The perfect naming scheme: Take the complete datasheet of your hardware, pull it through a hash function and you have a perfectly unreadable name wich totally makes sense.
and the best thing: you cant even reconstruct the data from it, you'll just have to guess until you get it right. And with all those different parameters you'll actually need a password cracker for that. great success.
@@defnotatroll it is genius. Its also anti consumer. Someone probably figured out less tech savvy people are more likely to buy laptops. Hence the horrible naming for marketing's sake. Make someone feel lime they're getting a really good ryzen 7 for a cheap price. Only for it to be super outdated. They won't know that or even notice it. Its disgusting. Intentional confusion for more sales
I think the issue they are having os that their older technology is still viable for certain use cases, but they are harder to sell with the old name. It's also not directly a scam. E.g. if you only need an office pc there is absolutely nothing wrong with the Gen 3 architecture for example.
About AMD and their "Portfolio Model Year", Not sure if this is elaborate, but each year is divisible by its respective "series" number: 2023/7=289 2024/8=253 2025/9=225 But wait, there's more! The digits of every year also add up to its "series" number: 2+0+2+3=7 2+0+2+4=8 2+0+2+5=9 If they were to follow this pattern, the 10 series would have to come out in 2080 at the earliest.
Samsung already figured this out. They skipped the s11 to go to s20 representing the year it came out in 2020, and then the s21 for 2021 and so on. This naming scheme is indefinite so long as years continue to pass.
I used to work in the Radeon department at ATI (pre-AMD) and the "marketing name" as we called it, was totally disconnected from the name we used. We were normally working on products that were nearly 2 years away from release, so there were many times where we'd have meetings about something that launched and we'd have Abbott and Costello level conversations where nobody knew what old (to us) product we were talking about, as we tried to convert marketing names back into engineering codenames from 2 years ago.
Would it be possible for you to share how those internal naming schemes usually looked like? I'm interested how much info those engineering names carried (and how useful/unwieldy were they). I realize it might be under NDA, but it's worth a shot asking (in case they were ever leaked, published or if the general naming scheme doesn't reveal much).
We have similar shit going on at Intel now. Somebody will mention that we did something interesting in "Devil's Canyon" and about half the room will have no idea what that means. We're trying to get the memo out to everybody to just refer to them by the generations.
they should name gaming equipment the same way they name American cars. i would certainly buy the new nvidia GPU if they named it "The Hellthrottler" or something similar
Well it depends. If you look at ferrari for example, their naming schemes are equally confusing, even though they have less products. But maybe that's the reason, if you want to confuse your customors, and you don't have lot's of products your confusing naming schemes have to be more confusing to confuse to the same amount. But they also are italian. So who knows. Let's look at their forumla 1 cars for example and start in 2010: 2010: Ferrari F10. 10 for 2010. F for ferrari. Thus technically the Ferrari Ferrari 10. 2011: Ferrari 150º Italia. Because the state of Italy had it's 150th birthday. 2012: Ferarri F2012. Straight forward. F to indicate it's a ferrari. 2012 because 2012. You can see, they don't want to confuse us and added a 20 because of this, to not confuse us with the ferraris from 1920, even though they didn't exist back then. Or they wanted to make their naming scheme future proof. Something they could use for over 100 years. And? Did they? 2013: Ferrari F138. No, they didn't. They changed it the next year. F stand's again for for ferrari, 13 for 2013 and 8 because it's an 8 cylinder. 2014: Ferrari F14T. Yes, you guessed the Ferrari part right, also the F. The 14 stand's for 2014 (the 20 in the name is gone again, maybe it was to confusing). The T for the turbocharger it now got. 2015: SF15-T. Technically the Racing Team is called Scuderia Ferrari and a daughter of the mothercompany ferrari. So the SF stands for scuderia ferrari, which mean Ferrari Racing team. The 15 stands for 2015, the T for turbocharger, but for some reason they added a hypen infront of it for this year. 2016: SF16-H. SF16, well you get it. But now it's a H instead of a T at the end. This stands for hybrid. Because it is got a hybrid powertrain with a turbocharched engine, like the last two years cars. I guess you need also somethin to differntiate to following models from each other? Although, they already got the year in the name. The letter in the end could also be, that noone confuses it with the nonhybdrid Ferrari SF16. Although none such cars exists. 2017: SF70-H. Still a hybdrid, but ferrari was founded in 1957. So it's a car that came out 70 years after they started. So they put a 70 in the name to celebrate the round birthday. 2018: SF71H. I guess they also celebrate unround birthdays. Or they thought, let's not confuse our fans with confusing naming schemes, and just add a one in the end. Note, they dropped the hyphen before the letter. To make it simpler. 2019: Which year is 2019? Right 2019. What comes after 71? Right: 72. So how to name a car which comes out in 2019 and whose predecessor had a 71 in the name? Right: The Ferrari SF90. Cause it's their 90th birthday, this time not as a cars manufacturer, but as a racing team. I guess easy naming schemes are just too...easy? 2020: 70...71...90.....1000. Ferrari SF1000. Well, in 2020 they had their 1000th race. 2021: SF21. It's still a hybdrid, but i guess they figured out that their is no ferri SF21 a SF21-H needs itself to differantiate itself from by name. So they dropped the H. 2022: Ferrari F1-75. F1 for Formula One, 75 because now it's their 75th birthday as a car company. 2023: Ferrari SF-23. So back to the old naming scheme, But since their is no hyphen infront of a letter at he back, they put one infront of the number. Maybe they liked the look of it from the last year. Don't get me started on their road cars. Their is even a whole section in their wikipedia article about this: en.wikipedia.org/wiki/Ferrari#Naming_conventions
5:34 "I left more confused than when I started" perfect description of trying to understand this conundrum, IT has a whole new terminology and rules of its own
NVidia use to have a suffix of "m" for laptop variant GPUs. They stopped doing this, so now they have both desktop and laptop GPUs with the same name while they have fairly different specifications.
I could be wrong but from what I remember, they did that starting with pascal because their mobile lineup more greatly matched their desktop counterparts. It’s kinda like how Nvidia have done away with GT cards, only offering GTX and RTX.
