Where do I learn these stuff?

[Haswellx86]

After years of just consuming YouTube content, I have become a pretty decent tech enthusiast. Yet still there is room for being even more geeky. While my main domain is software, and I am already spending all my time in it, the only reason I got into hardware is because of LTT. Like seriously, I had zero interest in hardware before I starting watching. LMG unknowingly is one of my greatest motivations.

 

What I want to learn is even more geeky stuff, like the terms this guy uses in his videos. I don't have an extensive PC setup where I could just see and learn the individual knobs and dials in the motherboard UEFI. And not just that, but there are other things to learn that make you a true PC enthusiast and hardcore overclocker.

 

Of course, nobody just becomes a true tech enthusiast by just watching some videos. I think you need like an engineering degree, and a lot of experience, while I am still just very young, but I am not calling myself a true enthusiast already, just wanting to increase the technical knowledge I have gathered through just random sources.

 

The point of making this post is because, I don't know what I need to learn. Like what, do whatever it takes to become a CPU engineer at Intel and finally understand CPUs from the in and out? No, that's like actually becoming an engineer rather than just an enthusiast. That would be like actually becoming a doctor when all I needed is some here and there knowledge about the human body so I could take care of my health better.

 

How did actual tech enthusiasts, become tech enthusiasts? Who teaches you to lift power limits in your firmware for better performance? For me, just videos. But before even something like YouTube was a thing, from where did enthusiasts get the knowledge, or were there no such people and just engineers? I am just asking such things because I have no idea.

[wasab]

Enthusiasts are consumers. They are knowledgeable about the products they consume. You get this literally by just watching these youtube tech channels. what more knowledge do you need?

 

Quote

Who teaches you to lift power limits in your firmware for better performance?

This is like asking who can teach you to attach wings to your cars and fly it like an airplane. you do know they are not designed for this right? If it is so easy and have no downside and are just free performance, then hell, why would intel and amd set things like a clock limit? why would they limit their cpus and give an edge to their competitors? why cant they just go our CPU runs at infinity gigahertz and are the fastest CPU in the whole world!!! it beats all the competitors, we are the market performance king! muahahaha!!!!

 

You believe those chip engineers working at intel, amd, and laptop manufacturers who work on optimizing and performance fine-tuning their hardware as a full-time job are deliberately limiting and handicapping their cpus for no reason? there is always a trade-off. I say regular people should not overclock especially if your device is a phone, tablet, and laptop. desktop have more leeway and is officially supported by both intel and amd for overclocking but even that is more trouble than what is worth in my opinion. 

[saintlouisbagels]

32 minutes ago, Gat Pelsinger said:

The point of making this post is because, I don't know what I need to learn.

 

How did actual tech enthusiasts, become tech enthusiasts?

Your problem is that you want to learn things without having a goal. That's not really a good way to learn anything. Just wanting to "know more" is not a goal.

It's the same with learning a programming language or learning a piece of software like Photoshop. You need an explicit project first and then you branch off from there.

 

Tech enthusiasts on YouTube MUST learn this kind of stuff since they presumably want YouTube to become the primary job. Linus has like nearly 20 years of tech experience and he has a massive team of writers to help him sound smart. He has more experience than you have years of life. He has more resources than you will ever have (well, unless you also become a very financially successful person).

[RONOTHAN##]

13 minutes ago, Gat Pelsinger said:

Of course, nobody just becomes a true tech enthusiast by just watching some videos.

You'd be surprised. Most, if not all of my tech knowledge is self taught, either from YouTube, googling questions, or experience from using random hardware (good and bad). The best way in my mind to learn this sort of stuff is to just go head on into it, so watching more technical YouTubers (AHOC is a good one, so is GN, Skatterbencher is also decent, they're the first few that come to mind but there are plenty others) and every time you have a question pause the video, see if they address it, and if not get to Google to see if you can find the answer yourself. If not, get into the comments/forums/whereever and just start asking questions for one of the fine folks over there to explain it. 

 

20 minutes ago, Gat Pelsinger said:

I don't have an extensive PC setup where I could just see and learn the individual knobs and dials in the motherboard UEFI.

