You are now in charge of making cpu installation foolproof. How would you do it?
>remove fools
>????
>profit
>>54801702
no installation is foolproof
I would pre-install it tho
but faggots would still destroy it while thermal pasting n shit
https://en.wikipedia.org/wiki/Slot_1
Solder it to the motherboard in the factory.
Have three options
Gaming Motherboard
Office Motherboard
Budget Motherboard
>>54801736
This. Selling mobos preinstalled with a CPU is just about the only thing i can think of. I don't think you'll be able to use this idea to get rich though. Sorry anon.
>>54801769
>no enthusiast motherboards
>>54801795
>no sport motherboards
>no ultra special extreme gaaaaaaaameeee MLG pro
>no Music love motherboard
>no Imotherbooord from Apple TM INC ORG NET
Fun fact: I fucked up my first CPU installation and it was an AMD CPU... bent all the needles, broke the thing off that literally just allows you to snap it in... I don't even recall how I fucked it up so bad. I believe it had something to do with me watching a youtube clip of someone installing an intel CPU.
Sad day, money down the drain.
>>54801830
enthusiasts are actually thing unlike the meme shit you're trying to pull.
>>54801769
>>54801788
Yeah, I don't think those are profitable as well. So many unused motherboard+cpu combinations will be junked
>>54801819
Why don't we have more of this?
cpu applicator that comes attached to cpu, a plastic device with detailed instructions that basically tell you to
1) line it up with big fucking arrows
2) place pins in motherboard heatsink screw hoes
3) press down
4) press button to release cpu from applicator
tada, cpu in motherboard
even better, it would have thermal paste already applied
when its in the motherboard the button would be pressed to detach the cpu and paste from the applicator
then you can just stick a heatsink on it and be done
hands wouldnt even have to touch the cpu or mobo
>>54801856
Implying your opinion on this matter matters to matter .
get back to /v gaymer
>Cover the top of the Cpu with some plastic closing that you can grab on,
>put on motherboard a 1 cm tunnel the shape of Cpu
>drop the Cpu there
>the CPU top plasticcover acts as closing top panel.
Literally 10 second idea.
>>54801702
It kind of is already. You put the CPU in the socket, put a drop of thermal paste on top of it and mount the cooler. Hell, even if you completely forget the thermal paste your PC won't exactly explode. Everything is already dumbed down for the stupid impressionable kids whose parents buy the crap. All you need are working fingers and the ability to operate a screwdriver.
>>54801856
http://www.neoseeker.com/Articles/Hardware/Reviews/aopenax4btube/
Meme you say?
>>54801893
>comparing enthusiasts with geymers
did i poke a hole in your little meme-bubble?
>>54801915
there's no need for improved audio on the motherboard because there are audio cards.
CPUs are already idiot proof. The wafer is notched so you can only put it in one way: the way that works.
>>54801875
>Why don't we have more of this?
Because there's not much of a market in helping people put the square peg into the square hole. Most of us know not to try and use the round peg.
>>54801875
The mobo I just bought,asus z170a, came with one of these. It was terrible. The CPU didn't want to stay in and after popping it in (the directions literally say "push until it clicks") the CPU still fell out onto the socket as I was positioning it. I nearly shat my pants. I tried not using it but the socket didn't seem to accept it sans the bullshit plastic guide thing. By the time I was done both sides of the CPU were covering in fingerprints.
>Inb4 retard sounds like you're just too stupid to assemble a pc
I have assembled quite a few and have NEVER had an issue with seating a CPU until this bullshit.
When will the pins bend? During clamping an incorrectly seated processor or when placing the cpu the wrong way?
Put a manual in the box with it and state that the warranty dose not cover physical damages to the CPU and Motherboard and that the company cant be held responsible for any damages caused by misuse.
Never give it a second thought for the rest of my life.
Really things don't NEED to be foolproof, I just have to make sure I cant be held accountable
>>54801702
Impossible.
As soon as you produce an idiot-proof product, nature will simply produce more robust idiots.