@@labombaromba It was only really the case for Pascal though, they shoulld have switched back like the Ti range, but that would impact sales I guess. But bless em for not trying to sell us a GT 430 rebrand for the 50th time :D
The laptop GPUs now have "Laptop" as suffix (from 30 series), but this makes searching for them more difficult as you'd be looking for example a 4060 Laptop laptop.
One trick I usually try to use is try to discover the real internal number used by the engineers, they usually are a lot more consistent, except when the entire division is fired.
I haven been learning about CPUs since 2018. I have learned that the best way of dealing with this is to just not trust the names and simply look up the specs. At least their official websites explain enough.
The fact that I thought you made up the naming scheme for AMD at first until I looked it up has me doubting there was even one combined brain cell in the room when they chose that system
The amount of times that someone tells me their computer is still *super high tech* because they have an i7 cpu, when their computer is probably nearing 10 years old, is enough to make me agree with you completely.
For server EPYCs I felt it was done on purpose to confuse, so only the sales people knew what to recommend to puzzled customers. For mobile laptops? Seems they actually want to bring it to this space too. A surprisingly good "not serious" video!
@@BWTHeuSeD Thanks, I looked at Wikichip's listing and it sort of makes sense but: - arbitrary placement of the F to denote the Frequency line (2nd or 3rd, since 9xxx Zen4 it's a suffix modifier!) - 2nd & 3rd sort of denote higher = better if not for F - The 7/9 leading number meaning is unconventional for the same reasons stated in the video (and I don't find Xeon numbering good either, but *nobody ever got fired for choosing Xeon's* numbering scheme right? ;) the only good thing is gen number is consistent
I don't think the EPYC CPUs have names done to confuse. The vast majority of buyers are companies which have dedicated people whose job is to select what to buy. And when buying they have to see the consumption, the number of CPUs per socket, memory support and number of channels. How the architecture works in specific workloads. They won't be fooled by a nice looking name.
@@VADemon Why not ? I think that Intel had a lot of success in the server space because there was no competition. Now it has, and it's not doing that well. I don't see the naming scheme to have any part in that.
if i ever started designing products in the tech world i would use ikeas naming scheme and give every product a real name. it would be incredible watching people argue online about whether the best value card on the market was last generations Morton that just received a price drop or the newly released budget card Tiny Francis
I'm a big nerd for firearms and weaponry in general, what's funny is a letter or two with a few numbers can be surprisingly easy to remember, like Intel's GPU names. One letter with a number gives you 260 possibilities, a letter and two numbers gives you 2600, etc. Do it twice and you can convey a lot of meaning, take M4A1 for example: approved model number 4, with approved variation number 1. New from AMD: the A1A1 CPU; Architecture, Generation, Model, Variant. Idk something like that.
I am more of a plane guy, and the way USAF and US Navy do it is quite simple: Letter to explain the use (attack, bomber, fighter, etc.) And a number. Then they add a word like Tomcat or Blackbird to make it more memorable. It is somewhat weird sometimes, but is really consistent
Or maybe even a standardized cross-company scheme like Japanese and American naval aircraft in WWII: Role/Model/Manufacturer/Variant A6M2 -> Fighter, 6th model from Mitsubishi, 2nd Revision
I feel like part of this confusing numbers is to purposefully confuse consumers to buy something with a higher number cause it's difficult to keep track of all these numbers
And the most recent entrant to the computer CPU game, Apple, already managed to make their naming confusing, too! They named their highest performing chip 'ultra' and the second best one 'max', even though maximum literally means the most possible...
The reason they skipped the Ryzen 6000 name I believe was to have the name of Zen 4 (desktop) match their new Radeon 7000 series. And the Ryzen 7000 mobile? I'm baffled, what a masterpiece of a naming scheme. They really did smoke something special over there at AMD.
That would be weird considering the 7900 XT and XTX are named as such to avoid having 2 things named simply 7900. The 7900XT is a 7900, 7900XTX is the 7900XT. The might have skipped the 6000 ryzen to avoid the same names with 6000 Radeon a bit longer
But what about the lower end GPUs Rx 7700, 7600, aren't they releasing those? In that case they would be lining up with the Ryzen 7700 and 7600. Seems more like them planning it rather than them trying to avoid clashing names. Who knows, that might change next generation with them skipping 8000 in either CPU or GPU lineup.
No, they skipped it to avoid the situation where Ryzen 6000 on mobile uses zen3 and DDR4 and zen4/DDR5 on desktop. Same reason we did not have Ryzen 4000 on desktop for a while, at the time they jumped to 5000 to correct the shitshow they made when they had Ryzen 3000 APUs on mobile and desktop based on zen+ and zen2-based 3000-series CPUs. Thanks to this correction the 5000 series on both mobile and desktop could be zen3 based, but they had to fuck it up by making some 5000 mobile CPUs zen2-based.
@@shepardpolska nonononnono we already have the 7900XTX and the 7900XT, it is only logical we'll have the 7900X, the 7900T, the 7900TX, and the 7900TXT, corresponding to, in order: 6900xt, 6800xt, 6700xt, 6600xt, 6500xt and 6400, and then there will be a 7800 series as equivalents to the non xt models, and a 7850 for shiggles and a 7950 just to confuse everyone just in time to compete with Arc B
Please dont forget that AMD FX processors had a number for core count that applied to almost all SKUs except Extreme CPUs that had 8 cores but a 9 in the name
Woo! Someone else who is excited for the Intel GPU naming scheme! It's the smallest least important thing ever, but it makes me happy that it looks like they're going to be doing it right
You know the naming conventions are cooked when philip says "I think" like the man who basically gets to the bottom of things was like "IDK that shit's too confusing" there's a special place in hell for marketing executives.
Ask for random people's pet's names and go with that, until you start getting death threats, and then hire a security team to keep your naming convention intact so Petey Parrot doesn't get taken off the shelves.