Look on the used market to see if you can find some older stuff for cheap to just use as learning tools if you want to learn things like this. Older X58 gear is generally great for this, especially since the CPUs cost nothing on eBay (you can literally buy X58 Xeons by the bucket) so you don't really have to worry about breaking them, and they're very unlocked platforms so all the knobs and dials are available to you.

 

22 minutes ago, Gat Pelsinger said:

But before even something like YouTube was a thing, from where did enthusiasts get the knowledge, or were there no such people and just engineers?

There were actual computer clubs and forums back in the day where you could learn things like this back in the 90s AFAIK. I'm too young to be in this crowd and give proper experience from these days, but I do know it existed and have seen stories about people heading to their local libraries with their computers and installing Linux from scratch with help from fellow enthusiests. YouTube has made the whole tech enthusiest scene waaaaaay more approachable and bigger though, plus PC gaming being a bigger thing helps. 

 

 

Basically, if you want to know something about a particular thing, just start googling, follow the rabbit hole and see where it takes you. 

[Haswellx86]

8 minutes ago, saintlouisbagels said:

Your problem is that you want to learn things without having a goal. That's not really a good way to learn anything. Just wanting to "know more" is not a goal.

It's the same with learning a programming language or learning a piece of software like Photoshop. You need an explicit project first and then you branch off from there.

No, I have a goal. It's in programming. But I am learning about hardware as well as just a side thing. And as I said, I don't want to become a whole chip engineer if I want to learn about hardware the fullest.

 

12 minutes ago, wasab said:

what more knowledge do you need?

Right now, I don't really know what stuff I should know as far as just being an enthusiast goes and not actually pursuing in that career. Where this point hit which made me write this post is the video I linked in the post. That guy is using terms like Load Line Calibration, and has a super geeky setup. I was with a face just like other non-techies have when they see me, so it shows I still have a long way to go, but for a career I am probably not going to purse, so might need to be in limit.

 

So other things aside, what mainly I am in the mood of learning right now is maybe most of the enthusiast knobs and dials in an enthusiast level motherboard. I don't even know what to look up, and I mostly got simple and clickbait videos. Look, I am no electrical engineer to exactly know about power or voltages or stuff. And probably, neither a lot of other enthusiasts who have gone ahead of me. I want to be in the enthusiast limit only and not want to actually become like an engineer of some sort.

 

It sometimes depends on my mood. For example, I recently started playing a flight simulator, and I started to learn a bit about aircrafts and flying them (because it was necessary to play the game). When I used to play a car simulator, I gained some knowledge about cars as well. And no, I am not going to be a mechanic or a pilot, just learning these stuff as just a side thing because of interest.

[wasab]

2 hours ago, Gat Pelsinger said:

It sometimes depends on my mood. For example, I recently started playing a flight simulator, and I started to learn a bit about aircrafts and flying them (because it was necessary to play the game). When I used to play a car simulator, I gained some knowledge about cars as well. And no, I am not going to be a mechanic or a pilot, just learning these stuff as just a side thing because of interest.

yeah, because these are for consumption by laymen. they are more for entertainment purposes which are what LTT and these tech YouTubers are. Anything more than layman entertainment becomes technical and boring really fast and you are likely to understand NONE of it, especially at your age and educational level.

 

consider a  documentary like this

 

if you are interested in physics, would you be interested in the above video documentary? pretty sure you would. this is laymen media, meant for laymen. it is entertainment in addition to surface-level education.

 

but how about this?

do you enjoy this video? This is what is beyond the content meant for a laymen. It has little entertainment value. it is all technical and meant for those who are interested in the fields at a "professional" level.

 

 

There is no such thing as wanting to learn "more" at a deeper level without stepping into the boring technical materials necessary at a deeper level. I know as a kid, you are not comfortable with those boring school like materials which sound more like classroom and homework than a video game like an aircraft simulator but that's what the advanced stuffs are. if you are not comfortable with it, you will just have to stay at the bottom surface level of what these tech YouTubers show you. 

 

Edit: I actually found a video by gamer nexus that explains load line calibration more thoroughly. Even I find this to be super boring and technical and this is still consider laymen level material. Will you find this to be enjoyable enough to do a classroom video watching to learn about the topics?  i doubt it. 

 

[RollyShed]

2 hours ago, Gat Pelsinger said:

How did actual tech enthusiasts, become tech enthusiasts?

