Its feels like a video making by AI using wikipedia, have so many mistakes. Personally, i prefer Branch Education and Lawrence Systems for this type of videos. Are so much better.
Well it's either use an AI voice, or use your own voice and have it stolen and used by UA-cam to train their AI without permission. I guess they could also hire a voice actor for a $50-100 bucks and let UA-cam steal their voice instead, but that seems like a pretty unnecessary expenditure on a channel that probably isn't really earning much.
@@slimjimjimslim5923 What exactly do you mean? The 7000s and 9000s x3D chips are faster than the 5800x3D. Despite that, the 5800x3D still is an absolute powerhouse for gaminig. :)
@@cristianovoa I mean, he skipped completely the Intel 13/14th gen degradation issue. The whole video felt that 90% of the real work has been done in the first 5 minutes then it was all AI script
Fun Fact: Intel hasn't changed their L3 Cache Topology design in their entire career as CPU maker. They only changed the size capacity and wiring, but not the Topology. That's why Ryzen's 3DV-Cache is absolute game changer in history of semiconductor
Like part of the video, this is not correct either. See Skylake-X Mesh Interconnect and the Ring Bus. With Sandy Bridge, the L3 cache is integrated/baked in with the cores through a bus, as opposed to being wired like it was with Nehalem. Also, the L3 cache on SB and following processors was actually sliced but in a pool, whereas it was a large pool with Nehalem.
@@kingeling do you know what "Topology" means? You failed basic English? Skylake-C Mesh interconnect & the Ring Bus doesn't mean revolutionary in terms of Topology if it's still Horizontal. Horizontal is Horizontal, no matter what kind of marketing words they use
You skipped over several important CPUs (and several less important ones). Beyond the iAPX ones and the Itanium ones (good riddance), you skipped the 4004, 8085, 80286, and i486.
I mean, it's okay, if you optimize software for it. Productivity and video encoding for sure used these extensions. But for most, Haswell didn't introduce much relevant things to justify an upgrade over the already excellent Sandy Bridge/Ivy Bridge.
@@yukinagato1573 Haswell introduced 22nm FinFET. I have a Haswell 35 watt sustained i7-4700MQ laptop, that mobile CPU is just as fast a desktop i7-3770. Sandy Bridge vs Haswell at stock is worse, Sandy Bridge is well known because of the ridiculous overclocking headroom until it was finally surpassed by Alder Lake (watch BCLK overclock of the i3-12100 and the i5-12400). Sandy Bridge also was the last CPU CORE gen before Alder Lake that did a 20ish% IPC increase.
8:09... that's incorrect. Core i3 and Core i5 processors debuted on the lga1156 platform which is part of Core i's 1st Gen family. All lga1156 and lga1366 processors belonged to Core i's 1st Generation family. Sandy Bridge is Intel's 2nd Generation
@TuxikCE I interpreted it as a monotone way of reading the script. I can see where people may think it's AI. However the AI's i've heard have a lot less stability and exaggerate words strangely. i don't hear that here.
4770K was a decent jump over the 3770K, about 15% extra IPC. Also, the most notable feature of the 5775C wasn't it's efficiency, was it's integrated graphics and the 128MB eDRAM/L4 cache, much stronger than anything else before it, even beating AMD APUs from it's time.
dude you shouldve used the same VA as the Nvidia and Radeoon ones. the AI voice is obvious and makes the video feel cheap, like it was made by a content farm youtube channel
Probably not to make the video too long. Which is a shame, since those have such interesting stories. I'm also bummed about it having no mentions whatsoever about NetBurst's problems.
The amount of errors and misinformation in this video is laughable. Please stop trying to making educational videos when you have no idea what you're talking about.
@@Byfroft "MilliHertz" lol. The pricing of the CPUs weren't anywhere NEAR what this video states. It clearly show how this stupid youtuber does not understand how CPU pricing worked back in the day. Not a single person in existence paid the prices he quoted even on Day 1 launch. LOL Truly laughable. I've been building computers for 30 years now and I've bought plenty of them back in the day. This video is so wrong it's sad. Further more all of this video is, is a AI BOT reading from the wiki page and providing zero insight. Pathetic.
@@Byfroft 8:07 It stated the Core i3 and i5 didn't make an appearance until 2nd gen (Sandy Bridge), which is false. Quad core Core i5 debuted with the Lynnfield architecture in 2009, and Clarkdale (which included both Core i3 and the dual core Core i5) debuted in 2010, both for LGA1156. These are all part of 1st gen, and are all based on Nehalem.
Even if core ultra processors perform a little bit less im still very happy because of their efficiency they make direct competition of arm based CPU's without the need for translation.
So intel kept the same 4 cores 8 threads from the 1st generation i7 all the way up to 7th generation i7. Intel increased the core count on 8th gen because of the rise of AMD Ryzen.