Great video! But why no mention of the biggest scam of all? Nvidia naming scheme in laptops. Before Pascal(1000 series) they added "M" to their mobile GPUs because they had different specs compared to desktop ones. In 1000 series and 2000 series, Nvidia equalized laptop and desktop specifications, so they removed the "M" from the name. But since 3000 series, they started to cut down mobile GPUs again AND they still have desktop naming as if they were the same as on desktops, when they aren't even close.
for my naming scheme, i would blend my product into a homogeneous goo and take its HEX colour for a name. If two have the same we add a little food colouring
If I were to start everything over For CPUs: 3140D "3" - generation, also last digit of year (2023) "1" - placing in the lineup (9 is top notch, 1 is the cheapest) "4" - number of cores divided by 2. 16-cores would have an "8", for example "0" - for marketing and it looks cool "D" - the letter at the end clearly denotes purpose, D is for desktop, M is mobile, MX for "juiced up" mobile, etc
For GPUs: G3640 "A" - at the start to distinguish from CPUs. Denotes purpose. G is for gamers, W for workstation (Quadro), etc "3" - generation, also architecture "6" - placing in the model range. Also spans the full lineup for the year. 9 is the top (like 3090 Ti), 1 is the lowest (like 3050) "4" - VRAM amount divided by 4. Would only allow 8 GB, 12 GB, 16 GB and so on, but would be consistent "0" - for marketing
I have always loved the old John Deere's numbering system which lasted from the early 90s to the late 2000s where you had 4 digits: the first being the market segment, the second being the power level and the third and the fourth would generation numbers. So you could have model numbers like "1500" for a first gen ("00"), mid level of power ("5"), entry level ("1") something; or even "9830" for a fourth gen ("30"), high level of power ("8"), high level ("9") something
For intel their scheming makes some sense since they have WAY MORE CPUs than they announce or document on their website. With many CPUs they have clock steppings they use the last number for and some server CPUs or dedicated usage CPUs there are different numbering for the last two numbers for cache
ok i saw your video on userbenchmark when it came out but i seriously still can't believe how batshit insane their reviews are, especially that one on the 7950x oh my god why are they going on a tangent about AMD GPUs this is supposed to be a review of the 7950x??? why does expensive RAM and motherboards mean people should wait to buy a different CPU with expensive RAM and motherboards??? im going insane
A lot of these feel like someone came up with a naming scheme that they think makes perfect sense (because they developed it), and then went over with exactly no one else.
The Intel naming scheme using letters is actually great! After having used the 26 latin letters, they could start using different alphabets, and we might get an omega graphics Card!
I've spent the past 3 months on and off designing a serial number system for a manufacturing project. this is an incredible headache to see, but it puts my struggle in perspective. At least I had ISO standards on my side..
Since you wanted to tell you what you missed - you missed a chance of talking about integrated GPU architecture in those mobile AMD CPUs (5:32) and how 7040 has RDNA3 while 7045 has RDNA2 (and that there is RDNA1 and obviously Vega included in the same generation, making it 4 CPU architectures and 4 GPU architectures in the same generation).
Thank you for this video. It's impossible for an average consumer to know how good a graphics card or a processor is. If you're buying AMD, I even find it difficult to find out which is a CPU and which is a GPU.
I never understood why they didn't use the last digit for the graphics cards, they could have avoided so much confusion. To be fair I did own and used the GTX 465 from Nvidia back in 2010, but that was literally the only card I can think of that they named as such. And even then I remember that a bunch of people claimed that the GTX 460 was faster than the 465 but I never really went and tested it myself. Also another fun fact, my processor from that time was an i5-661 so Intel did use the last digit for something and 2010 was a wild year. Nvidia could have avoided so much confusion with the DDR3 1030 vs GDDR5 1030, just call the DDR3 one the 1030 and the GDDR5 one the 1035. Same thing for the GTX 960 2GB and 4GB, call the 4GB model the 965 and wabam a lot of confused customers are no longer confused. The Ti business makes no sense either, its kind of naming something in between the segments, so the XX60 Ti falls between the XX60 and XX70, so why not call this thing a XX65 in the first place? I suppose confusing people that don't research what they are buying beforehand and selling inferior things for way more than they are worth is the point. While I really suck at naming things IRL, I think I would go the Ubuntu distro naming route. Literally put the year of the product in the first 4 digits and put a dash followed by 2 more digits for the performance tier. So a mid tier product would be named 2023-50 or a high tier would be 2023-90, the two digits allows for a lot of flexibility when putting products that are similar side by side and it is unlikely that 99 products would be made in a single year. It's not a fancy product name or anything, but it is straight to the point and avoids confusion.
Honestly, can‘t we just use a different version of semantic versioning? First number: Generation Second number: Model Third number: Revision So a chip might be an Intel 13.9.0 (i9 13900K). Or an Nvidia 40.90.0 (4090). We could also enforce not skipping a version or a model. So an i9 would actually be a x.4.x, i7 would be x.3.x, i5 would be x.2.x and an i3 would be x.1.x. Chances are, you‘re also not gonna care about the revision that much.
Honestly the best naming scheme i can think of is Product Grade + Gen for example 1 Gen 3 would be third generation of entry products while 3 Gen 2 would be mid tier second gen. But since that don't give much more room for models in between i think better one would be "logarithm" scale where each new step is order of magnitude bigger like entry level be 10, mid tier 100 and high end would be 1000. So if something in between comes up you shove it in between like 150 would be better than mid tier bit less good than high end.
My own naming scheme would be based on Minecraft mobs, with a number for the year it was released and an X or a Y for comparing it to the last year cards, where X indicates it's superior to any of the last year mobs, and Y for indicating it's only superior to the weaker mobs, like Ghast 24, Ender Dragon 25.
@@TrustyGun2 what? snapshot names make perfect sense tho. first 2 numbers are the year, w is week, next 2 are the week number in the year (1-52) and the letter after is the number of snapshot released in that week. so the 2nd snapshot released in the 30th week of 2023 would 23w30b
For GPUs, I'd go with something inspired by Intel's naming scheme First letter - architecture (A, B, C...) Second number - release year Third number - power class Fourth number - zero for marketing Nvidia's 3060 would be A120 (A for Ampere, 1 for 2021, 2 because it's the second down the stack, only 3050 behind it). The 4090 would be L370 (L for Lovelace).
I think Intel and Nvidia have generally done a pretty good job with their naming schemes and keeping them fairly consisten over time (though with Intel it gets a bit messy because of all the different types of laptop CPUs). AMD is the one that stands out continually having bad naming schemes and changing the naming schemes all the time. They literally have three different "6000 series" gpu series.
I'm so glad someone else understands! Ryzen CPU's had such a good naming scheme for the first couple generations, but it's just fucking awful now. Absolutely ridiculous, and difficult to follow.