By having a good tool kit. A good range of screwdrivers and spanners. Definitely a hammer (!!!!), pliers, sidecutters, soldering iron, power drill and drills and actually knowing how to use them. I started by building a one valve radio at age of about 12. This isn't something you will be doing but do you know anything about screws and the more than a couple of dozen types there are? Can you drill a hole and fit a screw? Know what type to use?

 

NOTE - I presume you don't know what a valve is. They were used in the very original computers as well as radios and TVs.

 

That's all REAL hardware knowledge.

 

Then you apply the software knowledge on top of that.

 

Then you get a grey beard and your hair falls out.....

[Kilrah]

3 hours ago, Gat Pelsinger said:

How did actual tech enthusiasts, become tech enthusiasts?

With time and experience. People who know "what dial to turn" in overclocking videos do because they've been reading, discussing the subject and poking at CPUs on their own for a decade and have built a "grasp" of how things tend to be, then they'll know where to look on a newer thing.

 

3 hours ago, Gat Pelsinger said:

I am no electrical engineer to exactly know about power or voltages or stuff.

Then learn at least the basics of that. When you want to understand something and read about it you'll inevitably fall onto other subjects you need some knowledge of to understand the original thing, so you'll have to branch into that for a while. Over time you'll know enough about many subjects to be able to figure things out. 

 

The keys are being open enough to delving into a variety of different subjects, and time. 

 

 

[ImorallySourcedElectrons]

2 hours ago, Kilrah said:

Then learn at least the basics of that. When you want to understand something and read about it you'll inevitably fall onto other subjects you need some knowledge of to understand the original thing, so you'll have to branch into that for a while. Over time you'll know enough about many subjects to be able to figure things out. 

 

The keys are being open enough to delving into a variety of different subjects, and time. 

 

 

The crux of the issue is that even a lot of the books, even for folks who major in electronics, only cover the absolute basics of topics like CPU design. They basically give up somewhere around the early to mid 90s because it gets quite complicated quite quickly. Like they'll talk about the very basic concept of micro-code and they'll walk you through a simple example of how you decode an instruction and generate the right control signals for different parts of the processor, but rarely do the exercises go further than that in most books (there are exceptions!) Then they'll add on things like pipelining and out-of-order execution conceptually and illustrate it with diagrams, but they won't actually walk you through how to implement these things. Maybe they'll go into some detail on how SIMD might be implemented at a hardware level in the ALU because it's quite sensible and straight forward for mathematical operations. But good luck finding a book that will tell you about things like dynamic clock generation, on-die temperature sensing, practical approaches to speculative execution, the reason why cryptographic functions are implemented the way they are, how to design and build a DMA controller and schedule for data to arrive just in time for when it's needed, strategies to (de)serialize data coming in at clock speeds higher than your CPU core is probably capable of running, when to decide when to turn off parts of a chip, floorplan optimization to avoid bottlenecks, how to handle crossing clock domains in conjunction with dynamic clocking, practical approaches to on-die voltage regulation in modern cmos technologies, etc. And then we haven't even gotten into things like the chip packaging.

 

Which is to say, even if you are an engineer working in the electronics industry, good luck knowing the details on all of these things. These devices are products of huge teams, and not even the system architects will know all the details. The difference of most folks with a relevant engineering background is that they know enough about the base topics to quickly learn enough to be dangerous if they run into the challenge. But at that point you indeed consult extremely boring books, talk to colleagues who will point you towards internal documentation or journal articles, or just the previous design. And it's not uncommon that the same problem has to be solved again and again as a result of all of this.

 

So how do enthusiasts get into it? A lot of them are professionals who dabble in these things outside of work hours, because their job doesn't really allow them to do the fun parts that initially got them interested in the subject in the first place. And they'll often release snippets of knowledge, which are then picked up by the community. For example, some of the best explanations I've seen about how to approach chip floorplans can be found on reddit, and they were clearly written by someone with decades of experience, but odds are that they just wrote about it because they find it genuinely interesting. 

 

And once you go outside of the realm of engineers outside of work hours, things get hazy to put it mildly, a lot of times the enthusiasts are often just "wrong". A good example is Gamers Nexus's authoritative blabbering about industrial supply chains, they literally said something along the lines of "it's unheard of that parts change frequently during a production run" during their ranting about Gigabyte. Meanwhile, I had colleagues who at that same point in time apparently had to approve component swaps on an almost daily basis because we could only get a few reels of each component due to parts shortages. Several folks in the comments were pointing out these issues with their statements, but they forged on ahead anyway. So it's not because they seem knowledgeable that they in fact know what they're talking about.

