AMD didn't talk about its new GPUs in the keynote, but yes, they do exist. Clarification: Our statement regarding the new 9950X3D and 9900X3D being "not downclocked" is solely based on the information provided by AMD and we will have to wait till we have them in hand in order to verify that the clock speeds are the same in practice when comparing the X3D variant to it's non-X3D counterpart.
everyone said the 5070 would be 8gb also, since 5070 ti has 16gb, they drew that conclusion =X maybe they will listen this time? o.O plus doesnt it appear frame gen is costing a lot of vram? or did i read that wrong in their presentation
I'm sick of it. Are corporations dumbing down model numbers and directly ripping off one another because they feel we are too stupid to think for ourselves? It's frankly insulting.
@aikiwaki even as someone who keeps up with bleeding edge hardware, amd will never win as long as they keep trying to steal everyone else’s naming scheme instead of being consistent and doing their own thing. Too bad, as their number one fan, I kinda just am ready to watch x86 die now that it’s just amd left. They are in no position to lead even with leadership in products, they lack the actual leader part of that.
I feel bad to say this, but this is what AMD really need; direct comparison naming scheme for OEM and pre-built along with exlusive feature like FSR4. I do not want to say AMD is better or worse, do not get me wrong. They just deserve a better consideration, their products are good (not better, good) but they do not sell primary for marketing and brand ricognition over Intel (on laptop side they still have the look of a bad stuff) and nvidia. If this is not true for you, with a passion or at least interest on tech, it is very true for avarage consumer and gamers, where you really make the money as company. It is a mess? Yes, but mandatory at this point
As someone who works in tech and follows consumer tech I am so confused by AMD's naming schemes for their products. I can't even imagine how people outside of the tech space feel.
I also do follow consumer tech, but not AMD's latest gpu's release (from 7000 series) on how they're doing, so i only questioned why they changed the naming scheme and what purpose is behind all this, though i'll be back on track once it releases.
@@tohtorrenta fair assessment for once. Yes amd gpus are certainly ok but I do appreciate the price to performance of them especially with the 7900xt.
Amd top of the line single GPUs during the years: - HD 7970 - R9 290X - R9 FURY X - RX 480 - RX 590 - RX Vega 64 - Radeon VII - RX 5700XT - RX 6950XT - RX 7900XTX - RX 9070XT ??
at the 8:10 mark i realized that amd's naming scheme put me in a trance. like i watched the entire video and still have no idea what the hell is going on.
I don’t know about the constant clicking sound whenever a graphic was shown or removed. I tried to just listen to the video first but felt like I had some audio issues. Even when watching the video I find them a bit annoying… But that might be me
is the classic second player advantage, they can slash the intended price if nvidia is price competitive, and keep it or rise it if Nvidia goes bananas high with their prices.and yes they can announce one price now and slash it later but that makes them look bad, and also removes the chance to ever rise it.
did they cut the GPU announcement last minute? lot of prerecorded press articles/videos said they announced the 9070(XT), but no actual word in the announcement other than "more information on rdna4 later this quarter"
probably, AMD was 100% expecting the 50 series to be revealed by now, really feels like both AMD and Nvidia are waiting for the other to make the first move
10:30 No, how NVIDIA's customers upgrade is "Nope, can't afford that... Nope, can't afford that... Nope, can't afford that... Lemme check one generation back... Nope, still can't afford that... Nope, still can't afford that... Oh, this is within budget. Wait, this is just the same GPU I already have."
I have been looking to get a new laptop recently, and all of the name changes are so confusing. The shear number of times I had to look at that chart from the video you put up on AMD's naming scheme to double check what I was looking at, and second guessing if I had the right numbers where they should be was insane. And now it has changed again!
@@lexavaritia7596 Yea. As I am looking at laptops now I am seeing a few at a slight discount which I might go for. But there is also a bit of CES news atm, which makes me wonder if I will regret anything.
The name change was implemented just to confuse people, they'll accidentally buy it thinking its Nvidia or will buy it thinking it's significantly better bc 9 is bigger than 5
@@AlleonoriCat you really think people spending upwards of 1000$ on a gpu are dumb enough to confuse the two? Maybe if they are buying a prebuilt and not standalone I can see that happening, even then most of the ones who do buy prebuilt at least check with someone else who knows a thing or two since these gpus are not cheap.
I think its mostly because, nearly a decade ago, they stole the I3, I5, etc variant naming from intel. They were mocked as the copycat trying to keep up. An entire decade later, they are now ahead of intel. They likely believe the naming scheme is a big part of success for them, I think they're hoping the same result with NVIDIA.
AMD Makes it hard to be faithful to them despite their price-to-performance ratio being a huge sweet spot. Every single time I want to build a new PC or upgrade a part, I have to learn everything about their hardware all over again and basically do a full research project from scratch.
@@JETWTF Intel and NVIDIA may be better at first glance, but they're not much better. NVIDIA: RTX 3060 TI basically same as 4060 TI, the cutting of ram, 6GB and 12GB variants, the different variants performing differently and being cut down despite same GPU variant. Shifting up naming schemes to increase pricing despite performance not showing the increase in value offered for the generation. (IE, that's why we have 3050's and 3090's now). You used to be able to tell that a new generation would match the next generation up (IE, 980 is 1070, 1070 is 2060, etc), it was rough but pretty close, nowadays its no longer as straight forward. Massive pricing increases despite performance stagnation. A few years back this was true, today, its a mess. Intel: Also has been a mess, no hyperthreading on 9th gen, to re-adding it, from "we won't do chiplet designs", to then copying AMD and releasing e-cores, to bad performance uplifts, wattage, reliability, to now they're cutting e-core design (and likely regressing in multi-threaded performance), to the new 'core ultra x xxx' naming scheme.
That's an odd take. How often are you upgrading specifically a CPU or GPU for this to be a problem? And if basic, well-documented specs and figures are an issue, you firmly belong in the prebuild market. You're, by your own admission, only fooling yourself here.
