Agreed. I think amd can do it but maybe they will have to have a team design a WHOLE new architecture beyond the rdna name. I hope they are working on something
But they did with the 6950XT didnt they. It was more efficient than the 3090 and was just as fast. The 3090 Ti was faster yes, but used a boat load more energy.
@@valenrn8657 in reality amd a bit lucky intel has instability issue in cpu but for radeon vs nvidia or even intel arc theyre struggling with their architecture for overall performance
@@Dave-dh7rt did they ??? NVIDIA have largest mindshare, its very impossible when NVIDIA does lot more in architecture than AMD simple FSR tech that seem to me it took them 3 rdna gens to make it work. Not liking amd doing vs NVIDIA did it in day one and not wait 3 architecture chips to fix it ..what disspointing from AMD ...they wasted everyone time to get RT tracing to work...5 gens of RDNA to fix it ...wow...why AMD struggle something simple ?? Now Path tracing is different story
Y'all are being mean. It'll catch up with turing. Seriously though.. I love AMD and all I've every wanted is to have solid AMD products to buy so I don't have to resort to buying an $800 super ti GPU. but that is exactly what I just did. At least it will be worth $400+ in a few years when I upgrade.
Heard it all before. If RDNA5 was this year, it would be something, but 2026... yeah they might have the performance crown for a few months tops before they are once again runner up. It doesn't matter though, as long as they maintain the lowest cost per frame, it's fine. Also Paul, put on a shirt already.
AMD has had the price-to-performance advantage for several generations in a row now when it comes to raster, but they’ve been behind on RT, upscaling, downsampling, and notably also CUDA. At this point all of that matters a lot more when it comes to pressuring Nvidia.
It needs to happen. Nvidia is like intel was in the early 2010s. AMD is holding back the whole industry by not making a card that competes with Nvidia's top perforforming gpu. If AMD didn't release ryzen, we'd probably still be on 4c8t with the i7s and i9s locked away on a much more expensive socket.
Basically Ryzen. AMD started making major changes with rdna2, rdna4 is an optimization change, and 5 is the big picture. I don't think Nvidia has any actual new ideas, only cares about AI, dropped old feature support, and will ultimately completely stagnate, while AMD is pushing hard to compete. Not hard to see a turn around like AMD vs Intel.
4080 perf seems likely since 7800XT is like 70-75% the perf already. Shrink down, go wider, go faster, go smarter, go mono, and yeah, doesn't 4080 seem doable and likely? Also exposes that 4080 is basically twice the price it *could* be.
@attepatte8485 ihace 7900xtx and its performance of between 4080 and 4090 but 4090 beats itvraytracing but its rasterisation is alot faster and fsr 3 gas made bug sooed inprovements
AMD really needs to focus on the value of their offerings. Nothing good can come out of AMD trying to meet or beat Nvidia. Nvidia could spend AMDs market cap in a quarter on R&D to release a faster product and not blink an eye. AMD cannot afford to have an entire generation of R&D negated before they can recoup costs. They're still not that on that level.
@@yasserq78 At what point were we ever talking about datacenters? They're also not winning because of performance. They are winning because of openness. The datacenter equivalent of value.
I would expect RDNA IV sooner or perhaps R9K will dribble out. The function of RDNA lV is R9K attach and their volumes need to sync and something from this report seems leading and lagging. I already register RDNA lV components on interposer or SIP produced on production in relation financial assessment. mb
ok , time for me to do some theorizing form a electrical engineering student . If amd does a mini rdna 4 gen , only launching mid tier rdna 4 and pushing rdna 5 as fast as possible. RDNA 5 might be the first gpu gen that will use Backside power delivery with N2P , A16 gets backside power in 2026/2027. That alone increases the power efficiency and clock speed(around 6-7% from intel's current chip testing with backside power) regardless of architecture . On top of a generational Uplift in performance . On top of the possibility to even add 3d Vcashe to improve communication between chaplets( just a theory ) + gddr 7 . Amd might do a moon shot in early 2027 , Sacrificing top end rdna 4 to Launch Rdna 5 early , before Nvidia can respond with 60 series , forcing them to do another super series . but this time Rdna 5 might just be Far enough in Silicon tech that it will beat it with no contest . AMD gpu department might finally grow a pair of balls and try to Out compete Nvidea before launch , and not with post launch price cuts. Trying to beat nvidea with there heavy connections with TSMC , while rdna 4 is just there for the driver team to slowly upgrading Fsr ,anti lag, and frame gen.
also your correct...they are using CDNA architecture as band aid architecture exit from rdna4....yes rdna4 will be last rdna architecture by raja koduri the lead architecture designer for navi first gen that amd keep beating the dead horse since ...man they really surprise me to keep gonna till rdna4...4 generation of raja work...while CDNA architecture been around since R9 series using older tech that seem work better in programmable shaders that they can make this work for Ray tracing tech ....but not in raja work they realize by now..By early ATI unfinished work long ago that been around since AMD continue since the islands architecture ...not polaris architecture...that lead to navi afterwards from star names brand for consumer while pro's still using furry X tech architecture compute heavy programmable shader tech that requires compilers from the drivers to program it first
I'm not nvidia fanboy by any means I really really really want someone to make a video card that kicks the 5080's butt but my guess is the RDNA 5 will barely equaly the 5070 with ray tracing on - oh but solitaire even at 4K is just as good on the RDNA 5... Remember where there were rumors that Intel's video card was going to be on par with nvidia's latest video card and it was when using notepad.
RDNA 5 mainly needs to fix issues that got RDNA 4 Halo SKU:s scrapped while delivering the advantages of peak Chiplet design it was planned to be. But in addition to that it will deliver simpler advantages it was planned to have over RDNA 4 Halo. The leapfrogging design teams will mean RDNA 4 cancellation issues probably have less impact on RDNA 5 than some people think. It was literally cancelled to move people to work on RDNA 5 to get it working faster.
I love how everyone here has a baseless suggestion for AMD. Sometimes things take time guys, if its not amd that delivers then it might be intel . Meanwhile why cant we be happy about having 3 players competing in the gpu market now and appreciate atleast two of them for trying?
People hating on AMD saying theyll never beat nvidia when they did with RDNA2. The 6950XT was within 1% of the 3090 while using 50-60 less watts. Also, THREE HUNDRED AND EIGHTY COMPUTE UNITS? Thats INSANE
RDNA4 isn't even out yet, can we hold the bulls*** until RDNA4 is launched at the very least, because even if I believe you and your sources have real information everything is highly subject to change, so there is no real point of leaks this early except for the most fundamental changes just around the architecture.
we didnt even see any of actual RDNA4 paper. only 3.5+new RT. and this is also according to amd says before "rdna4 will have new RT but other remain the same"
Love the work you do Paul but I’m sorry I have to disagree with the hype, the gpu market will never be as exciting as it once was coz what Jim of AdoredTV said a few years back still holds true: there’s just no point for AMD competing with Nvidia in the high end consumer market even though they can because nobody is buying their stuff even when they were technologically ahead. They make more money selling CPUs on the consumer side anyway (CPUs don’t require other components unlike GPUs that require ram, pcb, etc.), meanwhile high end Radeon GPUs only make sense in the data center where companies write their own programs and can utilize their cheaper solutions since CUDA has way better support for desktop/workstation applications and AI despite all their recent efforts with ROCm. What I think AMD is instead intent on doing is flooding the market with custom solutions for x86 handhelds and laptops, atleast that’s what I can assume why they are working on Strix Halo-level of APUs. I think they’re doubling down on very powerful APUs because they know they have an advantage in the handheld, mobile and mini PC markets since Intel graphics can’t compete, Nvidia can’t make CPUs that can compete with x86, and Apple can’t get AAA game studios to develop for their ecosystem coz people don’t buy Apple products to game. They will probably still invest in R&D for graphics and compute but it’s highly likely they will never compete in the high end discrete graphics cards market again. And I seriously hope I’m wrong about that. AMD needs to rebrand hard that they’re just not cheaper alternatives so they can sell more stuff because as consumers we need more options.