[Haswellx86]

@ImorallySourcedElectrons

 

Chip manufacturers are quite conservative when it comes to opening their knowledge and research to the people. If somebody wants to learn about these stuff, then their best shot is RISC-V or a major part of ARM.

 

As @wasab said, there is a line between entertaining and interesting knowledge, and actually studying it. The deeper you want to learn, the less people there are who themselves have such knowledge, the topic itself is too complicated, and this just gets pretty involved so that you need to find the way yourself and get full knowledge in it, rather than just seeing some videos.

 

But, I don't want to maybe become a chip engineer with just here and there knowledge (if that's even possible). I want to do and learn what just enthusiasts and done and gone through. And the GN video wasab gave is probably what I was looking for. Still would like a better source and with a guy who is more enthusiastic than GN from 6 years ago (oof).

[ImorallySourcedElectrons]

3 minutes ago, Gat Pelsinger said:

Chip manufacturers are quite conservative when it comes to opening their knowledge and research to the people.

I disagree, you need to go to more conferences and spend time in the exhibition hall or at the bar instead of attending presentations. Plenty of interesting conversations to be had, and folks within the industry are quite open with the challenges they face and how they solved them. The electronics industry is relatively small, and within a particular region most folks do tend to know of each other or have met at least a couple of times. Attending conferences is also how you end up in sketchy mountain villages with a bunch of drunk electronics professors and the largest bar tab I've ever seen, but that's a subject for a different time.

 

33 minutes ago, Gat Pelsinger said:

If somebody wants to learn about these stuff, then their best shot is RISC-V or a major part of ARM.

Let's say you get a pile of VHDL or Verilog for an ARM IP core, you're not going to learn how to design a CPU from that. Instead, you'll find one particular design with a whole lot of choices already made optimized for a specific target (e.g., a particular type of FPGA), but that will not teach you how to design when targeting actual silicon yourself. It's really a matter of knowing how to find the right resources for what you're hoping to implement. I've done small-sized things myself, like trying to reimplement an early i386 and seeing if I could generate the full mask set for a technology we had in-house, but the reality is that you very quickly notice that there are giant gaps in your knowledge and that you need to sit down and have a good long think about how to implement these things, or go and get some help. I was fortunate enough to have plenty of access to the latter over multiple long lunchbreaks, but it's really not a given to just look at what someone else did and base your knowledge on that. It took running into the wall multiple times, realising I messed up, and then getting the help I needed.

 

24 minutes ago, Gat Pelsinger said:

But, I don't want to maybe become a chip engineer with just here and there knowledge (if that's even possible). I want to do and learn what just enthusiasts and done and gone through. And the GN video wasab gave is probably what I was looking for. Still would like a better source and with a guy who is more enthusiastic than GN from 6 years ago (oof).

Check der8bauer's videos:

 

[Haswellx86]

8 minutes ago, ImorallySourcedElectrons said:

I disagree, you need to go to more conferences and spend time in the exhibition hall

Well I have never done that, but I have heard a lot that chip manufacturers are not very open about their research. Well yes, those videos by Intel teaching you branch prediction is something that you might call as "being open", that is something really basic about modern processors. If you go to learn chip manufacturing (and like actually hardware and not software side) in actual like deep level where you would actually achieve something, there is probably not much help. That is as far as I think. Mostly what you learn about chip manufacturing would be from open source documents.

13 minutes ago, ImorallySourcedElectrons said:

Let's say you get a pile of VHDL or Verilog for an ARM IP core, you're not going to learn how to design a CPU from that.

Of course not, but at least you will get the whole CPU architecture in front of you. Chip manufacturing is a different thing which I don't know about.

 

14 minutes ago, ImorallySourcedElectrons said:

Check der8bauer's videos:

Thanks for suggesting but I have already watched them. Enthusiasts should go more with engineers. That makes truly a great video.

[ImorallySourcedElectrons]

13 minutes ago, Gat Pelsinger said:

Well I have never done that, but I have heard a lot that chip manufacturers are not very open about their research. 

