDNA4 was supposed to be a MCM design, which is probably why they didn't have high end monolithic die. Turns out they weren't able to make MCM work within this generation thus they are stuck with smaller mid range chips that were supposed to be chiplets for the higher end. I think AMD missed their opportunity, RDNA5 should be MCM but next gen nVidia would also be on MCM design.
There is no strategy. AMD is lost and at already and least one generational performance leap behind. They don't matter anymore in the enthusiast and high end segment, unfortunately.
@@EFXVoila That's not how things work. RDNA1 only had mid range card, namely the 5700 and 5700XT. But RDNA2 had the full stack and were extremely competitive.
You don't compare mid range to high range cards and determine their generational progress in terms of GPU design. RDNA4 only have mid range cards, and you know so by the die size. If AMD wanted they could have made a bigger die and probably perform better than a 4090. There is no reason they can't. Even if RDNA4 design had zero architectural uplift, they can still get performance gain from the smaller node.
What AMD lost is only the high end segment this new gen, they are not behind a whole generation in terms of GPU design. Their mid range will still be competitive. And if they sort out the MCM design, they will still be able to challenge nVidia's next gen with RDNA5.
Problem is Nvidia is lacking everywhere except in the 90 SKU.
I am interested to see if the RTX 5080 really does out perform the RTX 4090, I mean there's no competition so there's no reason to, but if the RTX 5080 matched the 4090 with 16GB of VRAM, and they could sell it at a lower price, that would just slaughter AMD's 8900/8800XT/X, again they don't even have to do that, which is really sad, but you can't really blame a company for improving their products.
They are always competing with their own previous generation. They have to because a huge segment of the consumer base is always looking for an upgrade. If they don't offer them one they will sit on the card they have and their money.
These days "leaks" are more like a "get free advertisement from content creators" - mostly, not always. No matter what real performance those cards will have, they will not be worth the price. Nvidia offers nothing to be exited about - right now i'm just interested in AMDs pricing strategy since this will define the future market now.
The gpu market keeps getting better and better...
joke post?
So is the 5080 better than the 4080 Super or not?
AS I READ ALL the won't make much uplift simce the want that China market 😂 so at most a 4090 or cut down one idk
@@Oliver-sn4be Most likely 20% faster but also much more expensive.
yes. it's gonna be slightly faster than 4090 actually.
@@oogwaythesussyturtleIf the 5080 is faster or as fast as the 4090, then it's great for me. I Upgrade from RTX 2070 Super.
I can barely understand you with that anonymous witness voice you gave yourself
Same rumor based "news" Most likely same ridiculous pricing and sub-par performance when we will get the real products ;)
At the rate of how much more money Nvidia is making from A.I. compared to their gaming dept, in which A.I. is about 95% of their income compared to gaming, I expect this to be Nvidia's LAST generation of gpu's. As they transition from a "gaming" company to ... whatever it is that they're selling now. A.I. stuff.
It was already telling that during their computex keynote of two hours that they made NO mention of gaming, gpus or any other gaming related news. Just a moment where Jensen grabbed a GPU, with no markings, made some comment about it like "uh not this, this is a gaming gpu", then immediately took it around something else like he was embarrassed to even grab it and show it to shareholders.
I think they are too greedy to leave any money out there on the table for AMD and Intel to scoop up. They def prefer selling 40,000 dollar Ai cards but they want ALL the money, not to share.
Pascal was their last gaming GPU generation.
They only released productivity, mining and ai cards ever since.
5090 won't go that low on performance
AMD needs to be careful.if battlemage is around a 7900gre performance things might hit the fan.and if Intel release 1st AMD might be fighting an uphill battle.
Yes, if...
Arc also was supposed to perform like a 3070, but barely tied a 3060.
The only uphill battle is Intel fighting with themselves.
Do you even get day 1 support with Intel drivers for games?
@@MrNelahem I agree with you to a certain degree but Intel is just starting in the discrete GPU market and before you start rolling your eyes and think intel shill I'm an AMD fan both CPU and GPU but I can still see the threat coming down the road.