A-00-9-U A slot goes for generation 00 is core count 9 is your performance or market segment, like your ryzen 3 etc U is just bonus letters. M for mobile, G for integrated graphics, and so on. My theory is that reading left to right it will quickly tell you how interested in the part you are, and also hyphens would make it easy to split and filter through them all.
to make a good naming scheme you have to define things in the order of relevance. If the product refreshes yearly, you can put a year as a generational indicator, if not, make up your own numbers or what I find better - codenames for generations, in alphabetical order. Those usually work pretty well. then there should be core count, speed rating, tdp rating somewhere in there, either represented by number or letter in that order
I would go for the monitor manufacturer approach and just throw random numbers and letters together with absolutely no meaning at all so that consumers couldn't easily tell the difference.
If I had to name processors, I would do it like this: -Company Name e.g. AMD; -Product Name e.g. Ryzen; -Generation Number starting from 1 and incrementing in steps by 1 e.g. 4; -Market Segment ranging from 1 to 9 with a higher number indicating a higher class of price and performance e.g. 7; -Extra Features where its either 0 or 5 where 5 for example indicates an iGPU e.g. 5; -plus an optional "M" if mobile -> AMD Ryzen 475M = A 4th Gen Mid-High End Mobile CPU with integrated graphics.
this byfar is the best one i have seen so far. and if you overflow 9 then who cares? just shift everything else except the gen number to the right. so an amd ryzen 10th gen mid/high end mobile cpu with integrated graphics would be amd ryzen 1075m
2:30 "XFX AMD Radeon (Trademark) RX 580 GTS *XXX* EDITION 8GB" skull_emoji. AMD really did make an XXX Edition to a Graphics Card. Y'know, because it is "Graphic".
I propose that we use boring naming schemes for mass production CPUs, bigger number names for high-performance consumer parts, slightly confusing mixture of alphanumeric codes for military processors, and for the top-performing super expensive prosumer processors we use the really ridiculous names the gaming companies are currently using.
The perfect naming scheme: Take the complete datasheet of your hardware, pull it through a hash function and you have a perfectly unreadable name wich totally makes sense.
lmao!
and the best thing: you cant even reconstruct the data from it, you'll just have to guess until you get it right. And with all those different parameters you'll actually need a password cracker for that. great success.
That’s probably how they name monitors lol
Just put the FutureMark value as a name.
might as well salt them too!
"The great thing about computer standards is there's so many to choose from!"
I saw AMDs New naming scheme explained like three times and I still don't know what they were smoking while coming up with this.
It's a way to sell rebranded old CPUs is all. Genius from their POV
@@defnotatroll it is genius. Its also anti consumer. Someone probably figured out less tech savvy people are more likely to buy laptops. Hence the horrible naming for marketing's sake. Make someone feel lime they're getting a really good ryzen 7 for a cheap price. Only for it to be super outdated. They won't know that or even notice it.
Its disgusting. Intentional confusion for more sales
@@Derpynewb deserves a lawsuit IMO
I think the issue they are having os that their older technology is still viable for certain use cases, but they are harder to sell with the old name. It's also not directly a scam. E.g. if you only need an office pc there is absolutely nothing wrong with the Gen 3 architecture for example.
It's a deliberate attempt to make it difficult to identify which generation of processor is in which chip.
About AMD and their "Portfolio Model Year",
Not sure if this is elaborate, but each year is divisible by its respective "series" number:
2023/7=289
2024/8=253
2025/9=225
But wait, there's more!
The digits of every year also add up to its "series" number:
2+0+2+3=7
2+0+2+4=8
2+0+2+5=9
If they were to follow this pattern, the 10 series would have to come out in 2080 at the earliest.
@@2kliksphilip My autism told me to stare at the numbers
I wonder what the “series” number for 2027 will be ...
@@lawrencedoliveiro9104 Well it has to skip to 2027, naturally.
@@UchihaMadara401 High five
@@UchihaMadara401 THE NUMBERS -MASON- DONATO WHAT DO THEY MEAN?
Samsung already figured this out. They skipped the s11 to go to s20 representing the year it came out in 2020, and then the s21 for 2021 and so on. This naming scheme is indefinite so long as years continue to pass.
Unless it lasts a century, but let's not kid ourselves.
But then it resets in the year 2120 and people will get confused between the two S20s!
Sadly that's still a problem, because companies like to freshen up things for no reason.
Take logos for example. Why Pepsi changed it? Or Microsoft?
what if years stop passing though
@@L5GUK They could call it the S120
2:26 The "XFX RX XTX" is the funniest thing I've heard all April Fool's day, thank you Philip
Whoever thought that up must drive a Mitsubishi GTO GS R.
@@lawrencedoliveiro9104 either that, or a Nissan Sentra SE R Spec V
Didn't gamers nexus already include the more X=better as a graph in the review?
Is it as bad as the Xbox 1, xbox one, xbox one s, xbox one x, xbox series s, xbox series x
@@lawrencedoliveiro9104the same type of person that rides a Honda CBR1000 RR-R SP
I used to work in the Radeon department at ATI (pre-AMD) and the "marketing name" as we called it, was totally disconnected from the name we used.
We were normally working on products that were nearly 2 years away from release, so there were many times where we'd have meetings about something that launched and we'd have Abbott and Costello level conversations where nobody knew what old (to us) product we were talking about, as we tried to convert marketing names back into engineering codenames from 2 years ago.
I knew you worked in tech but it's fun to see you talk about it
Holy shit.
Would it be possible for you to share how those internal naming schemes usually looked like? I'm interested how much info those engineering names carried (and how useful/unwieldy were they).
I realize it might be under NDA, but it's worth a shot asking (in case they were ever leaked, published or if the general naming scheme doesn't reveal much).
out of all youtubers that i watch, i wasn't expecting njb.
We have similar shit going on at Intel now. Somebody will mention that we did something interesting in "Devil's Canyon" and about half the room will have no idea what that means. We're trying to get the memo out to everybody to just refer to them by the generations.
they should name gaming equipment the same way they name American cars. i would certainly buy the new nvidia GPU if they named it "The Hellthrottler" or something similar
They would probably end up calling the Ti version Godstrangler splitting the divide again.
Gets so hot it starts throttling?