In which sense? They're quite open I would say, they won't go into implementation specifics but I've seen plenty of open discussions about major architectural changes and process technology.

20 minutes ago, Gat Pelsinger said:

Well yes, those videos by Intel teaching you branch prediction is something that you might call as "being open", that is something really basic about modern processors.

That's my grievance with videos like that, they only cover the basics, that's like reading "Structured Computer Organization" by Tanenbaum, it's not a bad book but it won't really teach you how to implement it.

21 minutes ago, Gat Pelsinger said:

 If you go to learn chip manufacturing (and like actually hardware and not software side) in actual like deep level where you would actually achieve something, there is probably not much help.

There is plenty of help with that, just not neatly summarized in books.

22 minutes ago, Gat Pelsinger said:

Mostly what you learn about chip manufacturing would be from open source documents.

What do you understand as chip manufacturing? In terms of process technology, some PDKs go into excruciating depth on the technology side, in other cases you can find plenty of papers and conference presentations. No one really bothers to hide it, because even if you know enough of the process flow to setup a simulation of the fabricated devices (as in transistors/gates) and generate good device models, you won't be able to replicate the technology.

 

In terms of chip design, there are loads of good books on the subject, but they're quite dry. There's also plenty of good educational resources, check the electronics engineering department of your local university. Quite often the course slides are online, and they go into plenty of depth on both the design and manufacturing aspects. Meanwhile, most online summaries never really go beyond schematic diagrams.

 

28 minutes ago, Gat Pelsinger said:

Of course not, but at least you will get the whole CPU architecture in front of you. Chip manufacturing is a different thing which I don't know about.

Which you probably already had from the public datasheet  

[porina]

I think the nvidia crash thing was early 30 series, 3080 and 3090. Memory might be wrong but people started looking at caps used for power smoothing on GPUs and some media decided some were good or bad caps.

 

On the original topic, how I got where I am is a mix of a few things. I liked physics at school. I did Electronic Engineering at university and most of my working life. On the PC side, if it ain't broke, tweak it. You learn a lot through overclocking. I mean old school manual overclocking when you directly impact voltage and clocks though various methods. Even if you never get to a low level understanding, at least you get used to the relationships between various factors.

 

If you really want to go deep, the resources are out there. The best Zen 5 "review" is not from Youtubers complaining about gaming performance. Try something like this. Or look up Chips and Cheese. I personally don't find video to be great at going in depth at anything, and written articles help there. However one to look at is the Hot Chips conference. Not everything is made open but a lot of the presentations are. It's at the end of the month and I just realised I'll miss it live as I'll be travelling.

[ImorallySourcedElectrons]

1 hour ago, porina said:

I think the nvidia crash thing was early 30 series, 3080 and 3090. Memory might be wrong but people started looking at caps used for power smoothing on GPUs and some media decided some were good or bad caps.

Yeah, the decoupling on some of those early third-party GPUs was bad, but this was also during the peak of the parts shortage, so they were probably trying to minimize capacitor usage.

[ImorallySourcedElectrons]

3 hours ago, wasab said:

why would they create a "false" shortage? this hurts their bottom line and profit. the ideal price is where demand meets supply, this is where the max profit is at, not when price is so high that few people will buy it leading to unsold inventory and hence excess supply or when the price is so low that they sold out on everything but make little profit or even worse a loss.

He has no clue what he's talking about, doesn't even understand bandwidth

[wasab]

2 hours ago, ImorallySourcedElectrons said:

He has no clue what he's talking about, doesn't even understand bandwidth

Yeah. I mean it is economics 101. Companies will always want to produce just enough at a price that all of these "enoughs" will sell. Producing too little that there is a shortage means a loss of profits even if price is high because you have less inventories to sell. Producing too much means you can sell more but lead to oversupply which means lower prices if you wanna sell them all. 

[TVwazhere]

-Thread cleaned-

[Sauron]

On 8/11/2024 at 9:07 AM, Gat Pelsinger said:

Of course, nobody just becomes a true tech enthusiast by just watching some videos. I think you need like an engineering degree, and a lot of experience, while I am still just very young, but I am not calling myself a true enthusiast already, just wanting to increase the technical knowledge I have gathered through just random sources.