@@TheVirtualArena24 I think the right way of saying this is "manufacturing process of 10 micrometers", or "10 micrometers manufacturing process". Not the way it's said. But well. Just a grammar error.
@@yukinagato1573 by no means I am an English expert and English is not my native language also, but that sentence sounds right to me? Also I just checked online and there's no errors but idk.
@@TheVirtualArena24 I mean, neither my native language is English btw haha. But as a computer engineer myself, it feels more correct to me "manufacturing process of 10 micrometers" than the other way around. The meaning of it goes like this: "through the process of manufacturing, you get 10 micrometers transistors*"; and not "through the 10 micrometers* you get a manufacturing process". It's the tech you have that will make you 10 micrometers*, not the 10 micrometers* that'll make the manufacturing tech. *Quick, completely ignorable notice: lithography in general (5nm, 3nm, 2nm, etc.) currently has no relation with the actual size of the transistors themselves. It's just a convention (or rather, a marketing convention) to point out the manufacturing generation of a product. A 2nm process CPU has no actual 2nm transistors.
Most important part in a computer is the GPU nowadays. The CPU can bottleneck your system if it is extremely weak but not really. Even a quadcore from 5 years ago can perfectly do anything ordenary and high end game if the GPU is strong enough.
Also totally missed the Celeron line for Pentiums, and that the Pentium II and III, the "cartidge shaped" processors, can be connected in dual motherboards so you can have two CPUs working together.
Probably not to make the video too long. If they were gonna talk about all Atoms, Celerons, Pentiums, Core i3/i5/i7/i9, Core 3/5/7/9, Core Ultra 3/Ultra 5/Ultra 7/Ultra 9, Pentium II Xeons, Pentium III Xeons, Xeons and Itanium chips that ever existed, we would be here for months.
Is this real or AI voice? Appreciate the moment of nostalgia about how awesome we have it right now…especially the level of nostalgia when I look at Intel shares going back in time 20 years 😅😅
Can you run them in the first place? Sandy Bridge lacks AVX2 instructions, this is why i heared that the maximum oldest Intel CPU gen for general tasks is Haswell because it has AVX2.
@@yukinagato1573 What? Respect the Pentium 4, it was the first CPU with hyperthreading or SMT. Meanwhile you people forget about the Pentium D (a chiplet CPU), that makes Bulldozer look good. That's how bad it is. Gosh, and if we keep going the chiplet route, will end up with Pentium D crap again (i'm on Alder Lake i7 K with DDR5 right now).
@@saricubra2867 Pentium 4 only introduced SMT to keep its pipeline busy between each time it stalls. All the great ideas NetBurst introduced were only so to try to save the whole architecture from pipeline inefficiencies. And once they increased its pipeline to 31 stages (Prescott), not only they had very low IPC, but also extreme thermal dissipation and power consumption. At this point, there was nothing that could save NetBurst. Pentium D is not a chiplet CPU. It's just a multi-chip CPU. But most of its shortcomings come from NetBurst's shortcomings.
@@saricubra2867 Pentium 4 only introduced SMT to keep its pipeline busy every time it stalls. All of NetBurst's great ideas were introduced just to alleviate pipeline inefficiencies. But once they increased the pipeline to 31 stages (Prescott), not only it had very low IPC but extremely high power consumption and thermal dissipation as well. There was nothing they could do anymore to save NetBurst. Most of the Pentium D's shortcomings were because of NetBurst, not really because of itself.
@@saricubra2867 Pentium 4 only introduced SMT to keep its pipeline busy every time it stalls. All of NetBurst's great ideas were introduced just to alleviate pipeline inefficiencies. But once they increased the pipeline to 31 stages (Prescott), not only it had very low IPC but extremely high power consumption and thermal dissipation as well. There was nothing they could do anymore to save NetBurst.