@TatterNoho 1) Maybe dont assume someone belongs in a build or prebuild market if you dont know anything about my knowledge or preferences. 2) I build every 3-4 years which is exactly how often AMD has completely altered their naming and marketing scheme over the past 15 or so years, or more if you are using both CPU and GPU from them. 3) I also typically carry over the graphics card into a second build, then upgrade that after the initial build. So usually around 4 years on a card. I have a system that works for me from a financial and performance/future proofing standpoint (as much as that's possible). I'm not altering what I do just because some scrub random on the internet thinks I should buy a rebuild because a company's naming scheme is dog shit. Lmao
No it's not. Infinity is not a number, there's no concept of 'Infinity + 1'; countable infinity + 1 is still the same as countable infinity (see Hilbert's Hotel paradox).
It's been like this for years now. AMD always announces their cards after Nvidia. I assume it's to have a lower price point than Nvidias, and to show case which games run better than the new Nvidia card.
linus is the only dude i can think of who will subtly dawg on a company after they sponsor multiple tech upgrades, while simultaneously advertising them.
Most of this video feels like ragebait anyway. Shows a massive table with 20+ rows and 10+ columns for two seconds: "Did you understand that entire website worth of data in two seconds? US EITHER! Now enjoy this word from our sponsor!"
I feel like the RX line of _600 _700 _700XT _800XT _900XT and _900XTX was a perfectly fine set up for naming skews....Thanks AMD for making great products that much more confusing.....
They'll probably gonna recycle the name again after they missed on 5070/5060 performance target. It's kinda risky to naming your product based on your (unreleased) competitor GPU, it's hit or miss really.
@@titan_fx They're going to change again next gen because next gen they're going to release UDNA-based GPUs...this whole confusion is just to make them look better as soon as they will come out with a "win" in terms of performance. It's so annoying...
With all due respect, you... got it wrong 🤣 This gen it was actually 600, 600xt, 700xt, 800xt, 900gre, 900xt, 900xtx - and it is very stupid. Why is 'xt' the base version? What does '7700 xt' mean if there is no '7700'? And then when you want to add '-xt' you realise you already had it on base card, so you end up with 'xtx'. And GRE, because calling a card "golden rabbit edition" is apparently a better na than just... not doing that.
It worked, but a lot of people seemed to hate that they did a 7900XT/7900XTX instead of doing 7900/7900XT, people hated they later slotted in the 7900GRE, people got upset over the 7700XT/7800XT naming, and people online seemed to enjoy hating on the 7600/7600XT in a bunch of ways. I've also seen speculation that it might be something to do with disambiguation? Last gen we had a R5 7600 and an RX 7600 on the market at the same time, as well as an R7 7700, R7 7700X, and RX 7700XT.
@@jameslake7775 The only think I know is that now I hate the missed opportunity of releasing a new 9800XT. I would've loved a throwback at Radeon's golden era with a beast of a GPU...but no, nevermind. GG AMD, lol.
Love the little dots on the graphs showing which information you’re referencing. Not sure if you’ve used them before, but this is my first time seeing them, and they’re very helpful!
@@Mika-85 Incorrect. Both the Pro and non-Pro SKUs are exactly the same in both of the properties you listed. The top and bottom half seem to be the exact same for being a different SKU
@@intensifiedhipster Then what are they marketing? Xeon chips have a calculable difference between them when comparing different SKUs. If it's something not pictured in this specific graph, then cool. But then why make two separate sections for the graphic at all?
About the Z2 Extreme - it's important to remember that Zen 5 compact cores are NOT Intel E cores. They have the same architecture and IPC as regular Zen 5 CPU cores, they just don't clock as high and are slightly more compact for power and space savings. The Z2 Extreme is designed for handheld gaming PCs. Those are always extremely GPU-limited, so going with mostly Zen 5 C cores makes sense, and the "disappointment" mentioned in this video is misplaced.
They have less L3 cache per 8 core grouping. 32MB for every 8 Zen 5 CCD down to 8MB 8 Zen 5c CCDs. Windows scheduler has gotten better..ish? They still have to communicate through IF. Then we get weird core counts. Why 3/5 and not 4/4? Are we at a point where 4 full cores is wasteful after shouting from the rooftops that we want/need more multithreaded games and applications? Are 4 Zen 5c cores not enough for a smooth experience if the OS doesn't touch the "full cores" during desktop use for efficiency? I'm thankful we actually got all 16 CUs for the extreme model, but the core count split more than anything is making me question the whole part.
@@jongentile4933 E cores are quite literally a different architecture and IPC than P cores. In Arrow Lake, E cores are the Skymont architecture while P cores are Lion Cove. Zen 5 and Zen 5 C are the exact same architecture and IPC, just with differing clock speeds and cache.
Yeah, not even little bit of their specs looks like. It's just their announcement of the new (again) naming scheme. They probably still wondering what performance and price set by Nvidia for 5070/60.
I'm stuck debating if I want to build a new rig now, or wait for the new gen stuff and potentially pay maybe 25% more once tariffs for slightly better performance.
Admittedly, I don't mind the change in GPU naming scheme. It's gonna make knowing what to compare against WAY easier and less confusing, especially for the average person. "60 class", "70 class", and so on. On the CPU side, it's an absolute MESS though
@@tohtorrent I imagine industry knowledge about the rough generational improvements would give them a ballpark estimate by the time they need to name the card
It’s funny how marking a video as unlisted doesn’t really mean unlisted if someone adds it to a playlist-it’s like TH-cam’s way of redefining the word.
A video that's unlisted is still available for anyone to see. It just won't show on the channel or on the algorithm. That's also why anyone with the link to the video can watch it.
In mathematics, infinity is not a number in the traditional sense, but rather a concept that represents an unbounded quantity. When you add to infinity, the result is still considered infinity. This is because adding a finite number to infinity does not change the nature of infinity-it remains unbounded. In more advanced mathematics, such as set theory or different types of infinities, there are ways to discuss different "sizes" of infinity, but for basic arithmetic, infinity plus any finite number is still infinity.