AMD can make money by selling CPUs that can play with NVIDIA RTX graphics cards. For example, the RTX4090 sells for $1,699 and the 7800X3D also sells for $1,699.
Tbh I think they have to worry about CPUs too. It's not that Nvidia doesn't make them, it's that Nvidia does ARM. Under a limited shared power budget like a handheld, an ARM APU would trash x86 almost every time, and Microsoft is finally making the moves to slowly make this work. Apple has forced everyone to rethink and Nvidia's got years of advantage over AMD on ARM, and years of advantage over Qualcomm on GPUs (Intel's not even in the discussion). Even enterprise compute will go this way as electricity dominates the lifetime cost. AMD has to get on this because an artificial x86 duopoly is a house of 2 cards that cannot stand forever
@@michaelrose5696 "Apple has forced everyone to rethink"........Your "everyone" does not include the "x86 camp". Who authorized you to represent everyone? Apple just went back to the old path of isolation and happened to choose ARM. As for when will x86 die? Stop releasing new instruction sets and new technologies and AI can When it comes to completely replacing humans in writing programs, even if AI can completely replace humans in writing programs, the cost of communicating with the AI will not be reduced because the more powerful the AI is, as an administrator, you will become more of a loser, someone has been complain about AI generated code too difficult to read because it include too much abstraction layer...We are taught that the lower the dependencies of the code, the better. But when we encounter an AI that doesn't talk back, we dislike it so much.
NVIDIA has used geometry-based marketing pressure tactics since the original GeForce 256's hardware T&L vs 3DFX. NVIDIA used tessellation geometry marketing pressure tactics against the Radeon team. NVIDIA used BVH raytracing geometry (box, triangles) marketing pressure tactics against the Radeon team. With high-profile 3rd party developers, NVIDIA will focus on geometry pressure against competition. NVIDIA's one of its co-founders was SUN GX geometry processor designer.
Sorry, AMD releases RDNA 3 with chiplets, they SAY it costs less to manufacture even though the interconnect technology isn't cheap so I'm very skeptical AMD saved enough money for chiplets to be worth it, and then next once you jump up to high speed gaming, where having cache on another die is going to create a limit means Nvidia had an easy time handing AMD their asses. So, AMD moving to RDNA 4 SCRAPED their MCM parts and said they weren't going in the right direction. Meanwhile from Nvidia the ONLY rumors about MCM is the fact that they're testing it, but there is NO intention on it ending up in any product for the foreseeable future (the future that rumors can start spreading about new products). So, the company that has handed AMD their asses has no intention at this time, meaning their next 2 gen releases of using MCM, and AMD had to stop RDNA 4 MCM has pretty much shown that for now MCM isn't feasible for desktop gaming, means to me that MCM is DEAD for desktop gaming for now, and that INCLUDES RDNA 5 because the ONLY way AMD is going to pass Nvidia in performance is NOT making a design that has LATENCY ISSUES BECAUSE YOU HAVE TO PASS DATA BETWEEN DIE WHICH IS NEVER GOING TO MATCH THE SPEED OF PASSING DATA ON DIE!!!!!!!!!!!!!!! Sorry, don't believe it. The connection techniques that are feasible for MCM on graphics costs too much to make it worthwhile for one, and adds latency into the GPU for two and they will NEVER catch Nvidia using MCM. They can catch Nvidia using a monolithic GPU die with process nodes with ever smaller transistors and if they want to get an advantage over Nvidia it's going to be JUST LIKE what happened with RDNA 2, where it was on par or beating Ampere except in RT performance, and that was because AMD was on TSMC N7 and Nvidia was on Samsung 8N. But for current gen, Nvidia is spanking AMD because Nvidia NOW HAS the process node advantage. Process nodes has a LOT to do with everything, and if AMD is on TSMC N2 for RDNA 5 and Nvidia can't get onto that node, advantage AMD if they DON'T use MCM. And for those that don't get it, you can NEVER transmit data BETWEEN die at the speed you can transmit data ON die. That means latency is added for whatever data that is. Considering GPU processing is never fast enough and why people are always wanting new GPUs, for the LIFE of me I can't figure out how a product that adds latency is desirable. Then add to this, it takes more POWER to transmit data between die than ON DIE. It's why RDNA 3 MCM GPUs use more power than Nvidia. I remember the rumors that AMD was going to beat Nvidia with MCM and use less power. As soon as I saw AMD was on TSMC N5 and Nvidia was on a custom TSMC N4 I said no way in HELL, that it was going to be the inverse, that AMD GPUs would use more power and Nvidia GPUs were going to beat AMD. And that's what happened. And it was SO bad for AMD that Nvidia has been able to charge a fortune for Lovelace and are making crazy profits considering that sales aren't great for desktop GPUs right now.
8k. Past that the human can’t tell much difference. So once we get 8k at 60fps we’ll start hitting a wall graphics fidelity wise that most people wouldn’t be able to tell the difference. From then on it’ll just be niche people with insane amounts of money posting “x game at 16k!!!!” On TH-cam.
Translations: MCM is not good enough in RDNA 4, so AMD is shifting focus to hopefully get MCM working on RDNA 5. Nvidia has not gotten MCM working for gaming cards, either. Clearly, the challenges, likely power consumption due for data transfers, are too much for gaming MCM. We may need a tech breakthrough befit we really see it.
Considering Nvidia has several multiple times the resources for R&D than AMD has, it's not realistic to expect AMD to overtake at the high end. I think AMD would need to make certain breakthroughs that Nvidia does not if they are to streak ahead. That's a long shot, but not impossible. Therefore, perhaps AMD should not be judged by whether it overtakes or not. We should probably judge by whether or not AMD has become more relevant or less relevant against Nvidia as a competitor. There's several factors of relevance, plus an overall one, e.g., FPS scores, efficiency, price versus performance, etc, etc. If AMD gets closer every generation or so, then that is success I'd say, especially considering how much smaller they are. P.S. I switched to AMD several years ago because I did not like how Intel and Nvidia had effective monopolies which was not great for consumers, e.g., "you only need four cores for gaming" etc. AMD did not perform as well, but have managed to become more relevant overall over those several years -- and even took the lead in CPUs at times. I like backing the underdog if they deserve the support, which I'd argue that AMD mostly do. BTW. They did just appear to let Intel off the hook with the R7 9600X launch though. Mixed reviews so far. Good CPU, but R9 better do better if they want me to be compelled to buy it (I still use a R9 5900X). P.P.S. Your black singlet could be mistaken for you wearing a black dress and having man boobs. Not judging lol
We as consumers need amd to succeed so we can get good pricing on graphic cards with nvidia. I dont understand why everyone wants them to fail. In other words we need internet Explorer to get chrome.
After 10 years of "hoping" many are just jaded. Without congress getting involved AMD doesn't actually stand a chance. Which wont happen with Nvidia being a US based company and AMD not being tied here as closely.
You want used market prices...keep in mind retail prices have to account for distribution, warranty ect...for what they would have to sacrifice to launch a $100 card for the retail market, it is clear they are fine with leaving that to I-gpu's found on processors instead.
