I know we can probably already build them but they're not exactly well known or mainstream
So when do you think this trend will take off (again)?
You can build one already.
The trends are for small PCs, so all the advantages of having TDPs are removed since there's less to work with.
although it's best to instead of having a completely passive PC, have one or two large diameter fans spinning very slowly, producing a small airflow with below audible sound. Below audible fans are better than no fans.
>>53819634
>How far away are we from passively cooled desktop PCs?
They exist.
>they're not exactly well known or mainstream
Most people don't need them,
>So when do you think this trend will take off (again)?
Not for a very long time.
Fans are cheap, and greatly increase the performance per dollar of a desktop computer.
>>53819634
There's a lot of small passively cooled Atom-based boxes you can buy. They work for web browsing and office stuff just fine.
For desktop PCs, I don't really think ever. You can a lot more performance with active cooling.
For other tasks, like HTPCs, it makes snese
>>53819634
It's fairly easy to do as long as you don't need enormous power. Pic related have pretty good reviews as long as you use a case with loads of ventilation. Work well up to about 80w tdp
And fanless PSUs aren't that expensive, neither are SSDs so you can get a silent pc fairly easily
>>53819634
My i5-4590 is passively cooled right now.
Be Quiet Dark Rock 3, case only has a few ULV noctuas for airflow.
Doesn't go above 85c in synthetics, let alone real world.
With pascal/polaris coming, we'll have graphics cards that can do 1080@60 V-High/Ultra while passive.
Titanium PSUs don't need the fans unless you're really hitting them, SSDs make fuckall heat and the motherboard doesn't need a lot of airflow, it just doesn't like stagnant air.
>>53819634
That looks like it came from a nuclear power plant
You can already build them quite easily if you don't need to do heavy work, a big heatsink should be able to cool a low power cpu just fine.
>>53819634
We already have them, and they've been around for a while.
New CPUs & GPUs coming out run on lower wattage and with less heat than anything ever previously ... so people are just going to start taking fans off their cooling blocks and let general air circulation within the case handle cooling needs. This is probably a bit before your time, but I used to run PII and PII CPUs from about 233 MHz up to about 900 MHz on passive, with just one and sometimes two 90mm case fans pushing air into a case and a single PSU fan assisting in pulling air out, the remainder exhausting at opportunity through various openings in the case.
Also, technically, a completely passively cooled desktop/tower rig wouldn't even have case fans, and I can't imagine PSUs not using fans for some time to come (at least, not on power-hungry rigs; I know there are some passive PSUs available). I do not see that happening any time soon. Good, quiet case fans that push 60+ CFM would still be required or highly recommended even if you're CPU & GPU can run on passive.
> they're not exactly well known or mainstream
I think you meant to write there something like:
> I just heard about them
> what a cool idea
Anyone who's been building rigs since the mid-1990s is aware of hem and has used them ... just that CPU/GPU development went through a phase that generated so much heat it was not a viable option for most applications.
Tell me all about the CPU fan in your cell phone, or your Raspberry Pi, for example.
>>53821219
You can do extremely heavy work with ease, pair an i7-6700T with a NH-D15 or other xboxhueg cooler and you can 100% it all day every day right next to a reference cooled R9 290X without it giving a shit.
This looks comfy.
>>53819634
I like the concept, and I like silence, but the principle itself has problems.
Firstly, it's another step backwards in efficiency. Water-cooling targets the hotspots and removes the heat from them very efficiently, conventional air cooling with local fans is a little less efficient, but air cooling with fans remote to the heatsink is another inefficiency on top.
Since developers are continuously pushing to use all of the compute resources available to them, regardless of how advanced processors are, active cooling will always enable them to run at higher frequencies, and hence run applications quicker.
>>53821100
I just want a desktop that can play some decentish vidya and also be as quiet as my laptop
Not too much to ask right
>>53819659
>slow fans
Damn that sounds pretty nito, how have I never heard about this?
>>53822755
because you're an ignoramus
>>53819634
We had them some 20 years ago.
>>53822755
It's not that common unless you search specifically for quiet gaming computers
>>53822755
buy a case with space for 200mm fans on the top and front
>>53822730
Get a large case with lots of airflow, and use large low rpm fans.
Honestly, I used to seriously consider building a passively cooled PC, but when all my fans were new I couldn't even hear my PC when it was on.
It's 6 years old now and it is just loud enough to hear now, but still much quieter than any PC I have previously owned that used 80mm fans.
>>53822234
Why go "gaming motherboards" not have the components all in the same direction like server hardware? Should be good for cooling
You're better off having a bunch of large fans running at low speed.
>>53822833
because they don't have the fans that put out massive CFM's
those niggers are 30mm thick and sound like hornets
>>53822730
Normally it's other way around. Typically the smaller the fan is, the faster is has to spin to move the same volume of air, so it's more noisy.
My gpu fans don't spin unless I play a game.
This is probably the way to go rather than abandoning fans entirely.
>>53822906
?
that's common
not even worth mentioning
and that still doesn't make it fanless
why have a silent pc when it gets loud as soon as your GPU is under some load?