@@MrProteinK But he says "Even Apple's M4 ... chip [...] is gonna get..." so what should fill in there is really "Max" as he's talking about a chip there.
I couldn't decide if that was an error or not. The graph says Pro - but the numbers would put it near (but not beyond Max.) So Linux might have been extrapolating. Still - near Max performance at 120 watts is not that great....
@@colincornaby M4 Max is almost 2x M4 Pro in performance for GPU centric/heavy tasks, so for everything shown in the slide, with the exception of Vray, M4 Max performance would be wayyyy ahead of the AMD chip in this graph, no?
@@pkennethv Depends. Their Cinebench numbers are slower than Pro (which is not a great look at 120w.) The Vray numbers would put them towards Max. (But again, at 120w. I think Apple's Ultra line is basically a 120w part?)
I was going to upgrade my 2020 laptop this year, so I watched AMD keynote today so that i know what new chips are launching and I am hella confused these many processors launched, these many naming schemes, this is actually crazy
I don't even care now since at least in my market getting a notebook with even previously announced AMD cpus is practically impossible, it's as if everyone is lagging a year behind
5:51 it's a nitpick, but I find it funny how they spelled it "50 TOPs", like a "TOP" is one unit? Maybe they meant it as just "trillions of operations", without the "per second" part, but that would be weird as well. Intel has "TOPS" properly, rare Intel W
@@randfee I think they meant that it “could” give that chip a run for its money in the last graph because it was such a large lead over the M4 Pro. Was confusing though.
The AMD GPU naming hopefully hearkens back to the Radeon HD days: Radeon HD XX30 - Radeon HD XX50 - Radeon HD XX70 - Radeon HD XX90 .. They did that naming scheme for many years.
@9:30 Intel tried software unlocked features in CPUs once, and only once. It is a monumentally stupid and non-functional idea outside of true Big Iron such as mainframes.
Yeah, that's what I thought...it's dumb and not competitive at all. And at the end of the day the word Pro in front of a SKU it's the least of the problems of AMDs, Intel and Nvidia naming schemes 😂
I got some of those ear pad replacements for a Corsair HS70 and love them. Actually a pretty cool product. I hate the pleather that makes my head/ears all sweaty so the gel and microfiber go a LONG way to fixing that. Plus, they look neato.
Reminds me that when I was a kid, I always thought it was "dumb" how some cars where named with "boring, confusing numbers". Kinda silly in hindsight - but then again, most everyone could immediately recall what a "Bronco" was when Ford brought back that badge (not that the new models really deserve that title.....).
@@SethizFTWThe early brief they gave to the content creators included it and Z2/Z2 extreme/Z2 go slides, but they probably cut it from the event last minute for some reason
Meh, 9950X3D once again only has the extra cache on one CCD. I was really looking forward to a homogenous design with 16 cores, even if it meant a small frequency hit. Oh well, guess I'll get the 9950X and save some money.
You aren't buying either. Games will tank on cross CCU usage regardless of VCACHE being on both. You can already buy multiple CCU X3D thread rippers if it's really important to you...which its not. Its still going to be the second or tied fastest gaming CPU ffs you are just repeating nonsense you don't understand.
This would not be worth it for gamers, but maybe it’s good for enterprise and data centers? Imagine all the data crunching that can be done on 204-208MB of cache! Also, it’s definitely great for being able to host denser remote gaming servers with multiple GPUs. EDIT: Added a correct range to account for 9900X3D
@@samk2407 Linus: Look, Yvonne, a perplexus. The kids will love this! Little Man: Does this take batteries? Thing 2: I'm going to get the batteries! Thing 3: I want to play Minecraft.
I think you guys should have addressed the naming first as I found the first part in which you talked about the specs to be hard to follow along without knowing what any of the names actually mean. Other than that, great video though!
4:20 It's for console refreshes and steam deck. They're trying to make a chip that can go from 3W to 120W. It will require some radical new technology though.
I can’t wait for the podcast segment about this whole situation, I expect it to be 30+ minutes with a big rant from Linus and luke throwing a few jabs in. Definitely watching out for that podcast episode 👌🏻
No, intel arrow lake are the just released new ones, the 200HX series are for 'maximum performance' according to Intel. The 200V, 200U and possibly 200H are the more battery-life/efficiency oriented chips.
Having a proper naming scheme would bring them much more market share. Sick of tons of numbers and random looking letters which confuse customers. If I am getting a Mac for instance it is so easy, M3, M4, M4 Pro and so on... If it was AMD, they would probably name them something like Ryzen 36 M3089 X3D AI Pro something something... I know that Apple doesn't sell CPUs separately but you get the point
So great, AMD has some great chips for laptop. Great, they did last gen and I only saw Intel CPU based laptops in reviews. So for one, will they be available in enough quanties to where more makers will use them? And then, will they be priced low enough to where a person who's been buying Intel laptops for over a decade will switch? No matter how good the products have been the makers don't want to use much of them because they have to sell them for less.
AMD didn't talk about its new GPUs in the keynote, but yes, they do exist.
Clarification: Our statement regarding the new 9950X3D and 9900X3D being "not downclocked" is solely based on the information provided by AMD and we will have to wait till we have them in hand in order to verify that the clock speeds are the same in practice when comparing the X3D variant to it's non-X3D counterpart.
hey linus
thanks
sigma sigma boy
Ok
I'll read right after the segue, from our sponsor
Intel: Changes socket every generation.
AMD: Changes naming every generation.
nVidia: Changes prices every generation.
Nvidia adds a zero to the price with each generation
Legendary comment!
@@IDonotHaveAGoodName facts
Nvidia: disappoints on vram every generation
@@thegrumpydeveloper
RTX 7070 10GB $1100
At this point, I believe this year's CES presentations are AI generated.
*Neural rendered™
they used their ryzen ai chips for it
at the end: " Et voilà, everything was AI generated "
Gotta cut as much costs as possible
Generated by AI for AI, and we'll just get the summaries
AMD Marketing Department have an original thought challenge
AMD's AMD???