>>54801702
All CPUs now have to be connected through USB.
>>54801702
Two words: Wireless CPU
>>54802045
Working as a code monkey, every bit of the application should be fool proof and shouldn't even allow an invalid operation, because users are idiots, or so I've been taught. I think this can be applicable with hardware as well; an improperly seated cpu should not be allowed to be clamped down.
>>54802129
two words: bandwidth latency
>>54801910
Does the plastic cover come off once you install the cpu?
Have it installed like a game cartridge, people seem to be able to do that without fucking everything up most of the time.
>>54802143
LiFi CPU's when?
>>54801837
Quoting me, is this more common than I think or am I just a retard? Had no problem with the cooling fan.
What about a CPU that slides in like those Cable cards.
>>54801702
The CPU comes with a C4 charge. If you fail to install it correctly it will detonate. No idiot will live to tell he fail installing it.
>>54802169
Not necessary,and doesnt have to be plastic,maybe other material better at cooling.
The idea is to have the mounting point on the motherboard a tunell at least 1 cm long.
Magnetize the CPU's. No pins. Then Pins can't bend. Also, one side of the CPU is neon orange so you know where to face up (Arrow in corner is too difficult for autists)
>>54801938
Falling for the audio card scheme, holy shit.
Take the pins off the CPU, make it a pressure contact to the mobo.
>>54802182
>lelfi would solve any of these
but hold on, instead of using lga we could use a fiber optic array. transceiver on the prebuilt cpu block and receivers on the mobo. just make sure they are aligned with some simplest form of clamping and we're good to go.
>>54802175
>ya gotta blow on it first Steve!
>saliva shorts two connectors
>turns on new pc
>magic smoke comes out
>>54801728
/thread
>>54802244
>doesnt know about onboard interference
holy shit
>>54801746
>https://en.wikipedia.org/wiki/Slot_1
This tbf, one of the only slots a retard would probably not fuck up. Unless they tried installing it into a PCI slot...
>>54801887
>even better, it would have thermal paste already applied
But muh Arctic Silver
>>54802444
good luck adding a heatsink without bending it to fuck
>>54802462
overrated
>>54802175
old school is best school
see >>54801746
>>54802488
>overrated
B-but it's got FOUR different shapes of micronized silver
The only hard thing is applying thermal paste, you either put to much or to little. And removing it can be hard
>>54801746
>>54802497
the only thing that sucks about this is that modern CPUs are so power hungry and housefirey that you would only be able to do this to mobile-tier shit without having to use liquid cooling
>>54802540
no it's not. a small pear or large grain of rice is not a fucking hard size to estimate.
>>54801746
Why don't CPUs use PCI-E ports any ways? Is it just that the CPU needs to be close to the RAM or is the whole gorillion pins concept simply requires because of reasons?
![cook_hero[1].png](https://i.imgur.com/cekygZW.png "cook_hero[1].png")
>>54801702
solder it to the motherboard along with all of the other components
>>54802647
>>54802628
The PCIE lanes are on the CPU.
>>54802608
Cleaning it off can be a bitch
>>54802647
Isn't that what prebuilts are? :^)
>>54802700
only if you're a retarded faggot.
>>54802357
so equip the cartridges with spring-autoclose plastic covers, like there was on VHS
>>54801914
Intel CPUs and coolers even come with preapplied thermal paste. It's rubbish but good enough for an i3 and relatively foolproof. Not sure how you could make it simpler.
>>54802728
What is the most foolproof way
>>54802628
PCI-e is a slower bus compared to the FSB of a CPU if im not mistaken
>>54802175
we had something like that it was called "slot 1" it was great but only for the time when you didnt had a giant cooler hanging on it
>>54803267
>look mom I posted without reading the thread again
>>54802608
>a small pear
Hello, MSI.
>>54803231
Just as a sanity check to point out the retardation of this idea: The CPU has to host 40 PCI-e lanes. Are you saying you would host 40 PCI-e lanes using the 16 lanes of a single slot? That's not even counting the interface to the RAM...