This whole video isn't the "evolution" of Intel CPUs but a low-effort, AI-assisted slideshow of the specs of their highest end SKUs, with some mistakes here and there. - 4:05 millihertz - 8:20 "i7 two-hundred-sixty K" - leaving out information such as transistor count for numerous chips, even though it's a quick search on Google, e.g for the 8700K at 11:07 - lack of understanding of Intel's architecture reshashes, most notably Skylake being used up until they cmae up with Rocket Lake. It wasn't just 14nm, it was the exact same core architecture. - usual uninformed joe's meaningless comparison of clock speeds across different architectures, you'd think niktek would know better. Well, apparently not. - 13:24 E-cores have in fact been proven to be less efficient than P-cores. Reviewers view them as "cost-efficient cores" rather than the power efficient cores Intel claims they are. - 8:08 First gen i5 and i3 models do in fact exist - 6:15 no mention of P4 being a hot mess compared to P3. It pulled 2-3 times more power because of its inefficient design. - 7:48 SMT already was a thing in Netburst (P4), it wasn't new at all with Nehalem - 9:50 no mention of eDRAM - 14:00 no mention of the recent Raptor Lake degradation fiasco - 14:44 "This processor has less threads" - superficial and doesn't explain why, nor the consequences
You totally missed i186 (a microcontroller), i286 that was the 16 bit reference CPU for a long, long time for PCs at 16Mhz and 20Mhz, and I486 that also lasted several years and was much powerfull than 386 and introduced many features like instruction cache. Also, when you say "first 16 bits or 32 bits CPU" your are referring for INTEL CPUs, as higher bit cpus was manufactured from other comapanies for mainframes and industrial uses ( Alpha, SPARC, Motorola, etc)
And it was a flop because Skylake outperformed it by IPC alone and DDR4. Just like any Alder Lake i7 or i9K chip with DDR5 and is better than Zen 3 X3D
@@saricubra2867 What no the I7 5775C did even peform like Kabylake. Broadwell was bit under 50$ more and was on the older 14nm node. The first 14nm node that would follow for many Gens. It also came just few months before the 6th gen out so the Cpu was already in a situation that was rather wait for the next gen the new gen. That also why many people bought Skylake. Alder lake was jump from AMDs win of the 5000. But in gaming AMD still had someting with the X3D. It was head to head. Ofc Productivity was Intel way better. But as you can see now X3D is just way better and Intel cant do anything. Intel had a X3D CPU but it was not worked on to perfection.
@ Bro, Broadwell also falls behind Kaby Lake that increased clockspeeds. The 5775C was a dumpsterfire of an architecture. Too expensive, clockspeed dropped so much that hurted it's overall perfomance because it was so hot. "Alder lake was jump from AMDs win of the 5000. But in gaming AMD still had someting with the X3D. It was head to head" What reality?, Alder Lake still has significantly higher IPC and the DDR5 advantage over Zen 3 X3D. Reviewers were pairing high core count Alder Lake with @ss RAM vs AMD to artificially inflate the X3D scores (and AMD always does worse with memory scaling vs Intel). It's all a scam. I play Sonic Unleashed and my 12700K is over 50% faster than a 5800X3D. Zen 3 X3D is even worse than the 5775C. It was overpriced and early X3D suffers from overheating problems. Very overhyped as well, the 5800X3D has clockspeed jitter issues, it showed up in frametime graphs that for obvious reasons it was ignored because mainstream tech youtube is all about influencers. "But as you can see now X3D is just way better and Intel cant do anything" i9-13980HX is a raptor lake mobile CPU and has better power efficency than Zen 4 desktop X3D, faster than a 13700K, a reduction of power use above 100 watts. Raptor Lake is basically IPC from 2021. 3 years latter and AMD doesn't beat Alder Lake IPC. We have Lunar Lake that competes with the Apple M3 series in perfomance per watt. Then we get the 9700X, and the 12700K from 2021 still has 4% better IPC. Zen 4 and Zen 5 are a flop, just like 13th gen and 14th gen. Just because the X3D are the ones that suck less doesn't mean that they are magically ahead of everything else. For example, the Ryzen 9 7900X is criminally underrated. I watch how desktop CPUs are doing right now and are just boring compared to my 12700K. Also why no one is rereviewing Arrow Lake right now? When Alder Lake was new they gave it a pass with the horrible perfomance on the i9 (early adopter problems and optimization bugs) but Arrow Lake launches and basically nothing.
I'm sticking with my Intel Core i9-10900KF processor as long as I am using Intel Arc GPUs. I don't trust Windows use of P & E cores, yet which is a shame because I see the benefits on my MacBook M3 Pro. If anything, I might throw in the towel and make an entire AMD build when this one bottlenecks me hard on games.
Also totally missed Itanium processor line, the 64 bit intel architecture that totally failed. The 64 bit architecture of today's processors (Intel included) came from AMD.
Intel downfall was expected but it might not be what people think, I think part of it was their corporate dysfunction. When the market shift and they need to move fast the management seems so slow it was painful to watch.
This whole video isn't the "evolution" of Intel CPUs but a low-effort, AI-assisted slideshow of the specs of their highest end SKUs, with some mistakes here and there. - 4:05 millihertz - 8:20 "i7 two-hundred-sixty K" - leaving out information such as transistor count for numerous chips, even though it's a quick search on Google, e.g for the 8700K at 11:07 - lack of understanding of Intel's architecture reshashes, most notably Skylake being used up until they cmae up with Rocket Lake. It wasn't just 14nm, it was the exact same core architecture. - usual uninformed joe's meaningless comparison of clock speeds across different architectures, you'd think niktek would know better. Well, apparently not. - 13:24 E-cores have in fact been proven to be less efficient than P-cores. Reviewers view them as "cost-efficient cores" rather than the power efficient cores Intel claims they are. - 8:08 First gen i5 and i3 models do in fact exist - 6:15 no mention of P4 being a hot mess compared to P3. It pulled 2-3 times more power because of its inefficient design. - 7:48 SMT already was a thing in Netburst (P4), it wasn't new at all with Nehalem - 9:50 no mention of eDRAM - 14:00 no mention of the recent Raptor Lake degradation fiasco - 14:44 "This processor has less threads" - superficial and doesn't explain why, nor the consequences
3rd gen to 12th gen until AMD catched up and released probably the first 3D chip heading to 1nm. While Intel remain stuck at 2D for 3 yrs I thought they would retain their 12++ nm fab just improve on design to save money doesn't want to buy a new fab for a decade.