If AMD can at 3rd attempt get Chiplet working on GPU they could have better market segmentation with lower costs which should be better for consumers. Why 3rd attempt 1st was RDNA3 High end(Last minute bug ment couldn't run it as fast as planned) RDNA 4(Bugs caused Management to Abort it before birth) so RDNA5 will be chiplet relaunch but I don't think we will see the real gains till RDNA6 which will be like Zen3 moment. RDNA5 draw close/Level with Nvidia, RDNA6 pass Nvidia at lower price and Software stack match with new features.
So we are going back in time ... fixed function pipeline for fast execution but little flexibility. The GPUs since DX9/10 and OpenGL3/Vulkan went the other route, giving a lot of flexibility in how to program stuff and becomming more general compute units. That's what I got from your "improvements" in RDNA5 ...
Given RDNA4 is only mid to low end that should be kind of obvious. Hi end buyers are not interested in RDNA4 beyond how the generational improvements will roll into the higher CU/WGP count of RDNA5's hi end SKU.
I remember when people said rdna 3 will be huge leap but reality only equivalent with rtx 3090 in ray tracing, and zen 5 will be 24 cores and 25 percent ipc increase, but reality only 16 percent ipc and has no clockspeed increase plus it is still 16 cores😂😂😂😂
There is no Radeon GPU equivalent to RTX 3090 in ray tracing. Not Even Fkin Close. I ended up with 3090 after waiting for AMD for almost three years. And I hate nVidia.
Without a doubt, this year will be worse than the last one. I lost a lot of money last year as a result of poor investing decisions that I would never have made if I hadn't been so concerned about my portfolio. I kept investing, but I wasn't sure whether I should start saving for a house. In the end, I sold everything I had, and the house required more maintenance than I had anticipated. I'm not sure how much longer I can go on like this
It's true. Together with the help of a financial planner, I was able to diversify my portfolio across several markets and produce a net profit of over one million dollars from high dividend yield bonds, ETFs, and stocks. It is essential that you have a variety of exposures, especially with businesses that are already profitable
This is awesome! Curiosity made me look her up on Google. Found her bio, I’ve gotta admit her credentials are truly impressive, good enough to lay any doubts to rest.
I can see AMD going with Samsung for packaging RDNA5. or they are going to be using Samsung GAA for Next gen MIx00. Samsung already have packaging/silicon manufacturing
I would say... How AMD is (sadly) "almost there but not quite" like the last 40 Years. Every single GPU from ATI or AMD was always "almost there but it lacks that little bit to catch up".
I'm pretty sure most people would be willing to spend up to 500$ MAXIMUM. Spending more for graphics card is just bonkers insanity. But the problem is... today 500$ can barely buy you mid range GPU's
By the time Radeon 8000 Gpus arrive, the Nvidia 5080 will be released, which means the price for the 4080 Super should drop down closer to $500. Nvidia has not sold much of 4000 series cards, most have been 4090s with some being 4080s. Will Nvidia trickle out 5000 series to keep 4000 prices high? Retailers may want to get rid of stock, stats show only 160,000 4000 gpus have been sold.
More people have bought 4090's than any single AMD card...make sure to detail what currency your $500 is. $500 CAD gets you a 4060ti 8gb which while a bit short in vram, performs admirably at 1440p, and slightly faster than what you find in the current gen consoles. It's faster than the 2080s that launched at $1100 CAD.
Hey, remember when the narrative was that their chiplet architecture was going to mean that RDNA3 was going to leave Nvidia in the dust, AND be cheap to manufacture due to the smaller die sizes? Remember how that didn't happen, like by a long shot? Remember when they instead decided to play silly games with naming and pretended they had a card worthy of a 9x moniker when it wasn't even competitive with the neutered (small memory bus) 8x cards? Remember when they decided to play sillier games with pricing which has caused them to hemmorage market share to the point where they are an also-ran in a duopoly? Remember how this is like their fifth generation in a row where they've been on the decline? What makes you think round six is going to be any different? P.S. I'm a zen owner, and used ATi/AMD cards exclusively in my home PCs from the 1990s through 2017, but since then AMD has been just plain uncompetitive. I'd like them to give me a reason to choose them again, but they just can't seem to stop making themselves the inferior choice.
Its only going to match like a 4080 in ray tracing but then you have nvidia going above them in everyway yet again lol also rdna5 has been leaked to not be as good as even the 7900xtx in real world usage. Lets justhppe amd doesnt mess with image quality.
@raskolnikov6443 rdna5 leaks from leakers like kypo absolutely said it won't be as powerful as the 7900xtx but it will be 2x ray tracing performance. There goal is to do more what nvidia does and then bring in the power later I guess.
The technical advancements are certainly interesting ! But since Nvidias pricing is more an insult these days, my interest in Nvidia cards sadly went down to 0. RT is nice - but still in its infancy and more a gimmic than anything else. As for anything else AMD has the lead in the price/perfomance point of view. ... at least in my opinion
Bunch of nVidia fanboys in the comments. AMD is making huge strides, it just takes time. I'm sorry that AMD can't come out with a GPU next month that beats the 4090
@@patman3310 The only AMD card that has higher user count than the 4080 is the rx 580...let that sink in...The 4090 is ahead of ANY released AMD card in user count...Without some proof of AMD cards not being registered by steam, that is the most accurate measurement available to the public.
look, you're impressively swole and all, but we've seen this movie before and know how it goes. Radeon isnt Ryzen and and good lord Nvidia isnt Intel. what I mean by that is no matter how greedy and scummy Jensen is, Nvidia really never stops innovating and don't often slip. they certainly throw crumbs at us repeatedly but it's because they know theyre in a position to throw crumbs when AMD at best plays follow the leader when they have a good product on pricing while having worse features, while at worse put out garbage flaw messes that that still are overpriced. I know its super tinfoil hat, but I 1000 percent believe Lisa Su and Jensen Huang, who are actually literally family, have some kind of deal. because whenever Nvidia put out something mediocre and AMD has something in parity like RDNA2, AMD charge the same price to performance or less by only $20. whenever Radeon shits out something broken and awful like RDNA3, Nvidia sandbags their shit and renames their lower end dies higher end and sells them for more. and the latter is definitely just to keep Radeon in business so Nvidia doesn't get regulated as monopoly and justifiably broken up. nothing else to me makes sense any other way to me on how Radeon can be so dented head and miss the easiest and most critical lay ups, while Ryzen has been so successful and their climb so methodical. it's also why Nvidia has NEVER given a damn about AMD Radeon while the moment Intel ARC even became a concept Nvidia was shitting bricks. nothing else really makes sense about the handling of Radeon till you consider being mismanaged on purpose. just feels like radeon only exists to keep Nvidia from being broken up, price fix with Nvidia, and sell to Sony and Microsoft. nothing else makes sense and any other business would've folded Radeon by now and focused entirely on their crazy successful over a less than a decade CPU division
I have now counted your facial hair follicles. Your face hair amounts to a total of 114 single hair follicles with one double haired. A nice preteen lvl of 115.
Ok, so , you've got yourself some muscles, congratulations! Now, that doesn't mean you can just rock your wife's blouse and nobody is gonna say anything.
So AMD wants to (again) double performance between generations... And start releasing RDNA5 in the middle of Blackwell lifecycle? I dont like it. But if they can't beat NVIDIA in synchronized release cadence then I guess it is some way arund it. To be fair it is not giving me much confidence in Radeon, but maybe I'm an exception there.
RDNA 5 will only comes out when nvidia release 60 series. AMD in general did not like to shorten the life cycle of their product. If anything they like to extend them futher to get the best return from their initial investment.