@@NoCluYTthe A in AMD stands for AMD.
AMD goal: At least we can mock Nvidia when they release the 5060 8gb when we have the 9060 12gb.
RTX 9070 XT
everyone said the 5070 would be 8gb also, since 5070 ti has 16gb, they drew that conclusion =X maybe they will listen this time? o.O plus doesnt it appear frame gen is costing a lot of vram? or did i read that wrong in their presentation
“AI Max+ Pro“ What a crazy name
It's pretty good but I will wait for the AI max+ pro extreme edition to come out before I buy anything
forgetting an ultra somehwere win there with more X
Elite Edition
Hmm, that cpu sounds a little expensive for me. Is there an AI semi-Max + pro? Or at least a AI Max - pro?
I'm sick of it. Are corporations dumbing down model numbers and directly ripping off one another because they feel we are too stupid to think for ourselves? It's frankly insulting.
Customers: Finally starting to understand AMD's GPU naming schemes
Lisa Su: Let me introduce you to the RX 9070
Customers: 😮
their old naming scheme was better ngl
There is no gpu 😂😂😂
Lol they didn't announced gpu
How about the awesome new "AI" CPU garbage naming scheme
i would prefer 9700 XT
@@GameRTmaster they probably heard the backlash and decided to take a second look at their naming schemes XD
it feels unreal to see the name RX 9070, I'm still used to the RX 580, good times
Wow that's like going from my Samsung S8 to S24, that's a 16 gen jump!
Wow, it's 85 generations newer!
@aikiwaki even as someone who keeps up with bleeding edge hardware, amd will never win as long as they keep trying to steal everyone else’s naming scheme instead of being consistent and doing their own thing. Too bad, as their number one fan, I kinda just am ready to watch x86 die now that it’s just amd left. They are in no position to lead even with leadership in products, they lack the actual leader part of that.
@@ZaeTalkz what do you beam x86 is dead? What about Intel?
i changed my War Horse RX580 for a RTX 4070 Super 16gb and is a new level of power!
Probably makes it confusing on purpose. 9070 is a bigger number than the 5070. It's like if Pepsi decided to name color their cans red.
oh god, the next gen is going to be the 10070 isn't it.
I feel bad to say this, but this is what AMD really need; direct comparison naming scheme for OEM and pre-built along with exlusive feature like FSR4.
I do not want to say AMD is better or worse, do not get me wrong. They just deserve a better consideration, their products are good (not better, good) but they do not sell primary for marketing and brand ricognition over Intel (on laptop side they still have the look of a bad stuff) and nvidia.
If this is not true for you, with a passion or at least interest on tech, it is very true for avarage consumer and gamers, where you really make the money as company.
It is a mess? Yes, but mandatory at this point
As someone who works in tech and follows consumer tech I am so confused by AMD's naming schemes for their products. I can't even imagine how people outside of the tech space feel.
I also do follow consumer tech, but not AMD's latest gpu's release (from 7000 series) on how they're doing, so i only questioned why they changed the naming scheme and what purpose is behind all this, though i'll be back on track once it releases.
@@gateopssss gpus are ok. but laptop cpu names are insane.
Names are irrelevant just look at the price and if it fits the mobo. Then check it out on benchmark scores.
I just switched to amd for my next build and i had to learn all the confusing ass names😂 im cooked in the future
@@tohtorrenta fair assessment for once. Yes amd gpus are certainly ok but I do appreciate the price to performance of them especially with the 7900xt.
Amd top of the line single GPUs during the years:
- HD 7970
- R9 290X
- R9 FURY X
- RX 480
- RX 590
- RX Vega 64
- Radeon VII
- RX 5700XT
- RX 6950XT
- RX 7900XTX
- RX 9070XT ??
They skipped the HD8000 series too, but there was a 9800xt (ATi did anyway)
You forgot Radeon VII
@@tomstech4390 That's ATI, OP talking about AMD
The trash cards 😂
@@tomstech4390 Also the list title literally says "top of the line"
at the 8:10 mark i realized that amd's naming scheme put me in a trance. like i watched the entire video and still have no idea what the hell is going on.
AMD is garbage at naming products ...
Same, bro... ='^(
dude its genuinely incomprehensible what is amd doing
There are like 4 different amd laptop processor line up launching at the same time?
Yessssssss-LS
@@LinusTechTipsNVIDIA IS BETTER
We WILL BUY the 5090!
Only question I have is at what cost?
Well not exactly the same time. They are spread over 2 quarters in 2025 as far as I understood.
yup L nobody cares
I don’t know about the constant clicking sound whenever a graphic was shown or removed. I tried to just listen to the video first but felt like I had some audio issues. Even when watching the video I find them a bit annoying… But that might be me
Yes, I thought I was getting some strange notifications.
In physics, some infitnities are larger than others, some some maxes can be larger too
Did AMD forget to announce the GPUs onstage?
@@taziir443 they are waiting for NVIDIA to announce their new graphics card pricing, so that AMD can price lower
is the classic second player advantage, they can slash the intended price if nvidia is price competitive, and keep it or rise it if Nvidia goes bananas high with their prices.and yes they can announce one price now and slash it later but that makes them look bad, and also removes the chance to ever rise it.
@@lucaskp16 so they are doing Nvida -1 again huh? Disappointing
@@GuntasSingh To me, it's proof that they base their prices on competition rather than on actual cost of their product...
@@UltimateByte They do because Nvidia price with MASSIVE margins helping AMD to gain market.
did they cut the GPU announcement last minute? lot of prerecorded press articles/videos said they announced the 9070(XT), but no actual word in the announcement other than "more information on rdna4 later this quarter"
probably, AMD was 100% expecting the 50 series to be revealed by now, really feels like both AMD and Nvidia are waiting for the other to make the first move
They had to wait until Nvidia announce their pricing so they can be $50 cheaper with less features.
probably to wait for Nvidia pricing
bet they cut it because they don't have the prices decided
imagine nvidia cuts their rtx 5000 announcement now 😭😭🤣🤣
User benchmarks is gonna be so pissed
They'll just gaslight us harder.