Correction* The MHz in the CPU specifications are incorrect, it's supposed to be Megahertz and not Millihertz, I'm sorry for the confusion it caused 🙏
How dare you unsub and dislike (need to watch this first XD) btw when 2nd part of worst games?
@@WlyPotlol, I'm almost done editing it. Will post it tomorrow :)
@@NikTek
NikTek was supposed to be a meme channel...
My boy is all grown up. 🏋️♂️
@@NikTek kk
@@NikTek do amd one please
This is not NikTek I subscribed to
_slap_
As long as it is about pc, i watch it
True og's know he does these from time to time, and the spice is nice
Its feels like a video making by AI using wikipedia, have so many mistakes.
Personally, i prefer Branch Education and Lawrence Systems for this type of videos.
Are so much better.
So you subscribed to be a toxic fanboy?
Your AI mistakenly names MHz as milihertz instead of megahertz. And overall, it is a soulless junk compared to any human voice over
Found the Anti Ai person
@Space97. Your best rebuttal for AI
Well it's either use an AI voice, or use your own voice and have it stolen and used by UA-cam to train their AI without permission. I guess they could also hire a voice actor for a $50-100 bucks and let UA-cam steal their voice instead, but that seems like a pretty unnecessary expenditure on a channel that probably isn't really earning much.
Am i stupid or is this an ai voice
It is 😑
Literally since the Nvidia RTX evolution video but nobody minds that because it’s the same thing
Who cares?
@@i_cri_evertim your mom
I could tell by the "milihertz" clock speed.
4:04 16-33 millihertz? AI shat it right there
MHz = Megahertz, not Miliherts !
That’s off by a factor of 10^9!
My 8700k lasted me a good while.
Now i have 9800X3D and I'll never look back
still using :)
i am building a financial bare bones i3 6100 system.
Welcome to Ryzen, mate. Threw my 13600K out for a bargain deal on a 7950x3D at my local hardware store. Goes super smooth with my 4090.
@@slimjimjimslim5923 What exactly do you mean? The 7000s and 9000s x3D chips are faster than the 5800x3D. Despite that, the 5800x3D still is an absolute powerhouse for gaminig. :)
@@MichaeltheORIGINAL1 u answered ur own question
Hyperthreading was introduced way back at the Pentium IV era... this video is full of errors. Keep making memes, please
Or just do better? I wouldn't mind this kind of content with actual effort put into it
@@cristianovoa I mean, he skipped completely the Intel 13/14th gen degradation issue. The whole video felt that 90% of the real work has been done in the first 5 minutes then it was all AI script
@@cristianovoa Little effort
I'm glad I wasn't the only one who thought this video was kinda garbage.
Fun Fact:
Intel hasn't changed their L3 Cache Topology design in their entire career as CPU maker.
They only changed the size capacity and wiring, but not the Topology.
That's why Ryzen's 3DV-Cache is absolute game changer in history of semiconductor
Like part of the video, this is not correct either. See Skylake-X Mesh Interconnect and the Ring Bus.
With Sandy Bridge, the L3 cache is integrated/baked in with the cores through a bus, as opposed to being wired like it was with Nehalem. Also, the L3 cache on SB and following processors was actually sliced but in a pool, whereas it was a large pool with Nehalem.
@@m8x425 that doesn't mean anything if the L3 Cache layout still in Horizontal.
Horizontal means same Topology.
@@niezzayt3809 Now you're just saying a new design is meaningless lmao
@@kingeling do you know what "Topology" means?
You failed basic English?
Skylake-C Mesh interconnect & the Ring Bus doesn't mean revolutionary in terms of Topology if it's still Horizontal.
Horizontal is Horizontal, no matter what kind of marketing words they use
No, Ryzen X3D is successful because DDR5 and the infinity fabric suck.
I've waited for this for so long I'm like
"It's been 84 years"
You skipped over several important CPUs (and several less important ones). Beyond the iAPX ones and the Itanium ones (good riddance), you skipped the 4004, 8085, 80286, and i486.