Everyone talks about whether or not AMD can make a halo card that beats Nvidia’s but what they really mean is can AMD make a $1000 card that beats Nvidia’s $1600-3000 card and if they don’t it’s some great failure. Such a comparison obviously shouldn’t matter but most of the market is either fanboy or mindshare casual.
He got some compliment for being in good shape. Can't blame him. Also a surprising number of tech tubers film in a finished attic or side room with little to no ac.
Cos he looks dayummm good, i proper rate someone being in shape and also being a tech guy, its a really good to be healthy while being involved in a industry that doesnt require massive physical activity xd
Haven't even watched the video yet, but I'll believe it when I see it. There is always this hype train that describes how AMD is finally gonna beat Nvidia in the next generation. But it never happens. Don't get me wrong, they've come a long way. But they always either royally screw something up or they fall short somehow. The 7000 series was more of a screw up. It sure seems like the 7900XTX was originally intended to compete with the 4090 but when they realized it wasn't going to match up in performance, they decided to price it and compete with the 4080. The 7800XT is basically no improvement over the 6800XT except in price. But spec for spec, it seems like more of a replacement for the 6800 rather than the 6800XT. I'm glad their price to performance and VRAM is better than Nvidia. I just wish they would really compete with Nvidia across the entire stack.
Isn't it a bit depressing to have to say dumb things on a daily basis? Nvidia is significantly ahead, famous for being extra competitive, has largest war chest ever and some of the best engineers on Earth. But yeah, frogs will leap.
Red, stop that sh, man. Radeon still struggling to get drivers right since 2013. 7900xtx is one of the biggest disasters of the cards in graphics history. I never crashed as much as I did with 7900xtx, 2 of them by the way. The day Radeon beats Nvidia is about as far as their driver writting being complete, which is never man. I used to own Radeon 7950 HD in 2013 and it crashed less than 7900xtx We regressed. Budget gpu means cheaper, not faulty People need to address Radeon failures as they address Intel's 14900k failures. It is just as blatantly bad
Navi50 may have up to 9 shader engines chiplets + 1 central chiplet, so 360 is possible, 380 is not. Same ALUs per dual CU as RDNA4. This should stomp the 192 SM of the RTX 5090 TITAN. Keep in mind dynamic register caches mean the CUs will have a higher occupancy and thus will be more efficient.
@@louisfriend9323 Agree but I don't know about stomping Nvidia. Let's see if Jensen doesn't invent new ray tracing tomorrow so 5090 is 2x faster again. Also, even if it's not 13 chiplets but 4 (1 central + 3x120CU) that can be cut to smaller parts - that's a lot of chiplets - there will be penalty. Or I see it as possible that rdna5 will start unoptimized, below 30% on average, and, analogical to Intel A770, there could be 9x fps updates drivers that would use all CUs. My questions are: if the 13 chiplets gpu is a thing 1. does it mean 120cu is target for ps6, 2. will there be integrated but very expensive 120CU GPU in 2026, 3. will AMD not delay it launch further, 4. Could it mean the same chiplets for integrated and discrete gpus, 5. would tsmc be the only place to print it? You can share your guesses if you have a spare time.
The next RDNA 4 arch is not out yet, we know it will be far behind, and I’m supposed to hold my breath for the next-next one? I’m sure they are _really_ worried about this at Nvidia. Cut the BS.
![ILLSLICK - THE KING'S LANDING [Freestyle]](http://i.ytimg.com/vi/feIJJnNkTfk/mqdefault.jpg)
I'd really appreciate your support of my brand new fitness channel! Check it out here - www.youtube.com/@MuscleFoundryFitness
Your comment is 8 min ago but the video is somehow 1 min ago
Yeah, you can post comments on unlisted videos. It's just so I don't forget to pin!
@@RedGamingTech I didn’t know you could do that. Pretty cool
@@RedGamingTech You Juicing? You're looking very good, if you took some winstrol or dbol you would be a beast.
Cant believe we peer pressured him into it! im subscribing.
And when it doesn't accomplish it, we'll see videos on how RDNA 6 is the architecture that will get it done.
Agreed. I think amd can do it but maybe they will have to have a team design a WHOLE new architecture beyond the rdna name.
I hope they are working on something
Such is the cycle. Always been this way, always will be this way. TH-camrs gonna get their clicks in the meantime.
I am betting on 7 😂
All the talent AMD has is on the Ryzen side of the company. Not the Radeon side.
@@HablaConOwens, or they can buy Nvidia cards in bulk and brand it with RDNA name. No one will know
Every generation" amd is going to beat nvidia" buy actually they never do
But they did with the 6950XT didnt they. It was more efficient than the 3090 and was just as fast. The 3090 Ti was faster yes, but used a boat load more energy.
@@Dave-dh7rtactually it was only faster at 4K, at 1080p or 1440p even the 6900XT matched it with the 6950XT creeping about 10% ahead
@@Dave-dh7rt 3090 beats 6950XT in raytracing.
@@valenrn8657 in reality amd a bit lucky intel has instability issue in cpu but for radeon vs nvidia or even intel arc theyre struggling with their architecture for overall performance
@@Dave-dh7rt did they ??? NVIDIA have largest mindshare, its very impossible when NVIDIA does lot more in architecture than AMD simple FSR tech that seem to me it took them 3 rdna gens to make it work. Not liking amd doing vs NVIDIA did it in day one and not wait 3 architecture chips to fix it ..what disspointing from AMD ...they wasted everyone time to get RT tracing to work...5 gens of RDNA to fix it ...wow...why AMD struggle something simple ?? Now Path tracing is different story
...but which nVidia generation will it leapfrog?
Not even Ada tbh.
Pascal maybe
The Encoding quality of the GTX 970 Series (maybe).
Our host´s traps have made generational leap.
Y'all are being mean. It'll catch up with turing. Seriously though.. I love AMD and all I've every wanted is to have solid AMD products to buy so I don't have to resort to buying an $800 super ti GPU. but that is exactly what I just did. At least it will be worth $400+ in a few years when I upgrade.
Heard it all before. If RDNA5 was this year, it would be something, but 2026... yeah they might have the performance crown for a few months tops before they are once again runner up. It doesn't matter though, as long as they maintain the lowest cost per frame, it's fine. Also Paul, put on a shirt already.
AMD has had the price-to-performance advantage for several generations in a row now when it comes to raster, but they’ve been behind on RT, upscaling, downsampling, and notably also CUDA. At this point all of that matters a lot more when it comes to pressuring Nvidia.
Make the same AMD video every launch and it will eventually come true
😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂
True
It needs to happen. Nvidia is like intel was in the early 2010s. AMD is holding back the whole industry by not making a card that competes with Nvidia's top perforforming gpu. If AMD didn't release ryzen, we'd probably still be on 4c8t with the i7s and i9s locked away on a much more expensive socket.
Basically Ryzen. AMD started making major changes with rdna2, rdna4 is an optimization change, and 5 is the big picture. I don't think Nvidia has any actual new ideas, only cares about AI, dropped old feature support, and will ultimately completely stagnate, while AMD is pushing hard to compete. Not hard to see a turn around like AMD vs Intel.
@@JohnDoe-ip3oq That would be nice but the boy who cried wolf and all that. Won't get hyped until its in reviewers hands
I clicked on this video for the comedy
I remember one year ago I waited and hoped that AMD would make a chiplet beast, but they didn't.
Chiplet base gaming GPU is very hard to do.