>remove pins from CPU
>make a ball grid array socket on the motherboard
>CPU slides in from the top, touching the ball pins is harmless
>screwing the socket down puts the pressure on the ball grid, keeping it from shifting
>CPU is spring loaded so removal is easy as unscrewing
Was that so fucking hard intel? You literally have these tester type sockets in the fab house, I should know, I work there. You think by 20 fucking 16 they would find a way to get rid of bendable physical pins.
>>54803319
>>54801978
I have the same motherboard and the plastic bracket worked flawlessly for me, very easy.
Don't know where you were faulty, bro.
>>54802628
I suspect if Desktop PCs weren't growing more and more irrelevant, that would be the next step. Instead of having multiple different ports for different things, everything can be put into one multi-purpose slot. Pic related mainboard in the back and RAM, CPU and GPU go in orthogonally. The air can be provided like the arrow shows by a single fan. It's actually a lot like the mac pro, just with a little more accessibility.
>>54802605
That is why in the new versions we build we do something clever, like mounting screws that are hollow so coolant can flow through it from piping built into the motherboard which also acts as a sizable heat spreader.
Bonus points if both sides of the motherboard are uses to better separate components and add more cooling fins.
Honesty I could go on, but at some point it we have to start deciding between modular vs monolithic designs.
Still having a main board and 4 slot types for CPU, GPU, RAM and Other would be simple enough and be a sensible middle ground.
But really those slots should be designed by bandwidth and latency, not application, to give more options. Like how CPU and GPU are both PU but have diverged in many major traits to the point they are different.
>>54801978
You could have just not used it. It's purely optional. I have a Maximus VIII Hero Alpha that came with one and it worked perfectly for me.
>>54801978
I have the same mobo and this was the first computer I ever built and I got getting the CPU in correctly right on my first try. My only problem was how much I cringed putting the hinge back down due to have it felt traveling.
Literally all you have to do is open the socket, rest the CPU in, and close. How could you fuck that up to the point where shit was falling out?
When I installed my i5 2500k it was pretty fucking fool proof, there aren't any needles to mess up so you just drop the thing in and lock it in place and you're done.
It even comes pre-greased so you just click on the cooler and that's it.
I say just make everything all-in-one. Why do we need our memory, CPU, and motherboard separate at all? In the past, things were different, but today, there's almost no point to removable CPUs anymore since newer processors are very incremental upgrades and require board revisions to make use of new features anyway. The only situation where it would be useful is if you burn out your CPU and the board still works; but even then, most people will use that as an excuse to build a whole new system and the old board will be disregarded. You just new to release several different models, for example:
Weak processor, 8GB ram=Cheapo tier
Average processor, 16GB ram=Average tier
Strong proccessor, 32GB ram=Enthusiast tier
Weak processor, low energy consumption, 32GB ECC ram=Home server tier
>>54805284
>I say just make everything all-in-one
>One thing breaks
>Entire thing is fucked
>>54805284
apple posters please leave
imagine if cars were like this
>oops flat tire
>time to pull out 20,000$ and buy a new car
i mean tire upgrades are very incremental right? people shouldnt be allowed to replace bad ones without buying a new car
>>54802306
But then it's not wireless.
let the mainboard have a soldered processor
attach cpu+ram as a SOC via usb type c / pcie / whatever.
emulate x86 via software.
we do that shit pretty much with graphics cards for bitcoin mining and i hope this happens with ARM for cpus too
take what we learned from cluster server computing and lets put it in box-hardware.
would love to see the linux kernel patches compiler for this, would be fun
>>54805400
If the memory and CPU were built into the board, they could likely afford to make it cheaper than having to make separate parts and developing connection interfaces and so on. I've never had a CPU go on me in all the years that I've had a computer. I've had a couple of memory modules go, but that's usually after they've been in service long past their expected lifetime. Usually if the CPU is going to go, it'll go within the warranty period anyway, or you did something stupid with it like fucking around too much with overclocking.