But hey at least this time they will be around 3% right to do so ...
@@Onomere yeah lol
Bandwagon. Feel apart.
10:30 No, how NVIDIA's customers upgrade is
"Nope, can't afford that...
Nope, can't afford that...
Nope, can't afford that...
Lemme check one generation back...
Nope, still can't afford that...
Nope, still can't afford that...
Oh, this is within budget. Wait, this is just the same GPU I already have."
I have been looking to get a new laptop recently, and all of the name changes are so confusing. The shear number of times I had to look at that chart from the video you put up on AMD's naming scheme to double check what I was looking at, and second guessing if I had the right numbers where they should be was insane. And now it has changed again!
bought a last gen laptop this october 2024
i thought im gonna be so regretful. good thing they went on discount
@@lexavaritia7596 Yea. As I am looking at laptops now I am seeing a few at a slight discount which I might go for. But there is also a bit of CES news atm, which makes me wonder if I will regret anything.
The name change was implemented just to confuse people, they'll accidentally buy it thinking its Nvidia or will buy it thinking it's significantly better bc 9 is bigger than 5
Of course, duh. It even has all the letters of RTX with extra X! Must be hella better!
Not the first time AMD tries to pull scummy tricks like that, just like their motherboard chipsets are now copy-catting Intel's 🤡
@@norman6524 hold up, that's genius actually
@@AlleonoriCat you really think people spending upwards of 1000$ on a gpu are dumb enough to confuse the two? Maybe if they are buying a prebuilt and not standalone I can see that happening, even then most of the ones who do buy prebuilt at least check with someone else who knows a thing or two since these gpus are not cheap.
I think its mostly because, nearly a decade ago, they stole the I3, I5, etc variant naming from intel.
They were mocked as the copycat trying to keep up.
An entire decade later, they are now ahead of intel.
They likely believe the naming scheme is a big part of success for them, I think they're hoping the same result with NVIDIA.
I like how we got the gpu announcement from ltt rather than amd themselves
RDN4 ARE THE FRIENDS WE MADE ON THE ROAD.
@@ElibeniEB7 God bless you 🙏😭
God bless President Trump 🙏🌅
AMD Makes it hard to be faithful to them despite their price-to-performance ratio being a huge sweet spot.
Every single time I want to build a new PC or upgrade a part, I have to learn everything about their hardware all over again and basically do a full research project from scratch.
Pretty much why I stick with Intel and Nvidia, only have to learn it every 10 or so generations and it's close to the last so easy to learn.
@@JETWTF Intel and NVIDIA may be better at first glance, but they're not much better.
NVIDIA: RTX 3060 TI basically same as 4060 TI, the cutting of ram, 6GB and 12GB variants, the different variants performing differently and being cut down despite same GPU variant.
Shifting up naming schemes to increase pricing despite performance not showing the increase in value offered for the generation. (IE, that's why we have 3050's and 3090's now).
You used to be able to tell that a new generation would match the next generation up (IE, 980 is 1070, 1070 is 2060, etc), it was rough but pretty close, nowadays its no longer as straight forward.
Massive pricing increases despite performance stagnation.
A few years back this was true, today, its a mess.
Intel: Also has been a mess, no hyperthreading on 9th gen, to re-adding it, from "we won't do chiplet designs", to then copying AMD and releasing e-cores, to bad performance uplifts, wattage, reliability, to now they're cutting e-core design (and likely regressing in multi-threaded performance), to the new 'core ultra x xxx' naming scheme.
That's an odd take. How often are you upgrading specifically a CPU or GPU for this to be a problem? And if basic, well-documented specs and figures are an issue, you firmly belong in the prebuild market. You're, by your own admission, only fooling yourself here.
I would not trust any vendor with its next gen products being just better, so I would spend my time on reviews anyhow.
@TatterNoho 1) Maybe dont assume someone belongs in a build or prebuild market if you dont know anything about my knowledge or preferences.
2) I build every 3-4 years which is exactly how often AMD has completely altered their naming and marketing scheme over the past 15 or so years, or more if you are using both CPU and GPU from them.
3) I also typically carry over the graphics card into a second build, then upgrade that after the initial build. So usually around 4 years on a card.
I have a system that works for me from a financial and performance/future proofing standpoint (as much as that's possible). I'm not altering what I do just because some scrub random on the internet thinks I should buy a rebuild because a company's naming scheme is dog shit. Lmao
Not to be nerdy here, but saying "take a math lesson" at 9:15 feels a little wrong. Infinity+1 is a legitimate thing in higher math 😄
No it's not. Infinity is not a number, there's no concept of 'Infinity + 1'; countable infinity + 1 is still the same as countable infinity (see Hilbert's Hotel paradox).
So Linus announces new Radeon lineup instead of AMD themselves?
The circus tech tent brought to you by AMD
Instills great confidence huh
not only him. hardware unboxed also. the videos were ready i guess. amd pulled out their slides.
I was so disappointed to not see anything gpu related
same
Hopefully they're working on something big to compete with Nvidia but probably not
Cuz their new 9000 series sucks
It's been like this for years now. AMD always announces their cards after Nvidia. I assume it's to have a lower price point than Nvidias, and to show case which games run better than the new Nvidia card.
linus is the only dude i can think of who will subtly dawg on a company after they sponsor multiple tech upgrades, while simultaneously advertising them.
Most of this video feels like ragebait anyway.
Shows a massive table with 20+ rows and 10+ columns for two seconds:
"Did you understand that entire website worth of data in two seconds? US EITHER! Now enjoy this word from our sponsor!"
@@cougrrr I’ve Never Been This ANGRY and CONFUSED at a video title.