Threadripper names goes hard
Powercolor makes GPUs likes that, hellhound, red dragon, red devil
Well it depends. If you look at ferrari for example, their naming schemes are equally confusing, even though they have less products. But maybe that's the reason, if you want to confuse your customors, and you don't have lot's of products your confusing naming schemes have to be more confusing to confuse to the same amount. But they also are italian. So who knows.
Let's look at their forumla 1 cars for example and start in 2010:
2010: Ferrari F10. 10 for 2010. F for ferrari. Thus technically the Ferrari Ferrari 10.
2011: Ferrari 150º Italia. Because the state of Italy had it's 150th birthday.
2012: Ferarri F2012. Straight forward. F to indicate it's a ferrari. 2012 because 2012. You can see, they don't want to confuse us and added a 20 because of this, to not confuse us with the ferraris from 1920, even though they didn't exist back then. Or they wanted to make their naming scheme future proof. Something they could use for over 100 years. And? Did they?
2013: Ferrari F138. No, they didn't. They changed it the next year. F stand's again for for ferrari, 13 for 2013 and 8 because it's an 8 cylinder.
2014: Ferrari F14T. Yes, you guessed the Ferrari part right, also the F. The 14 stand's for 2014 (the 20 in the name is gone again, maybe it was to confusing). The T for the turbocharger it now got.
2015: SF15-T. Technically the Racing Team is called Scuderia Ferrari and a daughter of the mothercompany ferrari. So the SF stands for scuderia ferrari, which mean Ferrari Racing team. The 15 stands for 2015, the T for turbocharger, but for some reason they added a hypen infront of it for this year.
2016: SF16-H. SF16, well you get it. But now it's a H instead of a T at the end. This stands for hybrid. Because it is got a hybrid powertrain with a turbocharched engine, like the last two years cars. I guess you need also somethin to differntiate to following models from each other? Although, they already got the year in the name. The letter in the end could also be, that noone confuses it with the nonhybdrid Ferrari SF16. Although none such cars exists.
2017: SF70-H. Still a hybdrid, but ferrari was founded in 1957. So it's a car that came out 70 years after they started. So they put a 70 in the name to celebrate the round birthday.
2018: SF71H. I guess they also celebrate unround birthdays. Or they thought, let's not confuse our fans with confusing naming schemes, and just add a one in the end. Note, they dropped the hyphen before the letter. To make it simpler.
2019: Which year is 2019? Right 2019. What comes after 71? Right: 72. So how to name a car which comes out in 2019 and whose predecessor had a 71 in the name? Right: The Ferrari SF90. Cause it's their 90th birthday, this time not as a cars manufacturer, but as a racing team. I guess easy naming schemes are just too...easy?
2020: 70...71...90.....1000. Ferrari SF1000. Well, in 2020 they had their 1000th race.
2021: SF21. It's still a hybdrid, but i guess they figured out that their is no ferri SF21 a SF21-H needs itself to differantiate itself from by name. So they dropped the H.
2022: Ferrari F1-75. F1 for Formula One, 75 because now it's their 75th birthday as a car company.
2023: Ferrari SF-23. So back to the old naming scheme, But since their is no hyphen infront of a letter at he back, they put one infront of the number. Maybe they liked the look of it from the last year.
Don't get me started on their road cars. Their is even a whole section in their wikipedia article about this: en.wikipedia.org/wiki/Ferrari#Naming_conventions
5:34 "I left more confused than when I started" perfect description of trying to understand this conundrum, IT has a whole new terminology and rules of its own
NVidia use to have a suffix of "m" for laptop variant GPUs. They stopped doing this, so now they have both desktop and laptop GPUs with the same name while they have fairly different specifications.
I could be wrong but from what I remember, they did that starting with pascal because their mobile lineup more greatly matched their desktop counterparts.
It’s kinda like how Nvidia have done away with GT cards, only offering GTX and RTX.
@@labombaromba It was only really the case for Pascal though, they shoulld have switched back like the Ti range, but that would impact sales I guess.
But bless em for not trying to sell us a GT 430 rebrand for the 50th time :D
@@labombaromba But now the gap between mobile and desktop GPUs is widening with every generation.
Don't forget different laptops also have different GPU power targets so a "3080" running at 65W gets beat by a "3070" running at 120W
The laptop GPUs now have "Laptop" as suffix (from 30 series), but this makes searching for them more difficult as you'd be looking for example a 4060 Laptop laptop.
One trick I usually try to use is try to discover the real internal number used by the engineers, they usually are a lot more consistent, except when the entire division is fired.
I haven been learning about CPUs since 2018. I have learned that the best way of dealing with this is to just not trust the names and simply look up the specs. At least their official websites explain enough.
even better: look up benchmarks for the applications you're planning to use the CPU on.
@@-morrow I heard userbenchmark is a pretty good site ( ͡° ͜ʖ ͡°)
cpu world does a better job
at least there you won't be told that an athlon xp 3200+ is actually clocked at 3200mhz
but the average human eye can't even read specs!
@@-morrow like on userbenchmark? the road to hell is paved with good intentions.
The fact that I thought you made up the naming scheme for AMD at first until I looked it up
has me doubting there was even one combined brain cell in the room when they chose that system
I would simply give every CPU an UUID. Imagine buying the newest Intel 34950eb1-4384-440e-aa59-c446e66c2916.
Ah yes, the monitor route
Using the Uvuvwevwevwe Enyetwenwevwe Ugbemugbem Osas architecture.
@@zsideswapper6718 the legendary skit in the computing world.
@@zsideswapper6718 the legendary skit in the computing world.
so the Sony naming scheme?
The amount of times that someone tells me their computer is still *super high tech* because they have an i7 cpu, when their computer is probably nearing 10 years old, is enough to make me agree with you completely.
My laptop bears an i5.
It's more like an i3, and if we're going by current core/thread counts, well, does the Pentium ring any bells?
i7 860 💀
thanks intel naming, I can scam people buy a decade old i7. tech illiterate people rather buy i7-870 than Pentium G7400
"Have the outline of steve's hair burned into your retinas" - you got me, can't watch enough of tech jesus
For server EPYCs I felt it was done on purpose to confuse, so only the sales people knew what to recommend to puzzled customers.
For mobile laptops? Seems they actually want to bring it to this space too.
A surprisingly good "not serious" video!