>>54805511
CPU, memory, and motherboard do not have movable parts like a car, nor do they handle explosive gases or corrosive liquids.
>>54805644
memory is probably the most upgraded/replaced part on a pc you fucker
also iv had a motherboard go bad on me and what a fucking relief it was to replace just that instead of dropping $1000
iv had the same pc for 8 years and its still fine from getting well priced upgrades for it over the years. almost everything has be replaced, including the cpu. i think the case and psu are the only original parts
literally the problem with desktops right now is that they are too good and convenient for customers so they rarely need to buy new ones
>>54801702
The biggest problem is just getting the pins lined up correctly, with the CPU in the right orientation, and being able to place it down flat, without bending of breaking the pins.
> bending/breaking pins
Have the pins on the CPU surrounded by weighted telescopic cylanders (or even just a big rectangular one with holes for the pins), such that if you're holding it, it drops and covers up the pins, protecting them from damage; when you push it down in the right area, this telescopic covering gets pushed up, exposing the pins.
> Orientation
Imagine that you insert the CPU into a slot: I'd just make the CPU slightly rectangular so that it can only fit one way, then put little bits on the edges of one side so that it can't go in that way. By doing this, the CPU can only be inserted into the slot in one direction
> Lined up correctly, placed down flat
Make the slot mounted on a swiveling parallelogram type platform, so taht you just push it over and it lines up exactly with the motherboard platform, which pushes against the telescopic pin cover and allows the CPU pins to make contact.
Will soldered HBA maybe force a reintroduction of a slot 1 type connector or does it all fit on a modern sized CPU?
![MotherBoard_large[1].jpg](https://i.imgur.com/2qwUZuD.jpg "MotherBoard_large[1].jpg")
>>54803599
>we early 90's now
>>54801702
Solder it to the motherboard.
>>54805605
bandwidth
Here you guys go <3
>>54801702enforge LGA or better contact pin
Make cpu in a trapezoidal shape
Pre paste every heatsink and ban tubes of laste, if you nisplace it you have to clean with rubbing alcool and get a new heatsink
For no reason at all, anyone should apply anything on the heat spreader, not sticker nor protective films
>>54801702
>>54806103
>>54803599
There are actually computers back then which had CPU, memory and so on on card, and a simple connector board for the other cards
https://en.wikipedia.org/wiki/Single-board_computer
>>54802628
Okay a GPU is basically a chip on a motherboard that goes in another motherboard, real question is why can't we have motherboards that natively thake graphics chips, installing killer cooling would be simpler too
>>54807144
Because it would cost more for people who didn't need a discrete GPU. You'd have to add the power circuitry, and the VRAM sockets to the motherboard. It would probably double the cost of a motherboard. Besides, do you think that NVIDIA and AMD would ever agree to a standard for GPU sockets?
>>54801702
It is illegal to install your own CPU. Only authorized retailers, re-sellers and professionals are permitted to install the CPU on the motherboard.
>>54801702
Use magnets
>>54807185
Got me there but in retaliation I'll. Say it would look cooler to have two CPU coolers in you case
>>54801702
Host it on the cloud.
>>54801702
Solder it permanently.
>>54801702
Don't let monkeys do it.
>>54801856
>enthusiasts are actually thing
they waste their money on shit they dont need because their parent never taught them impulse control and/or money management.
They fall into the consumeristic bullshit memes and throw away culture that you live in.
In essence, good goy!
ps, this type of behavior shouldn't be encouraged.
>>54801702
The pins are on the board, just normal
But then the pins are surrounded by a rubber encasing type thing, put chip on top of rubber, apply pressure, rubber compresses, pins touch chip
This way you won't be poking the pins
>>54807337
Your a poorfag if you can't buy the 6th Generation Intel® Core™ unlocked processor
>>54807407
Also make the assymetry obvious to deter retards from trying to force it the wrong way, the tiny triangle/arrow is not enough
>>54807416
exactly..