@@fenxis I couldn't understand how angry and confused you were without an outrageous algorithm defining thumbnail tho :(
@@cougrrr You wanted him to leave the infographic on screen for 35 seconds while you read it? Bro has absolutely no idea how the industry works lol.
@@cougrrr You can pause. That is a feature of video streaming and has been since the dawn of the internet.
Nvidia: 4080, 4080 super, 4080 ti, 4060 8gb, 4060 16gb.. or whatever. I miss the days where you had a 760,770,780,780ti done.
well, in this naming you just need to understand what is cooler (super or ti)
@@vadym8713 Weren't there some super ti cards? For example RTX 4070 Ti super?
i ordered two different pairs of wickedcushions replacements for two of my headsets because of you, im incredibly happy with them :)
The AMD live stream was amazing but I got sick of them saying “Ai”
Next time, do the game where you take a shot when they say AI
@@moortu ur gonna be taking hundreds of shots lmao
@ 32 shots!
I lost count at 30
with the presentation being 90% AI, I can't recall a single mention of Linux and ROCm. maybe it happened, but I still have AI ringing in my head.
The naming used to be so simple and understandable... [tier]-[generation]-[sub-tier]-[some additional junk]
I feel like the RX line of _600 _700 _700XT _800XT _900XT and _900XTX was a perfectly fine set up for naming skews....Thanks AMD for making great products that much more confusing.....
They'll probably gonna recycle the name again after they missed on 5070/5060 performance target.
It's kinda risky to naming your product based on your (unreleased) competitor GPU, it's hit or miss really.
@@titan_fx They're going to change again next gen because next gen they're going to release UDNA-based GPUs...this whole confusion is just to make them look better as soon as they will come out with a "win" in terms of performance. It's so annoying...
With all due respect, you... got it wrong 🤣
This gen it was actually 600, 600xt, 700xt, 800xt, 900gre, 900xt, 900xtx - and it is very stupid. Why is 'xt' the base version? What does '7700 xt' mean if there is no '7700'? And then when you want to add '-xt' you realise you already had it on base card, so you end up with 'xtx'. And GRE, because calling a card "golden rabbit edition" is apparently a better na than just... not doing that.
It worked, but a lot of people seemed to hate that they did a 7900XT/7900XTX instead of doing 7900/7900XT, people hated they later slotted in the 7900GRE, people got upset over the 7700XT/7800XT naming, and people online seemed to enjoy hating on the 7600/7600XT in a bunch of ways.
I've also seen speculation that it might be something to do with disambiguation? Last gen we had a R5 7600 and an RX 7600 on the market at the same time, as well as an R7 7700, R7 7700X, and RX 7700XT.
@@jameslake7775 The only think I know is that now I hate the missed opportunity of releasing a new 9800XT. I would've loved a throwback at Radeon's golden era with a beast of a GPU...but no, nevermind. GG AMD, lol.
Wow! You had more useful information about those CPUs than GN. Great content and great progress!
that one thread sticking out of Linus' shoulder drives me crazy !!!
Love the little dots on the graphs showing which information you’re referencing. Not sure if you’ve used them before, but this is my first time seeing them, and they’re very helpful!
For the love of all that is holy, what it the name of bejeezus is that naming scheme? The world has gone mad...
I feel they looked at Apple’s naming scheme and went hmmmm 🤔
what makes this worse is that it was greenlit by multiple higher ups at AMD 😂
I feel like the list is pretty simple when you realize all the obscure ones are laptop chips.
We're THIS close to naming as moronic as TV naming schemes 😂
6:11 is the chart wrong or am i stupid there seems to be zero difference
The second. Look at core Count and frequency
@@Mika-85 Incorrect. Both the Pro and non-Pro SKUs are exactly the same in both of the properties you listed. The top and bottom half seem to be the exact same for being a different SKU
@@MomV2 q1 is AI chipset, q2 is AI PRO chipset. (With zero difference in performance)
Zero difference, but Pro chips should have additional capabilities for enterprise users, kind of like Xeons.
@@intensifiedhipster Then what are they marketing? Xeon chips have a calculable difference between them when comparing different SKUs. If it's something not pictured in this specific graph, then cool. But then why make two separate sections for the graphic at all?
It's ALMOST as if assigning confusing names promotes people to buy the most expensive one "to be sure".
Usually if people don't understand what they should buy, they buy something else from a competitor with a clearer offer.
why were the gpus not in the AMD Presentation at CES
to wait for Nvidia to drop, and then slash the prices down and then make a second part to the CES presentation
@@justanonexistentuser Not a bad conspiracy. Gotta keep the only thing they have going for themselves
About the Z2 Extreme - it's important to remember that Zen 5 compact cores are NOT Intel E cores. They have the same architecture and IPC as regular Zen 5 CPU cores, they just don't clock as high and are slightly more compact for power and space savings.
The Z2 Extreme is designed for handheld gaming PCs. Those are always extremely GPU-limited, so going with mostly Zen 5 C cores makes sense, and the "disappointment" mentioned in this video is misplaced.
Isnt the z2e just a ryzen ai 9 hx 370 without the ai stuff? Just like how the z1e is a 7840u w no ai stuff?
They have less L3 cache per 8 core grouping. 32MB for every 8 Zen 5 CCD down to 8MB 8 Zen 5c CCDs. Windows scheduler has gotten better..ish? They still have to communicate through IF. Then we get weird core counts. Why 3/5 and not 4/4? Are we at a point where 4 full cores is wasteful after shouting from the rooftops that we want/need more multithreaded games and applications? Are 4 Zen 5c cores not enough for a smooth experience if the OS doesn't touch the "full cores" during desktop use for efficiency? I'm thankful we actually got all 16 CUs for the extreme model, but the core count split more than anything is making me question the whole part.
Can you provide the link to where exactly you read up on E cores being different architecture to P cores?
@@jongentile4933 E cores are quite literally a different architecture and IPC than P cores. In Arrow Lake, E cores are the Skymont architecture while P cores are Lion Cove.
Zen 5 and Zen 5 C are the exact same architecture and IPC, just with differing clock speeds and cache.