EPYC's naming is more based on Intel's Xeon line. The first 3 numbers are used to denote segments, with the 4th to denote generation.
@@BWTHeuSeD Thanks, I looked at Wikichip's listing and it sort of makes sense but:
- arbitrary placement of the F to denote the Frequency line (2nd or 3rd, since 9xxx Zen4 it's a suffix modifier!)
- 2nd & 3rd sort of denote higher = better if not for F
- The 7/9 leading number meaning is unconventional for the same reasons stated in the video (and I don't find Xeon numbering good either, but *nobody ever got fired for choosing Xeon's* numbering scheme right? ;)
the only good thing is gen number is consistent
I don't think the EPYC CPUs have names done to confuse. The vast majority of buyers are companies which have dedicated people whose job is to select what to buy. And when buying they have to see the consumption, the number of CPUs per socket, memory support and number of channels. How the architecture works in specific workloads. They won't be fooled by a nice looking name.
@@Winnetou17 if this were always the case, Intel's sales strategy of wining and dining wouldn't have enjoyed the level of success it has seen
@@VADemon Why not ? I think that Intel had a lot of success in the server space because there was no competition. Now it has, and it's not doing that well. I don't see the naming scheme to have any part in that.
The fact that AMD's new mobile processors came with a decoder ring for the press really says a lot about society
litteraly 1984
I dunno what it's supposed to say about society, but it does say a lot about AMD's marketing arm
if i ever started designing products in the tech world i would use ikeas naming scheme and give every product a real name. it would be incredible watching people argue online about whether the best value card on the market was last generations Morton that just received a price drop or the newly released budget card Tiny Francis
I'm a big nerd for firearms and weaponry in general, what's funny is a letter or two with a few numbers can be surprisingly easy to remember, like Intel's GPU names. One letter with a number gives you 260 possibilities, a letter and two numbers gives you 2600, etc. Do it twice and you can convey a lot of meaning, take M4A1 for example: approved model number 4, with approved variation number 1. New from AMD: the A1A1 CPU; Architecture, Generation, Model, Variant. Idk something like that.
ah yes the all new B0O3 from Intel
I am more of a plane guy, and the way USAF and US Navy do it is quite simple: Letter to explain the use (attack, bomber, fighter, etc.) And a number. Then they add a word like Tomcat or Blackbird to make it more memorable. It is somewhat weird sometimes, but is really consistent
Or maybe even a standardized cross-company scheme like Japanese and American naval aircraft in WWII: Role/Model/Manufacturer/Variant
A6M2 -> Fighter, 6th model from Mitsubishi, 2nd Revision
I feel like part of this confusing numbers is to purposefully confuse consumers to buy something with a higher number cause it's difficult to keep track of all these numbers
6:06 THIS IS ALL YOUR FAULT. YOU JINXED IT.
And the most recent entrant to the computer CPU game, Apple, already managed to make their naming confusing, too! They named their highest performing chip 'ultra' and the second best one 'max', even though maximum literally means the most possible...
love the almost daily videos from all your channels and the quality is still great
There are other channels??!
I absolutely love the Intel ARC naming scheme, its fun to imagine what every letter could be
Same thing with android, i love when companies name their versions in alphabetical order
"Introducing our second generation graphics: Intel Bark!"
@@deevee5327 LMAO
@@deevee5327 I don't want to know what the third generation is
Intel Erika
"Youve watched every linus tech tips video this week"
I am shooked and appaled at this personal attack!
The reason they skipped the Ryzen 6000 name I believe was to have the name of Zen 4 (desktop) match their new Radeon 7000 series.
And the Ryzen 7000 mobile? I'm baffled, what a masterpiece of a naming scheme. They really did smoke something special over there at AMD.
That would be weird considering the 7900 XT and XTX are named as such to avoid having 2 things named simply 7900. The 7900XT is a 7900, 7900XTX is the 7900XT.
The might have skipped the 6000 ryzen to avoid the same names with 6000 Radeon a bit longer
But what about the lower end GPUs Rx 7700, 7600, aren't they releasing those? In that case they would be lining up with the Ryzen 7700 and 7600. Seems more like them planning it rather than them trying to avoid clashing names. Who knows, that might change next generation with them skipping 8000 in either CPU or GPU lineup.
@@Jord8043 we don't know if they will be called 7600 and 7700, there also might only be XT and XTX models of those
No, they skipped it to avoid the situation where Ryzen 6000 on mobile uses zen3 and DDR4 and zen4/DDR5 on desktop. Same reason we did not have Ryzen 4000 on desktop for a while, at the time they jumped to 5000 to correct the shitshow they made when they had Ryzen 3000 APUs on mobile and desktop based on zen+ and zen2-based 3000-series CPUs. Thanks to this correction the 5000 series on both mobile and desktop could be zen3 based, but they had to fuck it up by making some 5000 mobile CPUs zen2-based.
@@shepardpolska nonononnono we already have the 7900XTX and the 7900XT, it is only logical we'll have the 7900X, the 7900T, the 7900TX, and the 7900TXT, corresponding to, in order: 6900xt, 6800xt, 6700xt, 6600xt, 6500xt and 6400, and then there will be a 7800 series as equivalents to the non xt models, and a 7850 for shiggles and a 7950 just to confuse everyone just in time to compete with Arc B
Intel is apparently chaning their i -naming scheme :D
Philip's been cursed to make scripts that become outdated within a month.
Please dont forget that AMD FX processors had a number for core count that applied to almost all SKUs except Extreme CPUs that had 8 cores but a 9 in the name
5:44 I was laughing so hard when you proposed the 9915e. What a BEAST! XD
I've had coworkers ask me about building computers. At least half of what they don't understand is these horrible names.
Woo! Someone else who is excited for the Intel GPU naming scheme! It's the smallest least important thing ever, but it makes me happy that it looks like they're going to be doing it right
Intel does a lot of things right, things that many people never see or hear about, things that the operating system abstracts away for the user.
You know the naming conventions are cooked when philip says "I think" like the man who basically gets to the bottom of things was like "IDK that shit's too confusing"
there's a special place in hell for marketing executives.
One of the main reasons I'll never buy AMD is because my brain is full and I can't add their arcane naming scheme into my mind
Wait till he dives into monitor naming schemes
Ask for random people's pet's names and go with that, until you start getting death threats, and then hire a security team to keep your naming convention intact so Petey Parrot doesn't get taken off the shelves.