>>54801910
>bent pins during insertion
>>54801702
Didn't it used to be plug in cards?
Why not go back to that.
Outsource socket design to Lego.
>>54801702
Integrate the cpu into a wire. That's right, a wire. With a plug on the end. A reversible plug.
>>54801702
>You are now in charge of making cpu installation foolproof.
stop selling PC components to people who haven't demonstrated the capacity to build a lego house (age 8 and up)
I cannot fathom how a CPU installation could be fucked up other than bending pins, and the last three processors I've bought and installed didn't even have pins on them.
Unless you're that guy from the cringe thread who puts thermal paste on the wrong side, but other than that it's literally
>Place square thing into matching square socket.
>Pull lever down.
>>54807570
Because anything above ~1GHz signal noise becomes a huge problem
headphone-jack plugin for CPU
one line is power with attenuated phase-shift input bits, the other line is 1's and 0's in a stream to ram controller
How about multiple heatshields on the bottom. They encase the pins and me contact with the heatshields. The motherboard socket has the same heatshields layout and several cutouts that insure you can only put it in the right way. There it'll never have another bent pin. And the CPU is now more thermally manageable. Along with being indestructible.
I managed to explain the process to a private over radio in about 15 seconds.
If it's gruntproof, it's foolproof enough.
>manufacture all in one computers
>turn them into glorified remote client computers
>have them to connect to a single powerful server
>have them pay performance per hour like use i7 $5 per hour
>Problem solved
![intel-xeon-phi-card[1].jpg](https://i.imgur.com/XQKcC53.jpg "intel-xeon-phi-card[1].jpg")
CPUs now use the form factor of GPUs with similar fans and plug into a slot of the board.
Some reduction in performance, but hey you wanted idiot proof and you got it.
>>54807144
>real question is why can't we have motherboards that natively thake graphics chips
Some time down the road this will have to be the way for GPUs to go if they want to continue the power creep for the next 10 years. Slots can only do so much.
>>54803397
Sounds expensive to QC. Also not sure the balls would have a good enough connection for high performance.
>>54810991
So this basically
>>54811015
>CPUs stagnate so much that they're soldered on motherboards
>GPU socket now on motherboard
>GPU swapping and cooling becomes the same as CPUs today
The sad thing is I can see it actually happening
>>54810991
This puppy costs more than nVidia Tesla.
>>54811058
A really beefed up version of it, yes.
>>54810991
Do I need an x86 system to take advantage of one of these or is there some open source driver?
>>54811094
It happened with sound cards.
>>54801736
I just read a guy selling his "gaming PC" and it said in the description that he changed the thermal paste 4 times a year.
>>54801702
balls instead of pins
Only sell soldered CPU's
No installation means no installation errors :^)
>>54801702
Every CPU now comes with a container that has a 16 digit analogue combination lock and a small device with pre-packaged "cpu installation simulation" software on it. The only way to get the combination is to successfully install the CPU in the simulation twice in a row without fucking up.
>>54801769
> Implies Gaming is the most intensive process possible.
What about scientific simulation?
>>54811188
I have no idea. These things are sort of an enigma. Mostly all you ever hear is they are for "serious scientific research", whatever that means.
>>54810991
We'll just invent a better idiot.
>>54811464
Damn, I want one but I don't want to get it just to find out I need another computer
Make it non symmetrical like a semicircle or something, you would have to be a dumb ass.
The cpu has holes at the bottom.
The pins get pushed up from the socket as the cpu is inserted correctly.
The user should not have any way to damage the contact points on the cpu or the socket.
>>54801702
Guiding rails on the socket and matching holes in the CPU's PCB. Slide the CPU down the rails to the motherboard so it lines up perfectly every time.
>>54801702
The interface is circular, with alternating concentric ring contacts on both the CPU and mobo, so that you can't possibly misalign it.
>>54801702
>You are now in charge of making cpu installation foolproof.
Put each CPU on its own card, like the Pentium II.