We got informative Niktek before GTA 6👀
This is the private equity version of niktek. I could have read this on Wikipedia
8:58, you forgot to mention architectural improvements, which made Haswell really great, including the new AVX2 instructions.
I mean, it's okay, if you optimize software for it. Productivity and video encoding for sure used these extensions. But for most, Haswell didn't introduce much relevant things to justify an upgrade over the already excellent Sandy Bridge/Ivy Bridge.
@@yukinagato1573 Haswell introduced 22nm FinFET. I have a Haswell 35 watt sustained i7-4700MQ laptop, that mobile CPU is just as fast a desktop i7-3770.
Sandy Bridge vs Haswell at stock is worse, Sandy Bridge is well known because of the ridiculous overclocking headroom until it was finally surpassed by Alder Lake (watch BCLK overclock of the i3-12100 and the i5-12400). Sandy Bridge also was the last CPU CORE gen before Alder Lake that did a 20ish% IPC increase.
it's MEGAHERTZ!!!
8:09... that's incorrect. Core i3 and Core i5 processors debuted on the lga1156 platform which is part of Core i's 1st Gen family. All lga1156 and lga1366 processors belonged to Core i's 1st Generation family. Sandy Bridge is Intel's 2nd Generation
Thank you for this video and yes I would like to see AMD CPU History
02:00 i hate the sponsors 😠
we all do...
Maybe chapters are there to get you off
@@Pearloryx i do it
If you don't need it, just skip that part. There's no need to express your dislike.🥰
Well that's kinda disappointing
i love tech so much and learning about it i can never get enough thanks for the upload
I thought we were going to make fun of intel for 15mins straight, there is clearly a misunderstanding in this upload about the target audience 😅
Awesome Mini Documentary NikTek!
is the entire video is AI generated & its not good
No it's not, WTF are you talking about? The editing and voice are exactly what you expect from normal editing and production.
@@HM-rz8nv The voice
@TuxikCE I interpreted it as a monotone way of reading the script. I can see where people may think it's AI. However the AI's i've heard have a lot less stability and exaggerate words strangely. i don't hear that here.
Milihertz 😂
It's definitely AI, wtf are you talking about? It pronounced a lot of sentences very weird to the point of not even sounding human.
Man don't go the AI generated garbage route, this channel was so fun.
AI is the future though. If you can't keep up, stay out of the Internet boomer!! 😂
This is an informative tutorial of hiPDF, thanks for sharing🤗
4770K was a decent jump over the 3770K, about 15% extra IPC.
Also, the most notable feature of the 5775C wasn't it's efficiency, was it's integrated graphics and the 128MB eDRAM/L4 cache, much stronger than anything else before it, even beating AMD APUs from it's time.
That "instability issue" line got me coughing for a moment there🤣
dude you shouldve used the same VA as the Nvidia and Radeoon ones. the AI voice is obvious and makes the video feel cheap, like it was made by a content farm youtube channel
Dont do this again
Looking forward to the AMD version!
You missed Intel 80486 after 80386, Pentium variants, Itanium and other non-x86.
Probably not to make the video too long. Which is a shame, since those have such interesting stories. I'm also bummed about it having no mentions whatsoever about NetBurst's problems.
nice video bro
The amount of errors and misinformation in this video is laughable. Please stop trying to making educational videos when you have no idea what you're talking about.
Examples?
@@Byfroft "MilliHertz" lol. The pricing of the CPUs weren't anywhere NEAR what this video states. It clearly show how this stupid youtuber does not understand how CPU pricing worked back in the day. Not a single person in existence paid the prices he quoted even on Day 1 launch. LOL Truly laughable. I've been building computers for 30 years now and I've bought plenty of them back in the day. This video is so wrong it's sad. Further more all of this video is, is a AI BOT reading from the wiki page and providing zero insight. Pathetic.
@@Byfroft 8:07 It stated the Core i3 and i5 didn't make an appearance until 2nd gen (Sandy Bridge), which is false. Quad core Core i5 debuted with the Lynnfield architecture in 2009, and Clarkdale (which included both Core i3 and the dual core Core i5) debuted in 2010, both for LGA1156. These are all part of 1st gen, and are all based on Nehalem.
I have found some of these in older office PCs.
You also find cool stuff like custom built parts for a particular company.
Even if core ultra processors perform a little bit less im still very happy because of their efficiency they make direct competition of arm based CPU's without the need for translation.
So intel kept the same 4 cores 8 threads from the 1st generation i7 all the way up to 7th generation i7. Intel increased the core count on 8th gen because of the rise of AMD Ryzen.
This is NikTek I subscribed to
Me watching the whole video in anticipation for the punchline, now that is a real meme 😂
Musisz pamiętać że te procki to najdroższe modele i niewielu je kupowało, reszta z nich była całkiem przystępna cenowo.
0:47 "10 micrometers of manufacturing process" ai text to speech is fine, but an ai script is unacceptable
What's wrong in that sentence? I don't see any mistake..
