> Building a gaming PC in real life is way too hard
Huh?
Huh?
I just recently put together a new PC and I was getting close to done. I plugged everything in before putting the side panels on, hit the power button and...nothing. No fans, no beeps, nada.
Okay, I'll just disconnect the ATX power cable from the motherboard and short PS_ON to ground and see if the PSU fan will spin a bit...nothing again.
Hmm. Maybe the modular cable is bad? Google the pinout for Corsair's "Type 4" PSU to make sure it's what I assume it is, drop my bit of wire down the correct pins directly on the PSU, jam my multimeter probes down a 12v + ground and...hell yeah, getting 11.something volts.
Okay, bad ATX power cable, call Corsair, wait a week, problem solved. My wife was watching me do all this on the breakfast bar and kept asking "how did you know to try that?" and I really didn't have a good answer other than "been doing this or watching my dad do this since I was 8".
Clearly step 1 is going wrong somehow, but, I mean, I'm a fairly typical HN poster, you know, I've got at least intermediate-level search skills, you know?
One of the things I always do these days is look up the memory compatibility for the motherboard I've selected, and then only buy off of that list. There are so many variations in just DDR4 2400 memory, and matching up all the parameters (timing, latency, etc.) is a hassle.
Had it happen again a couple years later but that time it turns out the motherboard was just DOA. But yeah the time spent confirming that the motherboard was the culprit sucked, and required another PC to swap parts with.
Theres two things I hate dealing with the most when it comes to PC building. Those are (1) building a PC from scratch from all new components (2) debugging a PC when a part died
This is because diagnosing the problem and identifying what was wrong could be so many possible things. Back then as well we also didnt have these nice tools like pcpartpicker.com, all of these great youtube tutorials, just a bunch of forums to debug issues.
PC not turning on? Maybe the power supply is not working. Or maybe its the power button itself is having a problem, the switch might have disconnected. Or maybe the wall power outlet is having issues, and maybe can't handle the load. Maybe you changed out a part recently and didnt connect the pins correctly, or during transportation something disconnected. Have to manually check every wire. Are the fans spinning at least?
Is the OS not booting? Okay, is BIOs at least running? Before that, can you hook up a keyboard into the PC? Is that even registering? Does it need an adapter (older PCs didnt use USB 2.0 ports for keyboard mouse). Do you have compatibility issues with hardware components?
Is the computer registering the RAM? Did you check if you slotted it in correctly, by shoving the pins far enough in and latching it? Its possiblethe RAM you recieved was dead on arrival. Try swapping the outlets and see if anything registers
When you changed the CPU did you make sure you didnt bend any pins? If you did your SOL. Maybe you put too much thermal compound between the fan and CPU.
Between all of these steps, did you make sure to ground yourself and prevent static charge from frying your motherboard?
Is your OS booting but your having software issues? Run safemode and find the culprit.
The list goes on and on and it just never ends. Narrowing down the specific issue was always a PITA.
I used to go to a screwdriver shop (in Federal Way?) that had some memory and a power supply sitting around at the counter. They would POST a new CPU and board in front of you before you take them home. Less stress than ordering from some random in the Computer Shopper phonebook.
It's pretty straightforward and easy to do, so you don't get any bragging rights when you finish, but there also aren't a ton of safety nets in case you do something stupid to keep you from ruining an expensive component.
When I put together my computer, I dropped a CPU at an angle and bent one of the pins on the motherboard, then spent 3 hours flicking it with a needle to get it to realign.
It's like doing taxes - not really an accomplishment, so big whoop if you do it right, but just dangerous enough to mean you have to pay attention and take it seriously.
CS person: grabs part by hand, shuffles across the floor in socks, brushes dust off a pin, presses part into place
EE person: stares in horror, a grounding strap dangling from their hand
CS: "What?"
EE: "You know you're supposed to ground yourself to prevent static discharge from frying parts?"
CS: "Bah, it worked fine the last time I did it this way."
Point being there's a difference between something that works 99/100 times, and something that works 60/100 times. And when failure means a dead part, probably best to go the extra mile.
If it's your bud's money... it will probably be fine. Just hold the beer when you are installing the CPU :)
Your PC build didn't boot? Probably shocked it. It's blue screening or kernel panicking? Probably shocked it.
Drivers and OS are much more likely to be the cause of quirks and weird behaviour.
What I did instead was touching a radiator every once in a while. Is this sufficient?
The biggest problems come from the fans running continuously and getting worn down, and the coolers are by far the easiest part to replace.
Assembly a PC is pretty easy these days, most of the parts go together pretty easily, and your chances of everything working on the first try are pretty high. This is way different from ye olden days when it might take hours just to get a PC put together correctly unless you were lucky, hardware is a lot less fiddly these days than it used to be.