>>54811998
My favorite part is the holographic tag. Like you were going to accidentally buy a counterfeit back in, what, 1997?
the only thing i'd change is add 2 rods that the cooler slides down on, to ensure it goes down flat on the cpu
getting the cooler down such that that thermal paste is spread properly is the trickiest part imo
replace pins with contacts
>>54811673
>>54811803
>>54812776
These three together make up a great idea.
The CPU should have rubber grommets surrounding holes on each side. These grommets should slide down a slick polished rod with very exact machining to make the CPU line up precisely. A circular mating connector at the bottom will add a 2nd line of fool proof design. Pins will not emerge from the socket until the cpu is locked into place.
This would be expensive, but it would work.
>>54801978
I literally just finished building my first PC and have a z170ar with the same tool. I had zero issues. Are you sure you read the instructions?
>>54811803
it'd make more sense just to chamfer one edge, sort of like a sim card
>>54812058
http://www.cnet.com/news/america-seeing-pentium-ii-fakes/
>>54801702
Slot 1 (so BGA) with magnets so it aligns and connects itself
>>54801788
>"Gaming" motherboard
Hang yourself
>>54801702
solder them in like apple RAM next question
>>54813143
Or, instead of removing a chunk of cpu, why not just add notches to it?
>>54816937
the case has the notch, not the CPU chip itself.
No chip maker is going to make a non tessellating shape as it would dramatically raise costs.
This does make me raise the question of why not hexagons given the outer perimeter would allow more contact pins for better bandwidth, while still meeting all other know requirements
>>54801819
man the buzzwords in that picture almost killed me
>easy
ok
>simplicity
wow fuck off
>quick
shut up
>precise intelligent
holy fuck
>risk of downtime
what the actual fuck
>another HP Quality differentiator
fuck
you
>no one mentions the only good approach
>it was still shit and failed horribly
pic related
if that didn't work
nothing will
razer tried hard marketing that shit, yet it failed, even with gaymen
make it so that it can only be installed one way.
>>54817409
10/10 idea
wait
Cartridge.
>>54817421
CPU's are always square shaped, if you add something like a peg on it than people will never be confused on how you install it.
You will always have some kind of retard that didn't bother check the triangle.
>>54811419
scientist can put a pc together though
>>54803599
>It's actually a lot like the mac pro, just with a little more accessibility.
I hope you mean economically accessible, because the 1st generation Mac Pros could not be easier to work on.
>>54801702
Use direct contact instead of pins.
>>54817449
No matter how square you make the peg and how round you make the hole, someone will find a way to fuck up.
It's best not to cater to idiots since there is always a better idiot.
Just make it fairly obvious, like put a little triangle on the CPU to note how it should go in, and leave it at that.
>>54801702
>making cpu installation foolproof
I don't see what's bad with the current way
>>54811419
Scientists usually rent or buy pre-build supercomputers.
>>54811188
>Do I need an x86 system to take advantage of one of these or is there some open source driver?
I feel like you don't know what you are talking about.
>>54812058
I thought they just wanted to put a cool micrograph of the architecture on it.
>>54816937
>removing a chunk of cpu
The PCB is not the die, fuckhead.
>>54817531
I was assuming that if there was a closed source driver it'd only support x86. If there's an open source driver architecture would most likely not matter because it could be compiled for other architectures. My wording probably could have been better but the point stands.
>>54817560
What do you have with full size PCI-E slots and a non X86 architecture?
>>54817601
A PowerMac G5. I'll have a MIPS board soon as well. Also, you can use a PCIe x16 card in a PCIe x1 slot using either an adapter or a dremel.
>>54817630
>A PowerMac G5
Oh. Then definitely no support.
>Also, you can use a PCIe x16 card in a PCIe x1 slot using either an adapter or a dremel.
That would seriously limit performance of that card, if it worked at all.
>>54817670
>Then definitely no support
What makes you so sure?