The 4004 was in fact 10um, nothing wrong there
@@TheVirtualArena24 I think the right way of saying this is "manufacturing process of 10 micrometers", or "10 micrometers manufacturing process". Not the way it's said. But well. Just a grammar error.
@@yukinagato1573 by no means I am an English expert and English is not my native language also, but that sentence sounds right to me? Also I just checked online and there's no errors but idk.
@@TheVirtualArena24 I mean, neither my native language is English btw haha. But as a computer engineer myself, it feels more correct to me "manufacturing process of 10 micrometers" than the other way around. The meaning of it goes like this: "through the process of manufacturing, you get 10 micrometers transistors*"; and not "through the 10 micrometers* you get a manufacturing process". It's the tech you have that will make you 10 micrometers*, not the 10 micrometers* that'll make the manufacturing tech.
*Quick, completely ignorable notice: lithography in general (5nm, 3nm, 2nm, etc.) currently has no relation with the actual size of the transistors themselves. It's just a convention (or rather, a marketing convention) to point out the manufacturing generation of a product. A 2nm process CPU has no actual 2nm transistors.
You missed out on the Core 2 Quad
Well, given the Core 2 Quad is just a cut-down Core 2 Extreme (the quad core ones), it's kinda mentioned.
I love this, I hope you one day make an AMD version of this video ^^😊
Most important part in a computer is the GPU nowadays. The CPU can bottleneck your system if it is extremely weak but not really. Even a quadcore from 5 years ago can perfectly do anything ordenary and high end game if the GPU is strong enough.
Intel is transitioning into a GPU company the way it's going.
And it’s beating both Nvidia and Amd at their own game lol
Also totally missed the Celeron line for Pentiums, and that the Pentium II and III, the "cartidge shaped" processors, can be connected in dual motherboards so you can have two CPUs working together.
Probably not to make the video too long. If they were gonna talk about all Atoms, Celerons, Pentiums, Core i3/i5/i7/i9, Core 3/5/7/9, Core Ultra 3/Ultra 5/Ultra 7/Ultra 9, Pentium II Xeons, Pentium III Xeons, Xeons and Itanium chips that ever existed, we would be here for months.
Guys i think you dont know that NikTek does this from time to time. Same as AMD GPUs and Nvidia GPUs
Is this real or AI voice? Appreciate the moment of nostalgia about how awesome we have it right now…especially the level of nostalgia when I look at Intel shares going back in time 20 years 😅😅
Nice
That's a nicely done documentation NikTek. Also, do you have the version for AMD CPUs?
alternative title: How Intel went to top and made their own downfall decades later
MILLIHERTZ
You made Nvidia and Intel now AMD remains
I believe you could do better than this, i hope im not proven wrong, this kind of video would be a great watch with more effort and less ai
He’s a youtuber, they are not known for their effort lol
I used i7 2600k until last year, over 10 years old beast that can handle modern games(in reasonable settings, of course)
Can you run them in the first place? Sandy Bridge lacks AVX2 instructions, this is why i heared that the maximum oldest Intel CPU gen for general tasks is Haswell because it has AVX2.
0:53 it shows 750khz and ai voice says 740khz lol that's what happens when you use ai better to do yourself
Intel 4004 by Vicenza Inventor "Federico Faggin"
Esatto! 😍💅🏻🔥🇮🇹🔝 fiera di poter dire che il primo microprocessore al mondo l’ha inventato un uomo italiano!!!
I'm only here for the humor but today I learned something....
Niktek about to break the meme channel barrier
I had to wait 14 minutes for the punch line. Had me worried, I almost thought this wasn't a NikTek video.
P3 was my first cpu. Man it was the goat. Lasted till 2007 then switched to core 2 duo.
You managed to skip the entire Pentium 4/NetBurst era. Congrats. You probably don't know the hell it was.
@@yukinagato1573 What?
Respect the Pentium 4, it was the first CPU with hyperthreading or SMT.
Meanwhile you people forget about the Pentium D (a chiplet CPU), that makes Bulldozer look good. That's how bad it is.
Gosh, and if we keep going the chiplet route, will end up with Pentium D crap again (i'm on Alder Lake i7 K with DDR5 right now).
@@saricubra2867 Pentium 4 only introduced SMT to keep its pipeline busy between each time it stalls. All the great ideas NetBurst introduced were only so to try to save the whole architecture from pipeline inefficiencies. And once they increased its pipeline to 31 stages (Prescott), not only they had very low IPC, but also extreme thermal dissipation and power consumption. At this point, there was nothing that could save NetBurst.
Pentium D is not a chiplet CPU. It's just a multi-chip CPU. But most of its shortcomings come from NetBurst's shortcomings.