But buying all the components for a PC is more challenging for novices. This can be hard for techies to understand because we just assume that everyone knows what we know. But your J. Average prospective computer building may not know whether a Core i7 8700 is better than a Core 2 Quad or vice versa. They also may not know what the heck an 8th generation Core i7 is. There are so many generations of components now and so many gradations within each generation that it can become and informational overload. If you try to go at the problem directly by just heading to newegg or amazon and browsing for parts without doing the research you're going to have a bad time.
There are a zillion different socket formats, there are a zillion different motherboard form factors, there are a zillion different GPU lines and each one has a zillion different variations. How do you know which are compatible? How do you know which ones are the best? Or just the best for the price? Are higher numbers better? Worse? Not important? Is LGA 2066 better than LGA 1151? Is a GeForce 9800 better than a GTX 1080? There are a zillion of these little questions. Already knowing most of the answers makes navigating the problem vastly easier, having to find an answer for each and every one of them can be exhausting. You absolutely need some sort of guide or hand holding to get started on this problem from scratch but if you don't have any pre-existing knowledge you don't know where to go. (Personally, my favorite is http://www.logicalincrements.com but there are others.) You need someone to help you narrow the scope of the decision problem to a manageable level, otherwise you'll drown in choices and questions until you run out of will power to continue.
It's a really big matrix of possibilities that can be pretty opaque to people who don't keep up with it.
I wish I could specify other specific priorities, like video conversion or compression or data management, so it would trade off between CPU, SSD speed, SSD storage space, or RAM in slightly different ways (though maybe that level of optimization is overkill).
> I wish I could specify other specific priorities
I tend to think that's overkill :) But one can toggle between a gamer setup and a mainly more more cpu heavy app focus, it's in the advanced menu under the recommend button. Switching between gpu and cpu heavy builds catches a lot of those different usecases.
> given the availability of crowdsourced benchmarks (http://www.userbenchmark.com).
One problem with those is that they are often based on artificial benchmarks, which is most of the time not what you want. I use benchmark results from professional publications instead. The problem then moves to ordering, but that's manageable with enough data.
> I tend to think that's overkill :)
Yeah, I think you're right. It's like that old Knuth quote. Programmers waste enormous amounts of time worrying about the speed of noncritical parts of programs.
> I use benchmark results from professional publications instead.
Yeah, I generally do that too, I just thought userbenchmarks would be easier to API and be more comprehensive.
Though thinking about it now, I'm not sure why I distrust UB so much. In a sense I'm just choosing different biases. Even if I find a reliable reviewer that personally buys each part at a retailer to avoid "reviewer binning," the best case, I still risk a sampling error due to the small numbers (typically n = 1) tested. The bell curves on two different CPUs might overlap considerably. So buying a marginally better rated processor might only give you better performance in 51% of purchases.
No professional reviewer could afford the cost or time to individually buy and test a statistically significant number of devices to determine its variance in operation.
On the other hand, the errors in judgment that might creep in from variance due to small samples are almost certainly washed out completely by the leaps between generations, right? So trustworthy reviewers are probably "good enough" and worrying about that issue is another form of unnecessary optimization, as much overkill as my first idea.
I'm open to that possibility.
Though we might go further down this line... Maybe the performance exaggerations due to conformance to published benchmarks aren't much of an issue either. As long as we're talking about ranking cards for real world usage, the odds that a card will perform noticeably better at benchmarks but noticeably worse at tasks I care about is... Well it's nonzero, but small. Smaller still if we upgrade our threshold from "noticeably" to "significantly" or "substantially."
I mean, it's your own call where you pull your data from because no data is perfect, and we probably pull from the same places. But you really got me running thinking everything is needless optimization and now I can't stop. :)
I now look at cpus again.... yeah, there it would not work, that is too far off of what I see in gaming performance. But cpus are hard, with single and multi/threaded performance differing so much and having different effects in each game you look at.
So imho one could only use the gpu ranking, but the gpus are not hard to rank anyway :)
But now I wonder whether that would not be a good data source for SSD and HDD performance. Maybe even for RAM...
Thanks for your thoughts, you made me take a second look on something I had discarded earlier!
A shame.
That probably got missed by not being able to remove windows just by using the arrows at the left of the box, right?
Now I see there's just a windows logo, but no actual windows pre-picked.
Why not put something else than a windows logo in there, when it's not filled?
I don't keep up with current events in custom PCs and when I built a PC last year [2] I thought it was incredibly straight forward.
[1] pcpartpicker.com
For others, even just pcpartpicker can present an overload of information.
https://pcpartpicker.com/products/cpu/
Which CPU family and which series? Once you finish THAT, everything gets a lot more straightforward.
That list is terrible. At least order the families/series chronologically.