@@kazedcat They are learning. RDNA 3 was just the beginning of the journey. RDNA5 will include and interate upon those learnings.
Close but no cigar. They will get closer but can't compete
*RDNA 5 is saving money, to fully stick into RDNA 6, which will be an on a total new level !* Nvidia will shit their pants FULL !
Hmm an AMD Radeon GPU with 4080 performance at half the current price would not be bad not at all 😯
4nm RDNA 4 with AFMF
4080 perf seems likely since 7800XT is like 70-75% the perf already. Shrink down, go wider, go faster, go smarter, go mono, and yeah, doesn't 4080 seem doable and likely? Also exposes that 4080 is basically twice the price it *could* be.
Thats a 7900xtx
In 2026-7?
@attepatte8485 ihace 7900xtx and its performance of between 4080 and 4090 but 4090 beats itvraytracing but its rasterisation is alot faster and fsr 3 gas made bug sooed inprovements
AMD really needs to focus on the value of their offerings. Nothing good can come out of AMD trying to meet or beat Nvidia. Nvidia could spend AMDs market cap in a quarter on R&D to release a faster product and not blink an eye. AMD cannot afford to have an entire generation of R&D negated before they can recoup costs. They're still not that on that level.
Hey say it louder so the Radeon Fanboys on Reddit might hear it.
@@GRIGGINS1I am an AMD fanboy. I had to buy Super TI tho cuz... AMD 😭
AMD is focusing by ignoring high end gaming. Gamers with money only buy Nvidia and there is no point chasing after them.
Here's another non-believer. AMD is already taking the data centers. You need to do thorough reading my guy.
@@yasserq78 At what point were we ever talking about datacenters? They're also not winning because of performance. They are winning because of openness. The datacenter equivalent of value.
Haven't watched any vids on this channel for a while and bloody hell you went from a ghost to jacked. Awesome.
I would expect RDNA IV sooner or perhaps R9K will dribble out. The function of RDNA lV is R9K attach and their volumes need to sync and something from this report seems leading and lagging. I already register RDNA lV components on interposer or SIP produced on production in relation financial assessment. mb
I don’t care about it being better than nvidia but about how well it will do in the ps6.
Zen 6 and rdna5 with ddr7 will be awesome
ok , time for me to do some theorizing form a electrical engineering student .
If amd does a mini rdna 4 gen , only launching mid tier rdna 4 and pushing rdna 5 as fast as possible.
RDNA 5 might be the first gpu gen that will use Backside power delivery with N2P , A16 gets backside power in 2026/2027. That alone increases the power efficiency and clock speed(around 6-7% from intel's current chip testing with backside power) regardless of architecture . On top of a generational Uplift in performance . On top of the possibility to even add 3d Vcashe to improve communication between chaplets( just a theory ) + gddr 7 .
Amd might do a moon shot in early 2027 , Sacrificing top end rdna 4 to Launch Rdna 5 early , before Nvidia can respond with 60 series , forcing them to do another super series . but this time Rdna 5 might just be Far enough in Silicon tech that it will beat it with no contest . AMD gpu department might finally grow a pair of balls and try to Out compete Nvidea before launch , and not with post launch price cuts.
Trying to beat nvidea with there heavy connections with TSMC , while rdna 4 is just there for the driver team to slowly upgrading Fsr ,anti lag, and frame gen.
Saw this and had to watch immediately
Love the Mass Effect wallpaper
also your correct...they are using CDNA architecture as band aid architecture exit from rdna4....yes rdna4 will be last rdna architecture by raja koduri the lead architecture designer for navi first gen that amd keep beating the dead horse since ...man they really surprise me to keep gonna till rdna4...4 generation of raja work...while CDNA architecture been around since R9 series using older tech that seem work better in programmable shaders that they can make this work for Ray tracing tech ....but not in raja work they realize by now..By early ATI unfinished work long ago that been around since AMD continue since the islands architecture ...not polaris architecture...that lead to navi afterwards from star names brand for consumer while pro's still using furry X tech architecture compute heavy programmable shader tech that requires compilers from the drivers to program it first
I'm not nvidia fanboy by any means I really really really want someone to make a video card that kicks the 5080's butt but my guess is the RDNA 5 will barely equaly the 5070 with ray tracing on - oh but solitaire even at 4K is just as good on the RDNA 5... Remember where there were rumors that Intel's video card was going to be on par with nvidia's latest video card and it was when using notepad.
And what of the GPUopen development on raytracing on radeon. Or RDNA 4's many changes to RT cores and higher promised performance.
RDNA 5 mainly needs to fix issues that got RDNA 4 Halo SKU:s scrapped while delivering the advantages of peak Chiplet design it was planned to be. But in addition to that it will deliver simpler advantages it was planned to have over RDNA 4 Halo. The leapfrogging design teams will mean RDNA 4 cancellation issues probably have less impact on RDNA 5 than some people think. It was literally cancelled to move people to work on RDNA 5 to get it working faster.
I love how everyone here has a baseless suggestion for AMD. Sometimes things take time guys, if its not amd that delivers then it might be intel . Meanwhile why cant we be happy about having 3 players competing in the gpu market now and appreciate atleast two of them for trying?
People hating on AMD saying theyll never beat nvidia when they did with RDNA2. The 6950XT was within 1% of the 3090 while using 50-60 less watts. Also, THREE HUNDRED AND EIGHTY COMPUTE UNITS? Thats INSANE
we still didnt see RDNA4 paper so it is no real RDNA 4 then ? only 3.5+new RT ?
RDNA5 will be good but thats because just like with RDNA2 it will be co developed by Sony , MS and AMD on the nextgenerations of consoles XD
RDNA4 isn't even out yet, can we hold the bulls*** until RDNA4 is launched at the very least, because even if I believe you and your sources have real information everything is highly subject to change, so there is no real point of leaks this early except for the most fundamental changes just around the architecture.
Hey if you don't want this guy to track the whole industry then don't follow
we didnt even see any of actual RDNA4 paper. only 3.5+new RT. and this is also according to amd says before "rdna4 will have new RT but other remain the same"
RDNA 5 but... only 2026!
Love the work you do Paul but I’m sorry I have to disagree with the hype, the gpu market will never be as exciting as it once was coz what Jim of AdoredTV said a few years back still holds true: there’s just no point for AMD competing with Nvidia in the high end consumer market even though they can because nobody is buying their stuff even when they were technologically ahead. They make more money selling CPUs on the consumer side anyway (CPUs don’t require other components unlike GPUs that require ram, pcb, etc.), meanwhile high end Radeon GPUs only make sense in the data center where companies write their own programs and can utilize their cheaper solutions since CUDA has way better support for desktop/workstation applications and AI despite all their recent efforts with ROCm. What I think AMD is instead intent on doing is flooding the market with custom solutions for x86 handhelds and laptops, atleast that’s what I can assume why they are working on Strix Halo-level of APUs. I think they’re doubling down on very powerful APUs because they know they have an advantage in the handheld, mobile and mini PC markets since Intel graphics can’t compete, Nvidia can’t make CPUs that can compete with x86, and Apple can’t get AAA game studios to develop for their ecosystem coz people don’t buy Apple products to game. They will probably still invest in R&D for graphics and compute but it’s highly likely they will never compete in the high end discrete graphics cards market again. And I seriously hope I’m wrong about that. AMD needs to rebrand hard that they’re just not cheaper alternatives so they can sell more stuff because as consumers we need more options.
AMD can make money by selling CPUs that can play with NVIDIA RTX graphics cards. For example, the RTX4090 sells for $1,699 and the 7800X3D also sells for $1,699.