@@damyjuk yes down the road. I don't think either NV and AMD has anything to worry about from intel even with the next gen gpu coming out. They still need more time in the oven.
@@MrNelahem Day 500 driver support
ima just laugh at this video. gd thing leak is in the title as you can say anything u wanna and even fail hard to give any trust worthy source as you did and get away with it.
Yeah because the 5090 would be disappointing if this is true due to that terrible SM cut.
@ZackSNetwork highly doubt is true as again zero trust worth sources were stated. that and the fact none of the big tech tubers have reported on this says it all. he'll a factual picture of a single 5000 series card has yet to be released yet alone confirmed. this is the problem sith leaks they are for marketing most the time false information as with each so call leak information changes.
Only 40% in pure raster at 4k lines up with rumours of 5080 only being equal to the 4090, disappointing, I was hoping for the 5080 to hit the 2x performance improvement mark over my 3080, nevermind. If it's just 4090 performance it better not be more expensive than the 4080 Super.
Why in the world were you hoping for those gains... did you sleep through 4XXX Gen?
@@ocha-time I'm talking about a standard generational improvement. I didn't sleep through 40 series, the 4080 was a noticable performance bump over previous flagship 3090ti. I was just expecting that to happen again which would naturally make the 5080 to the 4090 what the 4080 was to the 3090ti. That would lead to the 5080 being 2x 3080, the 4090 was already close to that depending on the game. Instead it's going to be more like Turing where there is only 1 card that is actually faster than what was possible in the previous generation.
4080 +40% is actually around 15% faster 4090. 5080 should be close to 2x the performance boost over 3080.
@@skychaos87 RTX 5080 is gonna be 2X the performance of a 3080. RTX 5080 is rumored to perform similar to an RTX 4090 and the RTX 4090 is 2X the performance of a 3080.
Nvidia announced, the 50 series are begin 2025.
You saved yourself 15 minutes and a lot of of nonsense.
"""""""""""""leaked"""""""""""""
AMD will wait for Nvidia to price their new GPU'S next year so AMD can price their's. 😂
5000 series will be skiped. It happens a lot in the GPU space. RTX 3000, 4000 is more than enough to run anything. And it will be for a while... For more than a while. Untill Playstation 6 and new Series X comes out.
Lets face it if they're 5% faster people will buy them! They'll come with some sort of exclusive AI accelarator software I guess. They have no competition anyway, it's sad for all even if you're a fanboy prices will be crazy!
I have no idea what this guy talks about
For some reason i always thought you were a black guy
Welcome to my world. 😅
@@AwesomeBlackDude Naah, you sound like an old Japanese lady.
You thought there was black meat on white street? Get some.
I was hoping for 25Gb and they pushed all the way to 28Gb????
Thank you NVIDIA!!!
I mean, after the shameful 6gb drama, they had to apologize somehow. Now let's see if they keep the high memory for the rest of the line
the min they could go while going higher was 28 so thats the 25 basically...
nobody care for 1K-2K resolution this cards will be for 4K,this is main goal for nvidia!So no point to buy this card is you use 2K monitor for sure!
"R6" and "R6X"... No one calls it that bruv. "Rate6" and "Rate6X"... It makes you sound as if you know not of what you speak.
Not excited about this thing because I’m not buying it regardless. I ain’t paying more than $1k for a solid gpu no more. To me, gaming on a $50 Sega genesis is way more fun than whatever an RTX has to offer. I only need a gpu to run my desktop and for some casual gaming.
Unreal engine 5
You look fit good job 👏🏾
atleast rdna 4 will be ironed out to the max with it being the last of its kind and they are also taking their sweet time. tho never understood released mid winter. `no one` outside hc gamers buy hardware at this time of year. partially explains why nvidia move more units
I just don't care about $2500 video cards, so please don't bother hyping 5090, it's boring. Why not hype Google next datacenter? It will have 100x more cpu power than the 5090, even without GPUs ..