>That would seriously limit performance of that card
No shit
>if it worked at all
It would
>>54817630
>A PowerMac G5.
Super unsupported. They don't even support Mac Pros. It's X86, only.
>>54807249
you know this site is 18+?
>>54817838
>They don't even support Mac Pros
They don't support Mac Pros, or they don't support OS X?
>>54817864
They support 0 Apple hardware, not even x86 Apple hardware. They don't even list other 'enterprise only' architectures like POWER or SPARC. What are you even trying to do?
>>54817899
So you're telling me even if I run a supported OS it still won't support my x86 computer?
>They don't even list other 'enterprise only' architectures like POWER or SPARC
Well yeah, if it's x86 only I'd imagine there's no POWER or SPARC support.
>What are you even trying to do
Use a Xeon Phi
Why don't they make motherboards with the CPUs soldered on them? Nobody replaces one without the other anyway.
>>54817947
>Why don't they make motherboards with the CPUs soldered on them?
They do.
>Nobody replaces one without the other anyway
But that's wrong.
>>54817947
>mobo dies, oh, have to also replace the cpu
>cpu dies, oh, have to also replace the mobo
>find mobo you like, oh, it only comes with this cpu? bummer
>find cpu you like, oh, it only comes with this mobo? bummer
>>54817942
Are you being obtuse because English isn't your first language or is it because you're retarded?
>>54817982
What are you talking about? English is my only language and I'm definitely not retarded.
>>54801702
It's fairly foolproof as it is. I'd change the socket lever system slightly, such that the level supports a floating rim around the socket, onto which you load the cpu. As a result, the cpu does not touch the socket until it is fixed in a rim, which you lower in to place using the lever.
The geometry of the rim would prohibit placement in any but the correct orientation. After which there is a two step closing system - much as already exists - with the first step lowering the CPU and removing the protective cover, and the second step fastening it into the socket.
>tfw you were shaking when putting your first pc together
Haha, I'm such a pussy but at least I didn't fuck it up.
>>54801702
>intel asking the internet for advice
do your homework alone intel
>>54818008
Then why are you obtuse? Your old, obscure shit is unsupported, deal with it.
cpu installation is hard to fuck up, you have to be an idiot for that
cpu coolers are the bigger threat
>>54818163
How am I being obtuse?
>Your old, obscure shit is unsupported
So I've been told.
>deal with it
Show me where I had a problem with my PowerMac being unsupported. I understand, I don't need you to spell it out for me you fucking retard.
>>54801702
1: make the PCB and heat spreader circular, wide notches around the edge to prevent insertion in any way but the correct way.
2. socket is a rotating bayonet design, and the notches are on the rotating lock, meaning the CPU can't be inserted unless the bayonet is fully open.
3. with the CPU dropped in, you simply rotate the lock (no directional or uneven pressure) until it reaches a specified click-stop.
pic related is a shitty render I made to showcase what I mean.
>>54818191
Just be glad you computer doesn't use propitiatory cassette tapes to load things. I have to change out several tapes to get lynx running so I can get to 4chan on this thing. But it still works so I got no reason to upgrade yet.
>>54818250
wow, that's actually a cool idea
Just download the cpu
>>54818609
>mass produce FPGAs so they won't be as expensive
>use them as a CPU
>preloaded with an i3
>can download better CPUs or even different architectures
I like where this is going
>>54818750
That's barely relevant
>>54818783
Your life is barely relevant
>>54818808
At least mine is somewhat relevant, scumbag.
>>54818860
the only relevant thing you've ever done or ever will do is talk to me, bucko.
>>54818959
Sure thing kid
>>54818250
I want to add that it's probably not a good idea if the retention is purely friction-based, so there could be a spring-loaded locking pin inside the ring assembly that works like the one on a DSLR camera body - you press down a button and it releases the lock rotation - but this is adding mechanical complexity.
An alternative is that the ring itself has outward-facing flanges with holes in them that line up with the CPU cooler mounting holes when in the locked position, so that when you pass through the bolts, it becomes impossible for the ring to come loose.