Tbh I think they have to worry about CPUs too. It's not that Nvidia doesn't make them, it's that Nvidia does ARM. Under a limited shared power budget like a handheld, an ARM APU would trash x86 almost every time, and Microsoft is finally making the moves to slowly make this work. Apple has forced everyone to rethink and Nvidia's got years of advantage over AMD on ARM, and years of advantage over Qualcomm on GPUs (Intel's not even in the discussion). Even enterprise compute will go this way as electricity dominates the lifetime cost. AMD has to get on this because an artificial x86 duopoly is a house of 2 cards that cannot stand forever
@@michaelrose5696 "Apple has forced everyone to rethink"........Your "everyone" does not include the "x86 camp". Who authorized you to represent everyone? Apple just went back to the old path of isolation and happened to choose ARM. As for when will x86 die? Stop releasing new instruction sets and new technologies and AI can When it comes to completely replacing humans in writing programs, even if AI can completely replace humans in writing programs, the cost of communicating with the AI will not be reduced because the more powerful the AI is, as an administrator, you will become more of a loser, someone has been complain about AI generated code too difficult to read because it include too much abstraction layer...We are taught that the lower the dependencies of the code, the better. But when we encounter an AI that doesn't talk back, we dislike it so much.
NVIDIA has used geometry-based marketing pressure tactics since the original GeForce 256's hardware T&L vs 3DFX.
NVIDIA used tessellation geometry marketing pressure tactics against the Radeon team.
NVIDIA used BVH raytracing geometry (box, triangles) marketing pressure tactics against the Radeon team.
With high-profile 3rd party developers, NVIDIA will focus on geometry pressure against competition. NVIDIA's one of its co-founders was SUN GX geometry processor designer.
I just want AMD to be competitive.
Sorry, AMD releases RDNA 3 with chiplets, they SAY it costs less to manufacture even though the interconnect technology isn't cheap so I'm very skeptical AMD saved enough money for chiplets to be worth it, and then next once you jump up to high speed gaming, where having cache on another die is going to create a limit means Nvidia had an easy time handing AMD their asses.
So, AMD moving to RDNA 4 SCRAPED their MCM parts and said they weren't going in the right direction. Meanwhile from Nvidia the ONLY rumors about MCM is the fact that they're testing it, but there is NO intention on it ending up in any product for the foreseeable future (the future that rumors can start spreading about new products). So, the company that has handed AMD their asses has no intention at this time, meaning their next 2 gen releases of using MCM, and AMD had to stop RDNA 4 MCM has pretty much shown that for now MCM isn't feasible for desktop gaming, means to me that MCM is DEAD for desktop gaming for now, and that INCLUDES RDNA 5 because the ONLY way AMD is going to pass Nvidia in performance is NOT making a design that has LATENCY ISSUES BECAUSE YOU HAVE TO PASS DATA BETWEEN DIE WHICH IS NEVER GOING TO MATCH THE SPEED OF PASSING DATA ON DIE!!!!!!!!!!!!!!!
Sorry, don't believe it. The connection techniques that are feasible for MCM on graphics costs too much to make it worthwhile for one, and adds latency into the GPU for two and they will NEVER catch Nvidia using MCM. They can catch Nvidia using a monolithic GPU die with process nodes with ever smaller transistors and if they want to get an advantage over Nvidia it's going to be JUST LIKE what happened with RDNA 2, where it was on par or beating Ampere except in RT performance, and that was because AMD was on TSMC N7 and Nvidia was on Samsung 8N. But for current gen, Nvidia is spanking AMD because Nvidia NOW HAS the process node advantage. Process nodes has a LOT to do with everything, and if AMD is on TSMC N2 for RDNA 5 and Nvidia can't get onto that node, advantage AMD if they DON'T use MCM.
And for those that don't get it, you can NEVER transmit data BETWEEN die at the speed you can transmit data ON die. That means latency is added for whatever data that is. Considering GPU processing is never fast enough and why people are always wanting new GPUs, for the LIFE of me I can't figure out how a product that adds latency is desirable.
Then add to this, it takes more POWER to transmit data between die than ON DIE. It's why RDNA 3 MCM GPUs use more power than Nvidia. I remember the rumors that AMD was going to beat Nvidia with MCM and use less power. As soon as I saw AMD was on TSMC N5 and Nvidia was on a custom TSMC N4 I said no way in HELL, that it was going to be the inverse, that AMD GPUs would use more power and Nvidia GPUs were going to beat AMD. And that's what happened. And it was SO bad for AMD that Nvidia has been able to charge a fortune for Lovelace and are making crazy profits considering that sales aren't great for desktop GPUs right now.
I'm not sure if someone can help me. I'm still new. I have a quick question. I'd the final resolution end goal 8k or 16k, thanks ?
8k. Past that the human can’t tell much difference. So once we get 8k at 60fps we’ll start hitting a wall graphics fidelity wise that most people wouldn’t be able to tell the difference. From then on it’ll just be niche people with insane amounts of money posting “x game at 16k!!!!” On TH-cam.
Gpus won’t be needed past then except for vr as I understand it, so they’ll just pivot more to AI
@user-ri3ho1yh5p cool thanks big help I always had a feeling it was 8k but never quit sure
I think Paul has been replaced. The Robert Smith model has been substituted for the new Arnie one
Have you been working out? Looking jacked!!
Translations: MCM is not good enough in RDNA 4, so AMD is shifting focus to hopefully get MCM working on RDNA 5.
Nvidia has not gotten MCM working for gaming cards, either. Clearly, the challenges, likely power consumption due for data transfers, are too much for gaming MCM. We may need a tech breakthrough befit we really see it.
wanted rumors with RDNA, got a gunshow instead
Considering Nvidia has several multiple times the resources for R&D than AMD has, it's not realistic to expect AMD to overtake at the high end. I think AMD would need to make certain breakthroughs that Nvidia does not if they are to streak ahead. That's a long shot, but not impossible.
Therefore, perhaps AMD should not be judged by whether it overtakes or not. We should probably judge by whether or not AMD has become more relevant or less relevant against Nvidia as a competitor. There's several factors of relevance, plus an overall one, e.g., FPS scores, efficiency, price versus performance, etc, etc. If AMD gets closer every generation or so, then that is success I'd say, especially considering how much smaller they are.
P.S. I switched to AMD several years ago because I did not like how Intel and Nvidia had effective monopolies which was not great for consumers, e.g., "you only need four cores for gaming" etc. AMD did not perform as well, but have managed to become more relevant overall over those several years -- and even took the lead in CPUs at times. I like backing the underdog if they deserve the support, which I'd argue that AMD mostly do.
BTW. They did just appear to let Intel off the hook with the R7 9600X launch though. Mixed reviews so far. Good CPU, but R9 better do better if they want me to be compelled to buy it (I still use a R9 5900X).
P.P.S. Your black singlet could be mistaken for you wearing a black dress and having man boobs. Not judging lol
We as consumers need amd to succeed so we can get good pricing on graphic cards with nvidia. I dont understand why everyone wants them to fail. In other words we need internet Explorer to get chrome.
After 10 years of "hoping" many are just jaded. Without congress getting involved AMD doesn't actually stand a chance. Which wont happen with Nvidia being a US based company and AMD not being tied here as closely.
💀 i want a cheap card under 100. And it has to be a 3050 8gb like performance.
Just look for used stuff. For brand new or upcoming new release AIB most likely did not want to release anything cheaper than $300.
You want used market prices...keep in mind retail prices have to account for distribution, warranty ect...for what they would have to sacrifice to launch a $100 card for the retail market, it is clear they are fine with leaving that to I-gpu's found on processors instead.