I have an engineering degree, they don't teach you anything about mainstream consumer hardware... and neither should they because this stuff changes every few months and unless you keep yourself informed your knowledge will become outdated immediately. Everything I know about consumer hardware has come from watching tech videos and participating in this forum, with the occasional article or tutorial thrown in. In fact I know considerably less now than I did 5 years ago because I'm not following the industry much anymore and the landscape has changed since I stopped. I used to know all available desktop cpu and gpu models and their rough price point, now I don't even know off the top of my head how many cores the high end models have. Of course I could easily get back into it by simply watching some videos on the topic and checking out the current listings on pcpartpicker.

 

Linus doesn't have a degree, he just worked at a tech retailer for years and was personally interested in this stuff so he absorbed that information - you can't build hundreds of PCs without eventually learning a thing or two about them.

[Agall]

On 8/11/2024 at 1:07 AM, Gat Pelsinger said:

How did actual tech enthusiasts, become tech enthusiasts?

The hardest part about learning tech is sifting through the misinformation, but that's just like anything on the internet. You also won't know what's true and what's not unless you're challenging everything, which mostly involves cross checking it or validating it yourself (and hopefully not validating it with the same sources as everyone else...). So, always having an open mind and re-evaluating any information you come across or put out yourself.

 

On 8/11/2024 at 1:40 AM, RONOTHAN## said:

YouTube has made the whole tech enthusiest scene waaaaaay more approachable and bigger though, plus PC gaming being a bigger thing helps. 

My initial jump was seeded in by Youtube and LTT specifically and revolved around gaming initially, at least since 2010. 

 

20 hours ago, Gat Pelsinger said:

Chip manufacturing is a different thing which I don't know about.

 

Thanks for suggesting but I have already watched them. Enthusiasts should go more with engineers. That makes truly a great video.

 

There's tiering to knowledge and a ton that simply isn't required to meet the arbitrary 'tech enthusiast' threshold. The deeper you go, the more advanced your fundamental level of science and engineering has to be.

 

[Haswellx86]

34 minutes ago, Sauron said:

I have an engineering degree, they don't teach you anything about mainstream consumer hardware...

Not like that, I mean I know there is no tech enthusiast course or anything, and you yourself have to get into it, but having like an electrical engineering degree, should probably help you in your enthusiast journey. I mean it just like I said, you need more experience into this, and no specifically any degree or course because there is none for being an enthusiast, so just other external things help.

 

38 minutes ago, Sauron said:

Linus doesn't have a degree

Wait really? Would like to hear more. That makes him an even more motivating figure.

[Sauron]

10 minutes ago, Gat Pelsinger said:

Not like that, I mean I know there is no tech enthusiast course or anything, and you yourself have to get into it, but having like an electrical engineering degree, should probably help you in your enthusiast journey. I mean it just like I said, you need more experience into this, and no specifically any degree or course because there is none for being an enthusiast, so just other external things help.

I'm telling you - an engineering degree makes you an engineer, not a consumer tech enthusiast. Neither is a prerequisite for the other.

12 minutes ago, Gat Pelsinger said:

Wait really? Would like to hear more. That makes him an even more motivating figure.

What more do you want to know? He just doesn't have one, at least as far as I know. I think I remember him saying he started one but dropped out, and I believe it wasn't even engineering. If you want his biography you should ask him, you can tag him or message him here you know.

[Kilrah]

You don't do a 3/5 year electronics engineering degree to get a grasp of consumer tech, you do so if you want to be the one designing complex commercial appliances... Sure if you happen to be one already it'll help understand other things and there are many engineers around doing and publishing things for fun in their free time that are more advanced than usual by leveraging their knowledge, but you don't need that to do or understand things.

 

90% of people who do cool stuff you'll find on hackaday aren't electronics engineers, they'll have been interested in things they've probably seen randomly, looked them up, figured them out, found something they want to do, tried and failed a few times to make it and eventually succeeded.

Look at the "maker" communities for inspiration.

 

More than ever almost everything in the tech world but the most advanced stuff is within reach of anyone both in terms of information and access to the actual hardware these days if you put in some effort and can stay interested long enough to absorb what's needed.

Sure you can't fab your own 4nm silicon but you can build your own CPU from scratch with logic chips if you want to understand the basics of one, and there are plenty of resources guiding you towards it etc. 

[monopoly_hater]

Basically become a electrical engineer.