@@jongentile4933 Wikipedia is your friend. The writes there about the architectures also include link to sources as per usual.
They didn't even reveal the gpu's at the keynote
Bruh facts, what a joke. Wasted a whole hour watching that
9:15 Not to disagree with Linus's point here... But you can actually have different sizes of infinity, search for the infinite hotel paradox. 😅
Perfect sponsor for LTT! Keep your tech longer
Ryzen 4070 is not a myth anymore.
Ryzen AI 4070 xtx GRE Max+ Pro
Christ I miss when Mhz used to tell you everything you needed to know about a chip.
486DX4-75Mhz :D
Like Bulldozer?
They did not in fact even tease any of the RDNA 4 desktop GPUs.
Yeah, not even little bit of their specs looks like. It's just their announcement of the new (again) naming scheme.
They probably still wondering what performance and price set by Nvidia for 5070/60.
I'm stuck debating if I want to build a new rig now, or wait for the new gen stuff and potentially pay maybe 25% more once tariffs for slightly better performance.
Admittedly, I don't mind the change in GPU naming scheme. It's gonna make knowing what to compare against WAY easier and less confusing, especially for the average person. "60 class", "70 class", and so on. On the CPU side, it's an absolute MESS though
how would amd know that 9070 is in the same league as 5070?
@@tohtorrent I imagine industry knowledge about the rough generational improvements would give them a ballpark estimate by the time they need to name the card
3:44 you know what else is?
high taper fade
@@thephantomrealmi disliked your comment 💔
LOWWWWW TAPER FADE
The fact that the AMD Naming scheme explained video was only 6 MONTHS AGO really shows just how bad it's getting
Me with my Quad Core 2nd gen Core i5:
“Eh, I’ll wait another year”
@panagiotispappas1001 lol same, I’m still hanging on to my second gen i7
@@WindyKnight0822 did you guy overclock it?
I’ve been using wicked cushions for years. It’s pretty neat to see a sponsored spot for them. Great value replacements.
At least this naming situation gives you a quick easy video to make to talk about it in more detail!
It’s funny how marking a video as unlisted doesn’t really mean unlisted if someone adds it to a playlist-it’s like TH-cam’s way of redefining the word.
TH-cam doesn't know what their own words mean anymore...
A video that's unlisted is still available for anyone to see. It just won't show on the channel or on the algorithm. That's also why anyone with the link to the video can watch it.
to be fair.... infinity plus one is a thing.....
No matter how much I wish it was not......
While true, its a poor defence as it's still infinitely long. The smart play is to go for multiple infinites instead of trying to have a larger one.
Just scrolled down to the comments to look for this. This should be further up.
In mathematics, infinity is not a number in the traditional sense, but rather a concept that represents an unbounded quantity. When you add to infinity, the result is still considered infinity. This is because adding a finite number to infinity does not change the nature of infinity-it remains unbounded.
In more advanced mathematics, such as set theory or different types of infinities, there are ways to discuss different "sizes" of infinity, but for basic arithmetic, infinity plus any finite number is still infinity.
You can tell Linus is sincere at 0:59 because of the neck veins.
Aye, just bought myself a set of wicked cushions for my Astro headset. They feel great!
I hope that laptops finally become useful, equipped with a great long-lasting battery life.
Can't tell if early or somewhere I'm not supposed to be. 🤔
is this a mistake at 5:17? Linus says "M4 Max" but the label on the bottom of the slide says "M4 Pro" (twice)
Maybe he meant m4 macs lol 😂
@@MrProteinK But he says "Even Apple's M4 ... chip [...] is gonna get..." so what should fill in there is really "Max" as he's talking about a chip there.
I couldn't decide if that was an error or not. The graph says Pro - but the numbers would put it near (but not beyond Max.) So Linux might have been extrapolating. Still - near Max performance at 120 watts is not that great....
@@colincornaby M4 Max is almost 2x M4 Pro in performance for GPU centric/heavy tasks, so for everything shown in the slide, with the exception of Vray, M4 Max performance would be wayyyy ahead of the AMD chip in this graph, no?
@@pkennethv Depends. Their Cinebench numbers are slower than Pro (which is not a great look at 120w.) The Vray numbers would put them towards Max. (But again, at 120w. I think Apple's Ultra line is basically a 120w part?)
I was going to upgrade my 2020 laptop this year, so I watched AMD keynote today so that i know what new chips are launching and I am hella confused these many processors launched, these many naming schemes, this is actually crazy
I don't even care now since at least in my market getting a notebook with even previously announced AMD cpus is practically impossible, it's as if everyone is lagging a year behind
to the editor: i love the little click-y sound that plays when a graph/chart/whatever pops up. nice subtle way to throw my brain into reading mode :)
Opposite for me, it's driving overload and couldn't finish the video
8% uplift from the 7950x3D.. Yeah I think I'll hold on to that.
Ryzen AI MAX Radeon RX ULTIMATE GAMING XT DOUBLE AI CRUSHER OF INTEL AI Chip
5:51 it's a nitpick, but I find it funny how they spelled it "50 TOPs", like a "TOP" is one unit? Maybe they meant it as just "trillions of operations", without the "per second" part, but that would be weird as well. Intel has "TOPS" properly, rare Intel W
t0pS
mistake: at 5:25 you say that it's comparable to M4 Max, yet the graph shows M4 Pro.
@@randfee I think they meant that it “could” give that chip a run for its money in the last graph because it was such a large lead over the M4 Pro. Was confusing though.
@@randfee you’re doing gods work
@@mjb180Nah it’s not even beating the full M4 Pro let alone the M4 Max
again designed to be confusing on purpose 😖
The AMD GPU naming hopefully hearkens back to the Radeon HD days: Radeon HD XX30 - Radeon HD XX50 - Radeon HD XX70 - Radeon HD XX90 .. They did that naming scheme for many years.