I feel called out. I *have* watched every linus video, since his fire truck. God damn it that was visceral and raw Kliks.
fantastic video phillip. perfectly scripted, timed and narrated. the klik empire is still as beautifully unique as the day i started watching
Great video! But why no mention of the biggest scam of all? Nvidia naming scheme in laptops. Before Pascal(1000 series) they added "M" to their mobile GPUs because they had different specs compared to desktop ones. In 1000 series and 2000 series, Nvidia equalized laptop and desktop specifications, so they removed the "M" from the name. But since 3000 series, they started to cut down mobile GPUs again AND they still have desktop naming as if they were the same as on desktops, when they aren't even close.
for my naming scheme, i would blend my product into a homogeneous goo and take its HEX colour for a name. If two have the same we add a little food colouring
What if your hex colour already had a name? Like maybe the "brown" or "chocolate" from html colours, they are really misleading
April Fools video at 23:48 on April 1st? Impeccable timing, Philip
If I were to start everything over
For CPUs: 3140D
"3" - generation, also last digit of year (2023)
"1" - placing in the lineup (9 is top notch, 1 is the cheapest)
"4" - number of cores divided by 2. 16-cores would have an "8", for example
"0" - for marketing and it looks cool
"D" - the letter at the end clearly denotes purpose, D is for desktop, M is mobile, MX for "juiced up" mobile, etc
For GPUs: G3640
"A" - at the start to distinguish from CPUs. Denotes purpose. G is for gamers, W for workstation (Quadro), etc
"3" - generation, also architecture
"6" - placing in the model range. Also spans the full lineup for the year. 9 is the top (like 3090 Ti), 1 is the lowest (like 3050)
"4" - VRAM amount divided by 4. Would only allow 8 GB, 12 GB, 16 GB and so on, but would be consistent
"0" - for marketing
@Ренхай these are cool but we already have more than 16core and probably would see more than 36GB VRAM in few years.
"Sticks and stones will break my bones but names will never hurt me", nuff said.
I have always loved the old John Deere's numbering system which lasted from the early 90s to the late 2000s where you had 4 digits: the first being the market segment, the second being the power level and the third and the fourth would generation numbers. So you could have model numbers like "1500" for a first gen ("00"), mid level of power ("5"), entry level ("1") something; or even "9830" for a fourth gen ("30"), high level of power ("8"), high level ("9") something
Now you need to do a video on the ridiculousness that is the USB naming scheme
or monitor naming schemes which are even more retarded
For intel their scheming makes some sense since they have WAY MORE CPUs than they announce or document on their website. With many CPUs they have clock steppings they use the last number for and some server CPUs or dedicated usage CPUs there are different numbering for the last two numbers for cache
Love you and your videos 💜💜💜
ok i saw your video on userbenchmark when it came out but i seriously still can't believe how batshit insane their reviews are, especially that one on the 7950x oh my god
why are they going on a tangent about AMD GPUs this is supposed to be a review of the 7950x??? why does expensive RAM and motherboards mean people should wait to buy a different CPU with expensive RAM and motherboards??? im going insane
I only wish no manufacturer bought a single one of their chips and let them rot with all of their stock
A lot of these feel like someone came up with a naming scheme that they think makes perfect sense (because they developed it), and then went over with exactly no one else.
The Intel naming scheme using letters is actually great! After having used the 26 latin letters, they could start using different alphabets, and we might get an omega graphics Card!
Atleast the naming schemes arent as bad as 3kliksphilip, 2kliksphilip and kliksphilip
lol
Back here after Nvidia decided that naming the 4070ti Super was a good idea.
I've spent the past 3 months on and off designing a serial number system for a manufacturing project. this is an incredible headache to see, but it puts my struggle in perspective. At least I had ISO standards on my side..
I never realised Intel GPUs were going for the alphabet, great catch
Since you wanted to tell you what you missed - you missed a chance of talking about integrated GPU architecture in those mobile AMD CPUs (5:32) and how 7040 has RDNA3 while 7045 has RDNA2 (and that there is RDNA1 and obviously Vega included in the same generation, making it 4 CPU architectures and 4 GPU architectures in the same generation).
There are no RDNA1 iGPUs, only RDNA2, RDNA3, and Vega.
Thank you for this video. It's impossible for an average consumer to know how good a graphics card or a processor is. If you're buying AMD, I even find it difficult to find out which is a CPU and which is a GPU.
I think we should randomly pull from the library of babel
I never understood why they didn't use the last digit for the graphics cards, they could have avoided so much confusion. To be fair I did own and used the GTX 465 from Nvidia back in 2010, but that was literally the only card I can think of that they named as such. And even then I remember that a bunch of people claimed that the GTX 460 was faster than the 465 but I never really went and tested it myself. Also another fun fact, my processor from that time was an i5-661 so Intel did use the last digit for something and 2010 was a wild year.
Nvidia could have avoided so much confusion with the DDR3 1030 vs GDDR5 1030, just call the DDR3 one the 1030 and the GDDR5 one the 1035. Same thing for the GTX 960 2GB and 4GB, call the 4GB model the 965 and wabam a lot of confused customers are no longer confused. The Ti business makes no sense either, its kind of naming something in between the segments, so the XX60 Ti falls between the XX60 and XX70, so why not call this thing a XX65 in the first place? I suppose confusing people that don't research what they are buying beforehand and selling inferior things for way more than they are worth is the point.
While I really suck at naming things IRL, I think I would go the Ubuntu distro naming route. Literally put the year of the product in the first 4 digits and put a dash followed by 2 more digits for the performance tier. So a mid tier product would be named 2023-50 or a high tier would be 2023-90, the two digits allows for a lot of flexibility when putting products that are similar side by side and it is unlikely that 99 products would be made in a single year. It's not a fancy product name or anything, but it is straight to the point and avoids confusion.
Just name your hardware according to the tick in nano-seconds since January 1st. 1970. People will get it.
you didn't even mention the worst of the lot, monitors.
H in AMD, Intel, Nvidia, stand for Has a sensible naming scheme
"Unlocked, Full Alder Lake CPU clocked at 3.20GHz" sounds way better than i9-12900K in my opnion
the naming confusion is what forces me to use userbenchmark to compare performace
I gotta give credit to Intel for at least sticking to it (mostly) since I build PC's (started with an i7 920).