If AMD can at 3rd attempt get Chiplet working on GPU they could have better market segmentation with lower costs which should be better for consumers. Why 3rd attempt 1st was RDNA3 High end(Last minute bug ment couldn't run it as fast as planned) RDNA 4(Bugs caused Management to Abort it before birth) so RDNA5 will be chiplet relaunch but I don't think we will see the real gains till RDNA6 which will be like Zen3 moment. RDNA5 draw close/Level with Nvidia, RDNA6 pass Nvidia at lower price and Software stack match with new features.
bro looking beautiful 0:13
So we are going back in time ... fixed function pipeline for fast execution but little flexibility. The GPUs since DX9/10 and OpenGL3/Vulkan went the other route, giving a lot of flexibility in how to program stuff and becomming more general compute units. That's what I got from your "improvements" in RDNA5 ...
AMD Engineers = RDNA4 is coming out, CEO and Investors = Tease RDNA5 😈😈
Absolutely. They most certainly control leaks.
Given RDNA4 is only mid to low end that should be kind of obvious.
Hi end buyers are not interested in RDNA4 beyond how the generational improvements will roll into the higher CU/WGP count of RDNA5's hi end SKU.
I remember when people said rdna 3 will be huge leap but reality only equivalent with rtx 3090 in ray tracing, and zen 5 will be 24 cores and 25 percent ipc increase, but reality only 16 percent ipc and has no clockspeed increase plus it is still 16 cores😂😂😂😂
There is no Radeon GPU equivalent to RTX 3090 in ray tracing. Not Even Fkin Close.
I ended up with 3090 after waiting for AMD for almost three years. And I hate nVidia.
Without a doubt, this year will be worse than the last one. I lost a lot of money last year as a result of poor investing decisions that I would never have made if I hadn't been so concerned about my portfolio. I kept investing, but I wasn't sure whether I should start saving for a house. In the end, I sold everything I had, and the house required more maintenance than I had anticipated. I'm not sure how much longer I can go on like this
Every one of us has made mistakes at some point. Consider financial planning
It's true. Together with the help of a financial planner, I was able to diversify my portfolio across several markets and produce a net profit of over one million dollars from high dividend yield bonds, ETFs, and stocks. It is essential that you have a variety of exposures, especially with businesses that are already profitable
Would you mind telling me about your financial planner ?
Leah Foster Alderman.. You are most likely to find more info when you look her up online
This is awesome! Curiosity made me look her up on Google. Found her bio, I’ve gotta admit her credentials are truly impressive, good enough to lay any doubts to rest.
Nvidia believes the RDNA5 hype: why else have a RTX 5090 TITAN??
I will believe it when I see it. Color me skeptical on AMD Radeon Claims.
cant wait for nvidia to have competition at the top end, then we wont be rinsed quite so badly
I can see AMD going with Samsung for packaging RDNA5. or they are going to be using Samsung GAA for Next gen MIx00. Samsung already have packaging/silicon manufacturing
No chance purely for the fact that nvidia has a much higher budget than amd and they will actually put more money into to the development
charge twice as much, sell half as many, fire half the employees, profit.
I would say... How AMD is (sadly) "almost there but not quite" like the last 40 Years. Every single GPU from ATI or AMD was always "almost there but it lacks that little bit to catch up".
There's been several generations from ATI/AMD that have beaten Nvidia.
@@26GuenterR9 290X vs GTX 780ti, Iceberg tech just put out a side by side benchmark comparison today.
Just wait another generation....
DJ Khalid "And another one" on Repeat.
Just like rdna 3 was supposed to match rtx40? Yeah will wait for reviews before thinking anything about it
To be fair 6700xt slaps the 4060 and the 6950xt at $400 is better value than a 4070
@@starkistunathat’s RDNA2… it’s not a good look for RDNA3 if the best cards in AMD’s lineup are previous generation when next generation is coming up.
Whadda! This must be *_MoreVoltageMoreOxidated_* .
I'm pretty sure most people would be willing to spend up to 500$ MAXIMUM. Spending more for graphics card is just bonkers insanity. But the problem is... today 500$ can barely buy you mid range GPU's
By the time Radeon 8000 Gpus arrive, the Nvidia 5080 will be released, which means the price for the 4080 Super should drop down closer to $500. Nvidia has not sold much of 4000 series cards, most have been 4090s with some being 4080s. Will Nvidia trickle out 5000 series to keep 4000 prices high? Retailers may want to get rid of stock, stats show only 160,000 4000 gpus have been sold.
More people have bought 4090's than any single AMD card...make sure to detail what currency your $500 is. $500 CAD gets you a 4060ti 8gb which while a bit short in vram, performs admirably at 1440p, and slightly faster than what you find in the current gen consoles. It's faster than the 2080s that launched at $1100 CAD.
Hey, remember when the narrative was that their chiplet architecture was going to mean that RDNA3 was going to leave Nvidia in the dust, AND be cheap to manufacture due to the smaller die sizes? Remember how that didn't happen, like by a long shot? Remember when they instead decided to play silly games with naming and pretended they had a card worthy of a 9x moniker when it wasn't even competitive with the neutered (small memory bus) 8x cards? Remember when they decided to play sillier games with pricing which has caused them to hemmorage market share to the point where they are an also-ran in a duopoly? Remember how this is like their fifth generation in a row where they've been on the decline? What makes you think round six is going to be any different? P.S. I'm a zen owner, and used ATi/AMD cards exclusively in my home PCs from the 1990s through 2017, but since then AMD has been just plain uncompetitive. I'd like them to give me a reason to choose them again, but they just can't seem to stop making themselves the inferior choice.
Day 1 of asking Paul to start a youtube channel teaching us how to achieve washboard Abs in 60 days
Subscribed to MuscleFoundry.
Its only going to match like a 4080 in ray tracing but then you have nvidia going above them in everyway yet again lol also rdna5 has been leaked to not be as good as even the 7900xtx in real world usage.
Lets justhppe amd doesnt mess with image quality.
You’re thinking of rdna4 not 5
RDNA5 has not leaked
@raskolnikov6443 rdna5 leaks from leakers like kypo absolutely said it won't be as powerful as the 7900xtx but it will be 2x ray tracing performance. There goal is to do more what nvidia does and then bring in the power later I guess.
You are mixing RDNA 4 with 5.
And how can RDNA 4 match 7900 XTX when the most powerful GPU only has the same CU count at 7800 XT?
@@Shatterfury1871 its been known for some time now that the higher end rdna 4 card will have almost identical performance to the 7900 gre.
The technical advancements are certainly interesting ! But since Nvidias pricing is more an insult these days, my interest in Nvidia cards sadly went down to 0. RT is nice - but still in its infancy and more a gimmic than anything else. As for anything else AMD has the lead in the price/perfomance point of view. ... at least in my opinion
The only "RED tech" Nvidia CEO Jensen is afraid to fight against...
...is the Chinese.
Has rdna4 even come out yet??
Bunch of nVidia fanboys in the comments. AMD is making huge strides, it just takes time. I'm sorry that AMD can't come out with a GPU next month that beats the 4090
👍 for the algorithm!
I don’t see RNDA5 matching or beating Titan AI. AMD is always one or two high end GPU’s behind.
Who cares. Almost no one buys the highest end anyways.
@@patman3310 The only AMD card that has higher user count than the 4080 is the rx 580...let that sink in...The 4090 is ahead of ANY released AMD card in user count...Without some proof of AMD cards not being registered by steam, that is the most accurate measurement available to the public.