@@saricubra2867 Pentium 4 only introduced SMT to keep its pipeline busy every time it stalls. All of NetBurst's great ideas were introduced just to alleviate pipeline inefficiencies. But once they increased the pipeline to 31 stages (Prescott), not only it had very low IPC but extremely high power consumption and thermal dissipation as well. There was nothing they could do anymore to save NetBurst.
Most of the Pentium D's shortcomings were because of NetBurst, not really because of itself.
@@saricubra2867 Pentium 4 only introduced SMT to keep its pipeline busy every time it stalls. All of NetBurst's great ideas were introduced just to alleviate pipeline inefficiencies. But once they increased the pipeline to 31 stages (Prescott), not only it had very low IPC but extremely high power consumption and thermal dissipation as well. There was nothing they could do anymore to save NetBurst.
The AI must have forgotten about the 286 and 486 :(
Don't worry too much about the comments. The video's great. Hopefully you can do AMD cpus and gpus later on!
This whole video isn't the "evolution" of Intel CPUs but a low-effort, AI-assisted slideshow of the specs of their highest end SKUs, with some mistakes here and there.
- 4:05 millihertz
- 8:20 "i7 two-hundred-sixty K"
- leaving out information such as transistor count for numerous chips, even though it's a quick search on Google, e.g for the 8700K at 11:07
- lack of understanding of Intel's architecture reshashes, most notably Skylake being used up until they cmae up with Rocket Lake. It wasn't just 14nm, it was the exact same core architecture.
- usual uninformed joe's meaningless comparison of clock speeds across different architectures, you'd think niktek would know better. Well, apparently not.
- 13:24 E-cores have in fact been proven to be less efficient than P-cores. Reviewers view them as "cost-efficient cores" rather than the power efficient cores Intel claims they are.
- 8:08 First gen i5 and i3 models do in fact exist
- 6:15 no mention of P4 being a hot mess compared to P3. It pulled 2-3 times more power because of its inefficient design.
- 7:48 SMT already was a thing in Netburst (P4), it wasn't new at all with Nehalem
- 9:50 no mention of eDRAM
- 14:00 no mention of the recent Raptor Lake degradation fiasco
- 14:44 "This processor has less threads" - superficial and doesn't explain why, nor the consequences
Yes i do want to see history of AMD CPU
16.12.2024 - Niktek pivots to Tech Channel
what a year....
When you order Techquickie from Temu:
Can't believe you didn't touch any NetBurst drama ;-;
Also totally missed 8085 CPUs. It was big for 8 bit computers and educational devices.
Very underrated CPU btw.
don't care about the hate comments, the video is pretty good 👍
they killed the 'i' in proccessor names
You totally missed i186 (a microcontroller), i286 that was the 16 bit reference CPU for a long, long time for PCs at 16Mhz and 20Mhz, and I486 that also lasted several years and was much powerfull than 386 and introduced many features like instruction cache. Also, when you say "first 16 bits or 32 bits CPU" your are referring for INTEL CPUs, as higher bit cpus was manufactured from other comapanies for mainframes and industrial uses ( Alpha, SPARC, Motorola, etc)
Last generations were like several death blows for Intel in a row...
Thankfully I’m still rocking the good old stuff: I have an i7 8700! 😂 🤘🏻
*The Degradation of Intel
What many dont know the I7 5775C had L4 Cache the similar AMD had grown there Cache with X3D
And it was a flop because Skylake outperformed it by IPC alone and DDR4. Just like any Alder Lake i7 or i9K chip with DDR5 and is better than Zen 3 X3D
@@saricubra2867 What no the I7 5775C did even peform like Kabylake. Broadwell was bit under 50$ more and was on the older 14nm node. The first 14nm node that would follow for many Gens. It also came just few months before the 6th gen out so the Cpu was already in a situation that was rather wait for the next gen the new gen. That also why many people bought Skylake.
Alder lake was jump from AMDs win of the 5000. But in gaming AMD still had someting with the X3D. It was head to head. Ofc Productivity was Intel way better.
But as you can see now X3D is just way better and Intel cant do anything. Intel had a X3D CPU but it was not worked on to perfection.
@ Bro, Broadwell also falls behind Kaby Lake that increased clockspeeds.
The 5775C was a dumpsterfire of an architecture. Too expensive, clockspeed dropped so much that hurted it's overall perfomance because it was so hot.
"Alder lake was jump from AMDs win of the 5000. But in gaming AMD still had someting with the X3D. It was head to head"
What reality?, Alder Lake still has significantly higher IPC and the DDR5 advantage over Zen 3 X3D.
Reviewers were pairing high core count Alder Lake with @ss RAM vs AMD to artificially inflate the X3D scores (and AMD always does worse with memory scaling vs Intel). It's all a scam.
I play Sonic Unleashed and my 12700K is over 50% faster than a 5800X3D.