😂No no buy 4060 4070 4080 4090. its time for 50% off all rtx 4000 gpu now. Or Grab rtx 5000 better😘
yes it is true people suck
But those numers doesn't make sens 18% more at 1080p but 71% more in 4k I call bs
4k is important, 1080p is for DOS games or really old stuff
If this is true it’s because of the massive CPU bottleneck. Which is why the increase in 4K is significantly larger.
It make sense because opf the CPU limit. The 4090 performs way better in 4K than in full hd. That is working as intended
It actually does make sense when you look at 4090 1080p benchmarks from release day where the 4090 was only barely ahead in a lot of games until the resolution was changed to either 1440p or 4k as even 1440p was cpu bottlenecked a lot of the time.
It's just Micron looking at GDDR. Why would Micron who only make memory know what the current performance of a 5090 was?
Just work on your speech already, geezus.
Don't buy Nvidia. Buy amd!! We need to support amd!
Radeon is dead.
Geforce is dead.
Arc is dead.
In the chat - We can conclude that GPU´s are dead long live AI processors.
i checked they are alive.
@@Pleasiotic1 nvidia controlling the entire gpu market says otherwise.
5090 is boring, for real. I dont see any game where i need more than my current 4090
same, but I will sell my 4090 and buy a 5090 anyway. More fps are always nice :-)
4k high refresh for any RT game without having to use 1080p upscaling? Always a need for more performane, the price is what will make it boring.
Yh, I mean in terms of price. Offcourse more performance is always welcome
You can never have enough power. 8k 144hz is coming out next and my 4090 is only good enough for 4k 144hz, sometimes even much lower, say red dead 2 at ultra 4k max only 20 or so 8x msaa etc
@@EFXVoilathis. There is no overkill. That coming 480 Hz monitor gotta be used
3 actually's in the first 10 seconds. Youre just trolling at this point😂
the last few videos killed my hype tbh... 50% increase in FPS is kinda on the lowside , and the 28gb Vram is an insult... 4 extra GB of vram on the top GPU is just dumb, what are you running with an extra 4gb VRAM?? now if it was 8 (32GB) now that would be decent....
Wait 50% increase in fps is on the lowside?
So a 100fps game becoming 150 is low?
60fps game becoming 90 is low?
300fps game becoming 450 is low?
Im confused. What numbers are you expecting here.
@@CrystronHalq 50% increase in enormous
@@potatorigs2155 agree
50% increase in one generation is enormous.. As for the 4gb increase, who cares if it's not a big bump in capacity? 24gb is already total overkill. I'm running a 3090 and I don't think I've seen anything utilize over 12gb yet.. It's really the bandwidth we're after.
The let down is that whatever we get, is likely to be cut down from what we could have gotten. No one will compete with nVidia in the high end for at least the first half of the product cycle, so they are going to get top dollar for whatever they release, while holding as much back as possible to compete with RDNA 5 via Super and Ti versions.
@@CrystronHalq every 2 years perf increase is around 60% 70% for the top GPU. 50% IS INDEED LOW. instead of including 3 examples of FPS increasing by 50% (like we don't know BASIC math) try instead using that time to google past performance increases.
Fake news
who cares the games out there suck
They should name the new platform after you " ClickbaitRidge" SMH
All these 'leaks' of different configs of a GPU like the 5090 are total bullshit. Nvidia has absolutely no need to "test" all these configurations. Once they have the chip out of the fab and fired up, they'll know how it's performing and can calculate very accurately how it will perform with any given config. They don't need to test a million different configs. There will be some numbers to crunch related to yields and exactly where they want to place it in the market, and there will be some validation to do prior to retail release. But the narrative around loads of different configs being tested and experimented with is fundamentally clueless. It's not how it works. The different 'leaks' are just bullshit numbers being made up or at best misconstrued or misunderstood.
If he wouldn't use such a crappy voice I might even listen to his videos!