Several generations in. That's what the Broadwell and Braswell BGA parts were for.
For example, Gigabyte GA-N3050N-D2P, Asus N3150M-E, Asrock N3700-ITX, etc. They are fairly low-cost, and would all work fine for an office PC using Intel's integrated graphics and the cheap Realtek NIC.
There are also a whole bunch of industrial-spec PCs, along similar lines, particularly thin-ITX ones. They tend to bring out a lot of COM ports.
The Avoton SoCs - a generation or two behind unfortunately, the Denverton 16-core successor won't be out until late 2016 - allow for building ECC-supporting storage-oriented servers, with plenty of SATA2 ports - in particular the Supermicro A1SAi-2750F and Asrock Rack C2750D4I are worth looking into if you want to build a passive, low-power storage server. Both of those bring out the internal Intel NICs with the Marvell PHY, instead of cheaping out. (Of course, the disks are going to dominate your noise and power in that case, but they're still highly suitable for a NAS build - and I think that's broadly what the C2 builds from Scaleway are, too.)
If they were really going to compete, however, the low-end boards would need to be cheaper. Low-end builds are starting to switch to small cheap SSDs instead of small cheap hard disks, but the CPU price would need to come down to really compete. Maybe Zen will shake things up enough for Intel to lower their Atom prices (I don't anticipate a very low-power 2-core Zen being on the immediate roadmap).
For high-end passive builds, well, Zalman used to make this (>30kg!!) copper/aluminium beast, the TNN 500AF. It's a highly niche market, however, and the cost was astronomical. They didn't sell very well... except to music studios.
>>53822880
true, loud as fuck, but in a desktop they can use the larger fans going in 1 direction
>>53822996
gaming computers just don't have the component density to require such engineering
>>53822965
>except to music studios
reminds me of RAID cards. Atto specifically doesn't fans on their cards so that they can be used for music and movie production.
>>53819634
I've compiled kernel + modules looped 10 times while using every thread on a 3.3Ghz haswell xeon under a d14 inside a r3 powered by a fanless 400w PSU w/o any fans running at 78F room temp. The on die cpu sensor never topped 180F and compile times were equivalent to the system being actively cooled which means thermal throttling probably never kicked in.
>>53822956
Why not have the option? It's silent when I don't play a game, and when I do the sound of the game drowns out the sound of the fans.
If you don't have proper cooling, it's never going to be as powerful as it could be, unless technology changes drastically.
>>53819634
>How far away are we from passively cooled desktop PCs?
We've gone 20 years too far. Modern CPUs need big coolers designed for 120mm fans to stay cool, Old high end CPUs only needed a small heatsink cooled by case airflow. High end graphics cards either had no heatsinks at all, or a small aluminium one with no fan.
Some systems had tiny heatsinks with tiny fans. Sometimes they didn't even have heatsinks. Just a tiny fan in a cup. But you could replace those with passive heatsinks.
Harddrives and PSU fans were really noisy tho.
Current computers are faster, but I still feel they've regressed a lot in terms of top end power consumption and the cooling systems our components need.
I've measured my current desktop to sit at around 50 to 400W measured at the wall. My old computers measures at 60 to 120W, and that's with an old inefficient power supply.
>>53823052
>If you don't have proper cooling, it's never going to be as powerful as it could be, unless technology changes drastically.
obviously
>>53819634
There's already passively cooled desktops and laptops.
>>53819634
I built one with that very cooler OP. 6100T Skylake, Corsair switched PSU. Couple of points:
Motherboard hotspots - true fanless PCs lack the usual drafts that motherboards tend to expect, so they hotspot. Good motherboard, understressed, good heatsinks, no shields.
Use a vertical oriented case with lots of holes in it, PSU at the bottom to maximize chimney effect inside case, don't overload on peripherals, and holes on the top are a good idea. If you're have to graphics card, performance will be gimped, probably best in a low profile format with a very accommodating custom heatsink. Better to use the onboard tbqh.
>>53823402
Could I get away with doing this on a micro atx build? Haven't made a build for a long time now so sorry if that's dumb
>>53824602
>micro atx
Do you have any idea how large those things are?
>>53825039
Looks like a hamster wheel
>>53824602
Better to go with the quieter fans than with passive.
You could also invest in a quieter case (one with better insulation)
Case fans are also important.
That Nofan cooler could only realistically cool a T Series intel CPU, right?
>>53825365
Up to 100 watts TDP
>>53821264
>tfw actually have a cpu fan for my raspi
I kinda abuse that thing though and it gets noticably hot. Another raspi of mine though has a 'normal' load and sits at a comfortable temp with no active cooling at all.
>>53819634
Better question is how close we are to actively cooled phones?
Doesn't it become inefficient when they have the same temperature i.e. room temperature? Room temperature isn't bad in that case but it's not amazing either.
>>53822234
>MSI logo on top of cooling fins.
Who gives a fuck about cooling as long as it has our fucking logo on it.
Did that spinning heatsink meme ever come out?