If I'm a reasonably informed random, I know to click a handful of things with "lake" in the name and a handful of things with "zen" in the name and that's about it. Except you can't do that with the chooser as it gives you zero choices, since -zen is a series and -lake is a family and those are disjoint (I should be able to union those).
As a noob, I would miss any of the "gaming value" choices with that heuristic (Haswell/Broadwell that are quite nice).
I don't see any EPYC in that list, so I'm scratching my head (yes, I know it's a server but a random person won't) and I would miss the Threadripper choice (why doesn't it have Ryzen next to it).
The problem isn't assembling a PC. It's the fact that PC knowledge has a half-life of like 6 months, and you really can't trust review sites or web searches on this topic anymore.
The problem, as a noob, is that you don't even have enough knowledge to use the filters correctly.
Maybe that particular filter on that particular part isn't that easy to use but I think sites like this make building a PC extremely approachable, even for someone who has never built a PC before.
Should also be mentioned that the manuals that come with things like motherboards and cases are much more detailed than they used to be.
PCPP is really revolutionary. Building a PC in 2018 should be easy.
The only priblem is a lot of everchanging technologies and acronyms, e.g. NVMe, M.2, etc.
Pick a processor, that determines the motherboard socket. The motherboard and CPU arch will determine the memory type. The wattage of the CPU and video card determines your PSU lower bounds.
The only odd thing was the addition of SSDs on a stick, but the motherboard determines the options there too.
Long story short, a couple of resistors and diodes later my WD Blues were spinning up again, but it was a pain in the ass.
You really have to rewind to mid 90's pre-Windows 95 in the days before Plug 'n' Play when manually assigning IRQs etc was a thing before the process of building your own PC meaningfully changes in my opinion.
OTOH, before the main consumer storefronts were the kind of flea markets of no-name sellers that used to roll into county fairgrounds every couple months, you could actually find good retailers with not only strong return policies, but also focussed on quality parts with a high S/N ratio in their catalog, that even offered added value (including, e.g., mounting CPU and memory when bought with a motherboard.)
Some of those same places survived, but they don't do consumer retail anymore.
Or possibly you had money to buy higher quality components than I did at the time.
Still, before AGP we had VESA, also around the time of MCA and IBM's attempt to capture some copyright patent control back from the expanding PC market and BIOS clones ("Halt and Catch Fire" Season one well worth a watch if you missed that era in life).
Yes pre Windows 95 was a fine art, from tweaking DOS memory, to optimising windows .ini files and then the fun of having to dial up Microsoft's patch BBS for a TCP/IP stack.
But top-end gaming PC's back then, sure made less noise.
If the argument were "I have other things I prefer to spend time on" then I would completely understand, but rejecting it as too hard seems somewhat lazy unless I'm just overestimating the average person's ability.
It’s nerve-wracking as hell. Closing the swingarm to hold the CPU in place feels like—as the article’s author points out—you’re 100% about to destroy a really expensive piece of hardware. Installing the CPU cooler was surprisingly difficult, and I only noticed that the pins holding it attached weren’t fully seated because I checked and rechecked the mounting. Every time you plug a cable in there’s a moment of doubt if it’s the right cable or if you’re about to overvolt something and melt it. Even when you’re done plugging everything in there’s anxiety around whether or not you’ve missed something like a fan cable, that will eventually result in an expensive piece of hardware overheating and taking out the entire machine.
The act of pressing the power button when you’re all done isn’t an exciting and satisfying one. It’s a massive surge of anxiety followed by a wave of relief, followed by more nagging anxiety that gradually fades.
Even then you’re not done. After fiddling around in the BIOS for an hour to set things up, I ran a benchmark and discovered that I hadn’t used the right profile for my RAM, and it was only using half of its performance potential. Apparently many other people did this too, because the benchmark showed a graph of others’ performance in comparison and it had two equal-sized peaks: one where it was misconfigured to a slower (but presumably safe for all configurations) speed, and one where it was operating at its published speed. It took a few hours of research to discover something called XMP—which I still have no idea what it is—that apparently allows my motherboard to detect RAM timings in a way that unlocks this performance. I never would have noticed this without the benchmark.
I also noticed that my new NVMe SSD wasn’t performing well. Turns out the heat spreader my motherboard uses actually worsens performance for whatever reason. Simple fix: open the box back up, disconnect cables until I can get to the m.2 slot, and remove the heat spreader. More anxiety follows: am I just killing this expensive SSD?
Every step of the way is like this. I am not a naturally anxious person. I am financially well-off, and easily able to eat the cost of any mistakes. I am extremely technically inclined. Expecting nontechnical people to do this with limited budgets and expensive components is nuts.