The cpu is now sold with it sodered to the mother board. Problem solved
>>54819435
>sodered
?
>>54819435
This smart-ass solution has already been mentioned a few times now, although the way things are these days, your CPU may as well be soldered to the mobo, since the socket becomes obsolete every two generations courtesy of Intel.
The problem is that motherboards have the highest failure rate of any single solid-state component in a typical computer due to their inherent complexity, so the amount of wasted time and resources to ensure minimum DOA or flawed units is going to far outweigh any advantages to this design. It only makes sense on the type of products where this is already done anyway.
>>54801702
ZIF and LGA chips are nearly-child proof to install.
98% of /g/ never dealt with DIP chips.
Design the CPU and socket like a Lego block. Autists love Legos.
>>54819622
>what are pins
kys yourself
>>54819598
yep, they can barely get any easier at this point
>>54819635
But those are conductive metal, insulating plastic would be much more familiar to the idiot users out there.
It would also insure the idiots get it working.
>>54802497
I came to post this.
550mhz Ocd to 750mhz master race.
>>54812058
>not knowing about the "pentium2" made by celeron back then when intel didnt eat them
Learn to history.
>>54819636
"Should i keep pressing, it dosent fit.."
>>54820773
pressing? modern ZIF sockets are just that, "zero insertion force"
if the cpu isn't "falling" into the socket under its own weight, something's wrong
>>54820823
>Have to reinstall a CMOS chip
That one is a fucking nightmare, too much force and it bends like a spider.
>>54820895
you mean a bios chip? pictured is pretty common for those, and they're pretty easy to work with imo
>>54821026
those are more like the ones before socket 3.
Im talking this motherfucker right here.
>You are now in charge of making cpu installation foolproof.
You know there's only one option, solder them to the motherboard along with the RAM and call it a Macbook
>>54821048
oh, you mean the classic EPROM/DIP ones, yea, those were tricky without the proper removal tools
also, fyi, the "cmos" is a seperate chip to the bios, and is typically not socketed
Ship only assembled machines to fools. It's the future.
>>54821089
Ty anon, most of what i know its because experience, not much of a reader.
though Cmos and bios were that same fucking chip.
The swap was long ago with soyo mobos with amd 1700+
2 computers died, and one with a chinise voice said bios failure among other shit, the other said that the procesor might be damaged, so we tried the swap, but without proper "tools" and it didnt work, but the moment was like i was doing a surgery with stone tools.
>>54821179
Actually there is a small but growing trend to switch back to stone tools for surgery as obsidian scalpels have shown cleaner cuts and a reduction in scar tissue after surgery thanks to their super fine razor edge.
>>54821179
>Cmos and bios were that same fucking chip.
no they're not, the bios chip is EPROM/Flash that stores the system firmware (bios), and in normal operation, is read-only
the "cmos" is an SRAM chip that stores the bios configuration and time
>>54821337
I saw the microscope pics of the edges, and pics of the results after surgery.
Shit is fucking amazing, but i cant see my self checking if they are using obsidian blades because i will be KO if so, there must be for a selected group nowdays.
>>54821355
knew the concept of firmware, but not the split between the firmware and its config in 2 separated sectors.
>>54801702
Its easy enough i just wish there is limit how thin they can make anything. I have small tower coolers I used for for older generation processors without backplate, like scythe katana 3 - every single one of them bend the mobos. Mobos are like 0.5 mm thin especially the cheap ones that have are usually used for i5 6500 and such. I swear to god they have to have some standarts, they bend like a cheese. I remember the c2d c2q and older ones where tower coolers were mainstream, boards even cheap ones were more rigid. I have couple of p35 boards in my house, that were not even upper middle range, and they were build like a tanks!!! The 110$ range can get you what now days? asrock extreme 4? Thin pcb. MSI cheapest krait - wow pure black pcb, great work - wait the pcb is ultra thin again. It is not good when the 170$ mobo is the cheapest one build right now.
