What's with all that crap they stuck on the board. You can practically get a epileptic fit from it.
I don't like that X570 needs a chipset fan, but at least cooling's useful. I can't have as much charity toward LEDs. Although I do like 'em. The æsthetic is sick. Although not this one: https://smile.amazon.com/G-Skill-Trident-3600Mhz-PC4-28800-C...
Not at all true. It's around 0.1-0.2 watt per LED. A RAM stick is going to have maybe a dozen at most, so that's an extra 1-2 watts. Similar for a motherboard. Maybe a fully binged out build is burning 10-20 watts on RGB, give or take.
That's comparable to the power usage of just the VRMs or the X570 chipset (which is 10-15w), to say nothing of the CPU or GPU itself (both of which commonly blow past 100W under loads)
LEDs are very, very efficient. They do not take a "disproportionate ton of power."
10-20W is 10-40% of idle power consumption, perhaps, for a workstation. So it wastes a lot even in relative terms.
Then turn them off? The whole point of RGB lighting is it's controllable. That includes setting it to off.
(Also computer hardware vendors seem particularly fond of blue LEDs, which is the most annoying color to some people.)
And yes, I mean RAM, GPU, motherboard, fans, etc advertised as "RGB". If it's the same price and includes LED, then for all purposes it doesn't have LED.
Not a single vendor supports Linux. Luckily some things have been reverse engineered, but it's very far from "just turn it off". Even on windows it sucks if you're mixing brands. MSI Mobo, Asus GPU, Corsair RAM, gonna need 3 separate programs running at startup to control those LEDs.
I get that some object to non-open standards on principle, but RGB is not a serious impediment to availability or use. Looking, there are non-RGB versions of most current-gen hardware available from popular resellers, and there are plenty of OEM options that are RGB-free. What I don't have racked somewhere else is in solid 'silent' cases with as few moving parts as required, but I bought parts with minimal/no lighting anyway.
I certainly hope that everyone can find something amenable to their situation, but I also feel that objecting doesn't help when an hour's work or research beforehand can allieviate the cause.
The RGB stuff sorta grew on me and it's kind of interesting on Linux to be able to see the CPU fan stop spinning completely (never seems to happen on W10 though...).
It’s extremely hard to build a PC that would fit my personal aesthetics and even if I can turn those off the mere thought of having to configure that and the mere thought of those LEDs existing at all offends me on a deep level.
You have limitless choices when it comes to building a PC but the aesthetics of the things you can get are awful and tiring and just … ugly.
the point is to market the premium RGB'd hardware as some game specific hardware to kids that think "game cache" is anything but more memory
Garish LEDs aside, marketing that sort of blatantly false exaggerated spec should be illegal. 1M hours is over 100 years.
Also, keep in mind that MTBF is not expected lifetime or anything at all like that.
What a 1 million hour MTBF actually means is that if you had 1 million of these drives running at the same time, you'd expect one failure per hour. (Or similarly, if you had 1000 drives, you'd expect a failure every 1000 hours).
100 drives - failure every 10,000 hours
10 drives - failure every 100,000 hours
1 drive - failure every 1,000,000 hours...
How is this not expected lifetime?
[edit: split lines]
It would be nice to know the time window, but I'd expect it to be somewhere in the 5 to 10 year range.
That equals 0.86% Annual Failure Rate (AFR).
Backblaze regularly provides their real-world-usage drive stats: https://www.backblaze.com/blog/backblaze-hard-drive-stats-q1...
Seagate announced they would use AFR instead of MTBF in the future: https://www.seagate.com/de/de/support/kb/hard-disk-drive-rel...
Of course, it also depends on what you count as a failure. Is a drive that can no longer write but can still read a "failure"? Probably - but that's a considerably softer way to fail than a hard drive with a bad bearing or head.
So my comment is 100% nitpick but I do want to make a point here.
X year failure rate divided by X is not the same as AFR. AFR will always be higher, because it has to compensate for each year's pool of potential failures shrinking.
The correct equation is already there, 1-e^-(n/114), and the correct number is .87%. And while that's only a percent off, it's lucky to only be a percent off. If MTBF was 113 years it would be .86 vs. .88, which is the difference between two digits of accuracy and one and a half digits of accuracy. Losing a quarter of your accuracy is not great!
http://in-win.com/ makes quality cases without windows.
its gross
https://www.anandtech.com/show/14161/the-amd-x570-motherboar...
ah yes. The fabled colourful plastic heat sinks.
Why bother with copper when plastic with flashy graphics dissipates heat just as well.
Thankfully my tastes have grown since then, but I can't blame kids today for doing the same thing I was doing (though I can grumble that they have fancy cases and led controllers and aio coolers where I had beige boxes and holesaws and custom wiring and fish tank gear).
However, the difference is now that we are effectively being forced to pay for features we do not want and which are a waste of material resources and power. Previously, you would have needed to get those extra bits yourself.
I've seen only one tidy motherboard for these new chips and its from Asus.
https://www.asus.com/Motherboards/Pro-WS-X570-ACE/
Look at that, a normal looking VRM & chipset heatsink, and somehow they've resisted the urge to stick a fake plastic gun magazine onto the board!
Supermicro server boards are about as utilitarian as you can get, with off the shelf heatsinks slapped on them, its only the consumer gear that wastes money on this stuff.