On my latest Asus motherboard, an Asus Strix Gaming, choosing XMP timings with an otherwise mild overclock overvolted the 6850k in a voltage setting I wasn't even aware of (not vcore) in such a way that the CPU died over time - my last build was a 2600k, which is now son's computer (OCed, and working fine).
This happened twice. Computer shop couldn't figure it out and replaced the CPU on warranty. Not Intel's fault, yet they copped it. I should have insisted on a new motherboard as well, very frustrating.
Fortunately I somehow discovered the hyper aggressive setting not too far into my third. I know it is at least a bit degraded, but oh well...
In 3 years I'll try an AMD. ;)
Plus, there's still a lot that actually can go wrong. Crappy PSUs and built up static electricity have killed uncountable motherboards and GPUs. Misconfiguring your BIOS can actually kill your components, sometimes quickly and sometimes slowly.
GPU hit > 95˚C and is now a several hundred dollar paperweight.
I'm sure if you talk to someone working at a bike shop they could take your comment, flip it, and still have it be accurate:
Two of my friends build PCs from scratch, but they find building a bicycle difficult. I argue that their hobby is ten times more complex than building a bicycle nowadays, as everything is mostly plug-in-play and idiot-proof. Truing your wheel or overhaul a hub is a walk-in-the-park compared to installing the CPU.
And then there's those poorly documented nonstandardized fiddly little wire connectors for all of the front panel functions, plus another set for the USB ports and so on. To this day I'm never sure which way to plug in the LEDs, and even when I think I have it down I'll run into a mobo that does it the opposite way. And then your case has a three pin power switch (outer two wired up), but the motherboard only has two pins for it...
First of all, pcpartpicker and logicalincrements will help you choosing the right CPU with the right motherboard. Correct timings on RAM hasn't been a thing for many, many years, they work, at worst it won't use the maximum possibility the module offers. The PSU is hardly a problem, unless you are spending an extraordinary amount of money, the smallest PSU you can realistically buy is an overkill (CPU is 65W typically, very few GPUs go above 150W, and you can't even get a below 350W PSU really unless you have some office prebuilt from HP or Dell).
The front I/O pretty much has standardized on Intel https://www.intel.com/content/dam/support/us/en/images/mothe... and most cases have a single block now.
A couple of examples off the top of my head: AMD RX480 power draw exceeds the pci-express current limit specification and bricked motherboards, hence I bought/overclocked an RX470 instead. Intel 6700k overclocking. Wow. Reading and understanding the thermal and power envelopes described in the 300 page datasheet difficult for me, it would be impossible for a lay person to comprehend.
To me, almost anything that can be completed in a fixed amount of time with high probability isn't hard; it's tedious at most. But if it can be completed within one hour it's not even tedious, it's (IMO) 'easy'.
For something to be 'hard', there should be a reasonable chance of failure with associated non-trivial costs.
Any single part of a computer is a non-trivial cost if broken, and it's not too difficult to imagine the situations where that might happen: bending a pin on a CPU, giving a static shock to a component, misapplying thermal paste.
Even for enthusiasts, this is a task they do perhaps every 3 years, and every time with different components.
It takes willful ignorance and disregard for the simple instructions included with each component to even have a chance at breaking the expensive things.
Can confirm, got a little lazy and put spare components together into a working system on the carpet this past weekend. Everything still works well enough. :)
Pins are in the socket now and very small, pretty much impossible to bend without negligence. Thermal paste is also usually pre-applied to the stock cooler, and you won't need to apply your own unless you're using an aftermarket cooler and overclocking.
Static shock is probably the only real concern but honestly I've built at least 20 PCs, at least 5 while wearing socks on a carpet. I've just never seen it be an issue or even met someone in real life who shocked hardware.
Parts are also much more durable than you'd initially believe. I competed in a PC building race (weird, right?) and most builds took 2 minutes or less (some stuff was pre-assembled). Hardware was slammed into motherboards as fast as possible, clumsily dropped on the floor, no regard for static. No attendee had any hardware issues.
> Even for enthusiasts, this is a task they do perhaps every 3 years, and every time with different components.
Nah, for those of us who love it, we'll gladly build your PC for free (for friends) if you buy the parts, I'll even double check the build first. Though I do try to haggle a six pack of beer out of the deal.
Example of modern pins:
https://www.techpowerup.com/reviews/MSI/Z170A_GAMING_M7/imag...
I did this back in highschool as well. We actually did have a team one year who broke a computer but that's because they were bleeding all over everything. There's a surprising amount of blood in these competitions when you're racing to put components into a completely unknown case with sharp edges all over.
And you will put too much paste on, or not enough.
Back in the days it was almost recommended to increase the iron in your diet prior to commencing a PC build.
Otherwise it’s the usual dread of whether the damn thing will actually boot. These days, though, they’re much better at booting than they are at not booting, so most of that nail biting is gone.