Zen 3 X3D is even worse than the 5775C.
It was overpriced and early X3D suffers from overheating problems.
Very overhyped as well, the 5800X3D has clockspeed jitter issues, it showed up in frametime graphs that for obvious reasons it was ignored because mainstream tech youtube is all about influencers.
"But as you can see now X3D is just way better and Intel cant do anything"
i9-13980HX is a raptor lake mobile CPU and has better power efficency than Zen 4 desktop X3D, faster than a 13700K, a reduction of power use above 100 watts.
Raptor Lake is basically IPC from 2021.
3 years latter and AMD doesn't beat Alder Lake IPC. We have Lunar Lake that competes with the Apple M3 series in perfomance per watt.
Then we get the 9700X, and the 12700K from 2021 still has 4% better IPC.
Zen 4 and Zen 5 are a flop, just like 13th gen and 14th gen. Just because the X3D are the ones that suck less doesn't mean that they are magically ahead of everything else. For example, the Ryzen 9 7900X is criminally underrated.
I watch how desktop CPUs are doing right now and are just boring compared to my 12700K.
Also why no one is rereviewing Arrow Lake right now?
When Alder Lake was new they gave it a pass with the horrible perfomance on the i9 (early adopter problems and optimization bugs) but Arrow Lake launches and basically nothing.
I'm sticking with my Intel Core i9-10900KF processor as long as I am using Intel Arc GPUs.
I don't trust Windows use of P & E cores, yet which is a shame because I see the benefits on my MacBook M3 Pro.
If anything, I might throw in the towel and make an entire AMD build when this one bottlenecks me hard on games.
NikTek finally took a break from smoking intel
You forgot that 7700K had thermal issues, because they used terrible thermal paste on IHS and it's not soldered
Great video! I am waiting one for AMD!
Also totally missed Itanium processor line, the 64 bit intel architecture that totally failed. The 64 bit architecture of today's processors (Intel included) came from AMD.
Intel maintained Itanium for 20 years due to a contract with HP. It would take them a week to talk about all of them, and all of the CPUs they did.
Remember the box on i9 use to be cool 11:30
Next amd please
Pentium Pro, 486 and 286 are missing
10 micrometers, well now we have 10 nanometer considered as outdated tech 😊 well still getting to 10 piqometers will take some long time 😁
I think you forgot to add intel i5th Gen cpus
Intel downfall was expected but it might not be what people think, I think part of it was their corporate dysfunction. When the market shift and they need to move fast the management seems so slow it was painful to watch.
That is the best summary of intel CPU I've ever seen
Huh? This is sub-par at best.
This whole video isn't the "evolution" of Intel CPUs but a low-effort, AI-assisted slideshow of the specs of their highest end SKUs, with some mistakes here and there.
- 4:05 millihertz
- 8:20 "i7 two-hundred-sixty K"
- leaving out information such as transistor count for numerous chips, even though it's a quick search on Google, e.g for the 8700K at 11:07
- lack of understanding of Intel's architecture reshashes, most notably Skylake being used up until they cmae up with Rocket Lake. It wasn't just 14nm, it was the exact same core architecture.
- usual uninformed joe's meaningless comparison of clock speeds across different architectures, you'd think niktek would know better. Well, apparently not.
- 13:24 E-cores have in fact been proven to be less efficient than P-cores. Reviewers view them as "cost-efficient cores" rather than the power efficient cores Intel claims they are.
- 8:08 First gen i5 and i3 models do in fact exist
- 6:15 no mention of P4 being a hot mess compared to P3. It pulled 2-3 times more power because of its inefficient design.
- 7:48 SMT already was a thing in Netburst (P4), it wasn't new at all with Nehalem
- 9:50 no mention of eDRAM
- 14:00 no mention of the recent Raptor Lake degradation fiasco
- 14:44 "This processor has less threads" - superficial and doesn't explain why, nor the consequences
5775C is the Intel greatest APU that nothing succeeded it...
5775C was a flop, Skylake came later and with the IPC alone defeated Broadwell.
@@saricubra2867 Skipping the IPC, nothing for PC has an OP iGPU like this. Also the massive 128MB cache just bring us the moment of X3D on Intel
mhz is megahertz, not milihertz
Is it your voice? Can you do AMD like this in detail.
Wow
I just watched an Intel 15 min publicity
Is this some sort of sarcasm about AI?
Do this kind of video for AMD
Core i7 980k 6 core 12 thread ....
3rd gen to 12th gen until AMD catched up and released probably the first 3D chip heading to 1nm.
While Intel remain stuck at 2D for 3 yrs I thought they would retain their 12++ nm fab just improve on design to save money doesn't want to buy a new fab for a decade.
im sure those TDP's are correct per generation too
You skipped the 80286
Damn
Video with the history of Ryzen. Why not with amd in general
Rip intel CPU