My favourite part about this channel is when linus does his parrot impressions 6:28
Why even schedule an early announcement if you were going to wait for NVIDIA anyway?
can we get a shoutout to the sound design in this video? some of the cleanest y'all have put out yet. feels like im watching a game of balatro
@9:30 Intel tried software unlocked features in CPUs once, and only once. It is a monumentally stupid and non-functional idea outside of true Big Iron such as mainframes.
Yeah, that's what I thought...it's dumb and not competitive at all. And at the end of the day the word Pro in front of a SKU it's the least of the problems of AMDs, Intel and Nvidia naming schemes 😂
0:55 that’s exactly why I’m here
I got some of those ear pad replacements for a Corsair HS70 and love them. Actually a pretty cool product. I hate the pleather that makes my head/ears all sweaty so the gel and microfiber go a LONG way to fixing that. Plus, they look neato.
Why didn't AMD cover some of the stuff in CES announcement live stream??
We're is the AMD ryzen 4070 😅😅
Flashbang at 4:58
your not wrong
My eyes!
"I'm getting hot here, boys", is a great little sound bite... 😀
Reminds me that when I was a kid, I always thought it was "dumb" how some cars where named with "boring, confusing numbers". Kinda silly in hindsight - but then again, most everyone could immediately recall what a "Bronco" was when Ford brought back that badge (not that the new models really deserve that title.....).
where is the tease in the keynote??
I have no clue, I rewatched the keynote and I did not even hear the GPU's
@@SethizFTWThe early brief they gave to the content creators included it and Z2/Z2 extreme/Z2 go slides, but they probably cut it from the event last minute for some reason
Meh, 9950X3D once again only has the extra cache on one CCD. I was really looking forward to a homogenous design with 16 cores, even if it meant a small frequency hit. Oh well, guess I'll get the 9950X and save some money.
You aren't buying either. Games will tank on cross CCU usage regardless of VCACHE being on both. You can already buy multiple CCU X3D thread rippers if it's really important to you...which its not. Its still going to be the second or tied fastest gaming CPU ffs you are just repeating nonsense you don't understand.
@@backgammonbacon ?
This would not be worth it for gamers, but maybe it’s good for enterprise and data centers?
Imagine all the data crunching that can be done on 204-208MB of cache!
Also, it’s definitely great for being able to host denser remote gaming servers with multiple GPUs.
EDIT: Added a correct range to account for 9900X3D
@@fujinshu the 9950X3D has 144 and 9900X3D has 140MB of cache
@sorrowdoa Oh, right. Thanks for the correction!
1:05 are those perplexus balls in the backround lmao
Probably, this is shot in his basement where the Linus toys are.
Lmaooo they fucking are what a snipe that's such a throwback I had those as a kid
@@samk2407 Linus: Look, Yvonne, a perplexus. The kids will love this!
Little Man: Does this take batteries?
Thing 2: I'm going to get the batteries!
Thing 3: I want to play Minecraft.
I think you guys should have addressed the naming first as I found the first part in which you talked about the specs to be hard to follow along without knowing what any of the names actually mean. Other than that, great video though!
Have you folks done any ads with Wicked Pads before? Is this the first time I'm seeing their ad in here?
Definitely not the first lmg sponsorship
4:20 It's for console refreshes and steam deck. They're trying to make a chip that can go from 3W to 120W. It will require some radical new technology though.
Seems like it could also be useful for something like the Mac Mini
@@pin65371 Yeah don't see that happening.. Why would apple move away from their own architecture, when they've just innovated on something -.-
@@clawwer4404 I didnt mean the actual mac mini. I just meant a micro computer like the Mac Mini.
@@pin65371 oh XD haha... There are already some really great mini computers out there XD Mb I guess.
uhh why do we have the PDF about the GPUs but NO ANNOUNCEMENT ON STAGE?
BTW they are probably going to change naming again unless they will settle with radeon 10070 or 1070, which I doubt
I can’t wait for the podcast segment about this whole situation, I expect it to be 30+ minutes with a big rant from Linus and luke throwing a few jabs in. Definitely watching out for that podcast episode 👌🏻
That slide is hilarious, did they seriously just compare graphics performance between their workstation SoC to Intel’s thin and light?
No, intel arrow lake are the just released new ones, the 200HX series are for 'maximum performance' according to Intel. The 200V, 200U and possibly 200H are the more battery-life/efficiency oriented chips.
@ right, but they’re not SoC’s, they’re meant to be packaged with a dGPU. It’s a dumb comparison
How did this come out so fast the show just ended like 5 minutes ago lol
companies give media outlets like LTT the information before they announce it publically
They had press release materials under NDA obviously.
I thought of that as soon as I posted the comment lol forgive my slow brain I am super sick today... =(
45 minutes waste of time... no desktop GPU announcement.
4:55 YOU KNOW WHAT ELSE IS MASSIVE????
🗣️🗣️🗣️🔥🔥🔥💯💯💯
Time for the yearly "linus rants about hardware naming schemes"
Having a proper naming scheme would bring them much more market share. Sick of tons of numbers and random looking letters which confuse customers. If I am getting a Mac for instance it is so easy, M3, M4, M4 Pro and so on... If it was AMD, they would probably name them something like Ryzen 36 M3089 X3D AI Pro something something... I know that Apple doesn't sell CPUs separately but you get the point
1:57 yea but most games aren’t as. CPU driven as CS2.
At 1.5x, Sponsor sounds like Wicked Christians
This segue was sponsored by Nero and the Roman Empire
I knew I was not the only one who heard that :D
I found you baby, quite a scroll though!
"We are expecting new GPU's"
AMD:
nope
Having a pair of wicked cushions, I can admit this sponsorship is warranted, extremely good comfort and quality
So great, AMD has some great chips for laptop. Great, they did last gen and I only saw Intel CPU based laptops in reviews.
So for one, will they be available in enough quanties to where more makers will use them? And then, will they be priced low enough to where a person who's been buying Intel laptops for over a decade will switch? No matter how good the products have been the makers don't want to use much of them because they have to sell them for less.