@@patman3310 Completely false.
look, you're impressively swole and all, but we've seen this movie before and know how it goes. Radeon isnt Ryzen and and good lord Nvidia isnt Intel. what I mean by that is no matter how greedy and scummy Jensen is, Nvidia really never stops innovating and don't often slip. they certainly throw crumbs at us repeatedly but it's because they know theyre in a position to throw crumbs when AMD at best plays follow the leader when they have a good product on pricing while having worse features, while at worse put out garbage flaw messes that that still are overpriced. I know its super tinfoil hat, but I 1000 percent believe Lisa Su and Jensen Huang, who are actually literally family, have some kind of deal. because whenever Nvidia put out something mediocre and AMD has something in parity like RDNA2, AMD charge the same price to performance or less by only $20. whenever Radeon shits out something broken and awful like RDNA3, Nvidia sandbags their shit and renames their lower end dies higher end and sells them for more. and the latter is definitely just to keep Radeon in business so Nvidia doesn't get regulated as monopoly and justifiably broken up. nothing else to me makes sense any other way to me on how Radeon can be so dented head and miss the easiest and most critical lay ups, while Ryzen has been so successful and their climb so methodical. it's also why Nvidia has NEVER given a damn about AMD Radeon while the moment Intel ARC even became a concept Nvidia was shitting bricks. nothing else really makes sense about the handling of Radeon till you consider being mismanaged on purpose. just feels like radeon only exists to keep Nvidia from being broken up, price fix with Nvidia, and sell to Sony and Microsoft. nothing else makes sense and any other business would've folded Radeon by now and focused entirely on their crazy successful over a less than a decade CPU division
lets go we need another vega moment hahahahaha
JFC, our boy is R I P P E D.
AMD already proven its reliability through Ryzen CPUs over Intel so it would just take some time to make it for their GPUs over Nvidia.
This is 4 years from now nvidia will have caught up
As he started to get naked on his videos i stoped watching the videos
Please name it RedGainingTech
I have now counted your facial hair follicles. Your face hair amounts to a total of 114 single hair follicles with one double haired. A nice preteen lvl of 115.
Wow, I used to be able to follow this stuff 🤠
I meam RDNA 2 was a BANGER maybe theyll repeat it around rdna 5, but ill believe it when i see it
That's because nvidia are using samsung 8nm back then.
Ok, so , you've got yourself some muscles, congratulations! Now, that doesn't mean you can just rock your wife's blouse and nobody is gonna say anything.
So AMD wants to (again) double performance between generations... And start releasing RDNA5 in the middle of Blackwell lifecycle? I dont like it. But if they can't beat NVIDIA in synchronized release cadence then I guess it is some way arund it. To be fair it is not giving me much confidence in Radeon, but maybe I'm an exception there.
At least there will be a point were they are faster for once.
RDNA 5 will only comes out when nvidia release 60 series. AMD in general did not like to shorten the life cycle of their product. If anything they like to extend them futher to get the best return from their initial investment.
Everyone talks about whether or not AMD can make a halo card that beats Nvidia’s but what they really mean is can AMD make a $1000 card that beats Nvidia’s $1600-3000 card and if they don’t it’s some great failure. Such a comparison obviously shouldn’t matter but most of the market is either fanboy or mindshare casual.
No lets talk about rdna 8. The ps6 architecture (unless sony goes nvidia and does 8K actually)
Rdna 8 will beat nvidia 5090
What's with the vests everyday?? it's not been hot here in the uk??
He got some compliment for being in good shape. Can't blame him. Also a surprising number of tech tubers film in a finished attic or side room with little to no ac.
Cos he looks dayummm good, i proper rate someone being in shape and also being a tech guy, its a really good to be healthy while being involved in a industry that doesnt require massive physical activity xd
You mean dress? 👗 😅
Does it matter how promising AMD Radeon is, AMD will screw it up somehow.
Haven't even watched the video yet, but I'll believe it when I see it. There is always this hype train that describes how AMD is finally gonna beat Nvidia in the next generation. But it never happens.
Don't get me wrong, they've come a long way. But they always either royally screw something up or they fall short somehow. The 7000 series was more of a screw up. It sure seems like the 7900XTX was originally intended to compete with the 4090 but when they realized it wasn't going to match up in performance, they decided to price it and compete with the 4080. The 7800XT is basically no improvement over the 6800XT except in price. But spec for spec, it seems like more of a replacement for the 6800 rather than the 6800XT.
I'm glad their price to performance and VRAM is better than Nvidia. I just wish they would really compete with Nvidia across the entire stack.
Isn't it a bit depressing to have to say dumb things on a daily basis?
Nvidia is significantly ahead, famous for being extra competitive, has largest war chest ever and some of the best engineers on Earth. But yeah, frogs will leap.
Yawn. Wake me up when it actually happens.
0% chance AMD leapfrog Nvidia.
Be competitive for a generation, sure.
Get out in front, no.
AMD "kicked in the balls" by NVIDIA with the 50 series? Does that mean AMDs' response with RDNA5 will be "AM Deez Nuts"?
Red, stop that sh, man.
Radeon still struggling to get drivers right since 2013.
7900xtx is one of the biggest disasters of the cards in graphics history. I never crashed as much as I did with 7900xtx, 2 of them by the way.
The day Radeon beats Nvidia is about as far as their driver writting being complete, which is never man.
I used to own Radeon 7950 HD in 2013 and it crashed less than 7900xtx
We regressed. Budget gpu means cheaper, not faulty
People need to address Radeon failures as they address Intel's 14900k failures. It is just as blatantly bad
_"rDnA 5 LeApFrOgS nViDiA"_
lolololooooolololoool
here we go again 😁😁🤡🤡🤡🤡🤡
well need to see it to believe it especially beating Nvidia
is 360CU number confirmed for rdna5? 256alus per cu for rdna4?
Navi50 may have up to 9 shader engines chiplets + 1 central chiplet, so 360 is possible, 380 is not. Same ALUs per dual CU as RDNA4. This should stomp the 192 SM of the RTX 5090 TITAN.
Keep in mind dynamic register caches mean the CUs will have a higher occupancy and thus will be more efficient.
@@louisfriend9323 Agree but I don't know about stomping Nvidia. Let's see if Jensen doesn't invent new ray tracing tomorrow so 5090 is 2x faster again. Also, even if it's not 13 chiplets but 4 (1 central + 3x120CU) that can be cut to smaller parts - that's a lot of chiplets - there will be penalty. Or I see it as possible that rdna5 will start unoptimized, below 30% on average, and, analogical to Intel A770, there could be 9x fps updates drivers that would use all CUs. My questions are: if the 13 chiplets gpu is a thing 1. does it mean 120cu is target for ps6, 2. will there be integrated but very expensive 120CU GPU in 2026, 3. will AMD not delay it launch further, 4. Could it mean the same chiplets for integrated and discrete gpus, 5. would tsmc be the only place to print it? You can share your guesses if you have a spare time.
Yeap these titles every generation.
Is rnda5 be pci e 5?
It most likely will be
The next RDNA 4 arch is not out yet, we know it will be far behind, and I’m supposed to hold my breath for the next-next one? I’m sure they are _really_ worried about this at Nvidia. Cut the BS.
like for the fitness channel
We seriously need to report videos like this for "untrue informations" or something. Maybe then all these tech-tubers will learn to not spread BS.
rdna 4 should launch in august 2024 , not along side rtx 5000
So the callout at the end is the real reason for the shirts lately! hehe
Let's hope AMD won't pull an Nvidia.
Brolic Paul FTW 💪