Honestly, can‘t we just use a different version of semantic versioning?
First number: Generation
Second number: Model
Third number: Revision
So a chip might be an Intel 13.9.0 (i9 13900K). Or an Nvidia 40.90.0 (4090). We could also enforce not skipping a version or a model. So an i9 would actually be a x.4.x, i7 would be x.3.x, i5 would be x.2.x and an i3 would be x.1.x. Chances are, you‘re also not gonna care about the revision that much.
Honestly the best naming scheme i can think of is Product Grade + Gen
for example 1 Gen 3 would be third generation of entry products while 3 Gen 2 would be mid tier second gen. But since that don't give much more room for models in between i think better one would be "logarithm" scale where each new step is order of magnitude bigger like entry level be 10, mid tier 100 and high end would be 1000. So if something in between comes up you shove it in between like 150 would be better than mid tier bit less good than high end.
My own naming scheme would be based on Minecraft mobs, with a number for the year it was released and an X or a Y for comparing it to the last year cards, where X indicates it's superior to any of the last year mobs, and Y for indicating it's only superior to the weaker mobs, like Ghast 24, Ender Dragon 25.
I would rather go with the snapshot version naming scheme, at least it's a different brand of illegible
@@TrustyGun2 snapshots make sense though, they are named by the week they are released
@@TrustyGun2 what? snapshot names make perfect sense tho. first 2 numbers are the year, w is week, next 2 are the week number in the year (1-52) and the letter after is the number of snapshot released in that week.
so the 2nd snapshot released in the 30th week of 2023 would 23w30b
Hahaha, so much for intel’s consistent naming scheme…
Finally someone mentioned that FX 5900XT, it was bugging me for years how they chose to do that because they were losing to 9800XT
The third number was, at some point, indicating the percentage of the die size of the gpu.
Coming up with a useful naming convention is the reason why I haven't made my own line of CPUs and GPU
Unix timestamp of when the CPU is released.
And now Intel is getting rid of the i-naming lmao
For GPUs, I'd go with something inspired by Intel's naming scheme
First letter - architecture (A, B, C...)
Second number - release year
Third number - power class
Fourth number - zero for marketing
Nvidia's 3060 would be A120 (A for Ampere, 1 for 2021, 2 because it's the second down the stack, only 3050 behind it). The 4090 would be L370 (L for Lovelace).
I thought I was the only one struggling with this naming scheme. I also can't comprehend phone cpus.
Can't understate the value of the Intel Ark website either. It's so nice to google a model number and see all the specs and be able to compare them.
Came back to this after seeing Intel's new naming scheme. The pain
I think Intel and Nvidia have generally done a pretty good job with their naming schemes and keeping them fairly consisten over time (though with Intel it gets a bit messy because of all the different types of laptop CPUs). AMD is the one that stands out continually having bad naming schemes and changing the naming schemes all the time. They literally have three different "6000 series" gpu series.
@@2kliksphilip is it that hard to tell a single letter apart?
0:32 there's no need to keep watching past this point. The video officially peaks here, calling out my retinas
I'm so glad someone else understands! Ryzen CPU's had such a good naming scheme for the first couple generations, but it's just fucking awful now. Absolutely ridiculous, and difficult to follow.
A-00-9-U
A slot goes for generation
00 is core count
9 is your performance or market segment, like your ryzen 3 etc
U is just bonus letters. M for mobile, G for integrated graphics, and so on.
My theory is that reading left to right it will quickly tell you how interested in the part you are, and also hyphens would make it easy to split and filter through them all.
yes thats good and all, but there's this guy named kliksphilip who has named his music atomic amnesia's remake as "Atomic amnesia MMX". crazy guy
to make a good naming scheme you have to define things in the order of relevance. If the product refreshes yearly, you can put a year as a generational indicator, if not, make up your own numbers or what I find better - codenames for generations, in alphabetical order. Those usually work pretty well.
then there should be core count, speed rating, tdp rating somewhere in there, either represented by number or letter in that order
He starts the video and singlehandedly destroys user benchmarks reputation in a single breath.
I saw AMDs New naming scheme explained like three times and I still can't wrap my head around it.
When he talked about me being a tech genius, i felt that
I'm still waiting for the RX 7969 xxxtx
6:20 so uhhh, about that...
this videos giving me a head ache man
I would go for the monitor manufacturer approach and just throw random numbers and letters together with absolutely no meaning at all so that consumers couldn't easily tell the difference.
I never ever understood AMD naming scheme and I got familiar with Nvidia and Intel easily enough. I just gave up on keep up with their tech years ago.
This is why you should never speak to common people.
I just found out yesterday that my father ordered an AMD RX 580 2048SP and not an AMD RX 580. The AMD RX 580 2048SP is a rebranded AMD RX 570.
3:00 small error, zen 2 in bottom line is already visible
If I had to name processors, I would do it like this:
-Company Name e.g. AMD;
-Product Name e.g. Ryzen;
-Generation Number starting from 1 and incrementing in steps by 1 e.g. 4;
-Market Segment ranging from 1 to 9 with a higher number indicating a higher class of price and performance e.g. 7;
-Extra Features where its either 0 or 5 where 5 for example indicates an iGPU e.g. 5;
-plus an optional "M" if mobile
-> AMD Ryzen 475M = A 4th Gen Mid-High End Mobile CPU with integrated graphics.
this byfar is the best one i have seen so far. and if you overflow 9 then who cares? just shift everything else except the gen number to the right.
so an amd ryzen 10th gen mid/high end mobile cpu with integrated graphics would be amd ryzen 1075m
The XFX RX XTX... I wasn't prepared for this.
2:30 "XFX AMD Radeon (Trademark) RX 580 GTS *XXX* EDITION 8GB" skull_emoji. AMD really did make an XXX Edition to a Graphics Card. Y'know, because it is "Graphic".
I propose that we use boring naming schemes for mass production CPUs, bigger number names for high-performance consumer parts, slightly confusing mixture of alphanumeric codes for military processors, and for the top-performing super expensive prosumer processors we use the really ridiculous names the gaming companies are currently using.
Can't wait to try out the RTX 1204090Ti Super OC Edition in 2090!