X299 confirmed housefires

Reports are showing most x299 motherboards are burning out VRMs at even modest OC's.

>Skylake-x was not rushed at all, all part of the plan.

When will Intel learn?

youtube.com/watch?v=f7BqAjC4ZCc

Other urls found in this thread:

legitreviews.com/intel-x99-motherboard-goes-up-in-smoke-for-reasons-unknown_150008
youtube.com/watch?v=3VV_It9kdrs
evga.com/thermalmod/
gamersnexus.net/guides/2691-final-evga-vrm-thermal-torture-test-and-analysis
youtube.com/watch?v=VfgbYSuZC2U
twitter.com/SFWRedditVideos

Well, when you put hundreds of watts through a socket, that tends to happen.

this.
intel needed the CPU to work for show and under normal conditions, they didn't care for the OC people enough to raise the limit.

It still hits 100c on synthetic load (Prime95) even with an AIO, though. This platform is really just such a mess.

What are they even going to do about the 12-18 core CPUs? They're gonna have anemic clockspeeds, or they'll have to put out a new motherboard to handle it. Hell, sounds like they might need to put out a new motherboard anyway.

AMD Vega overclocked + Skylake-X overclocked when?

Make some charcoal

Oh god it's debauer again, you know this is gonna bring in a major shitstorm

I knew I bought that 1500 watt PSU for a reason.

Make sure to get the 375 watt watercooled version. 300 watt aircooled version is for pussies.

>Reports
>Youtube video

Brought to you by AMD

nah, water one will be less silent at 100%, I want WHIRRRRING noises, combined with the 400mm double rad I need for cooling skylake it's gonna sound like an airport.

It's debauer, he's a renowned intel overclocker.

pls think before you sitpost

>x299 confirmed housefires (plural)
>Reports (plural)
>It's one guy who intentionally overclocks shit past reason for youtube views

120-130C VRM TEMPS ON NON AVX

WHAT A DISASTER
I don't know which one will be a biggest shistorm come tomorrow, this or Vega

Time to buy stock in uranium mines I guess.

>4.5 1,2v
>past reason

Did you even watch the video?

>cables get so hot they might catch fire and blow up your PSU
Lmao

>Brought to you by AMD
t. fedora

>frequency jumping from 4.6GHz to 1.2GHz because everything is close to blowing up

This is crazy

>everyone designed their VRM heatsinks to look cool rather than perform their required function
How the fuck does this even happen?

>4.6GHz
>1.25V
>beyond reason

>assume it's some random shitty kiketube channel making a fake video for clicks
>it's actually der8auer

Hahahaha, Intel on suicide watch.

Intel rushing shit, this will be a chomp on the ass instead of a regular bite Skylake-X reviews brought them.

>$300+ motherboards throttling, blowing up VRM and generally having disgusting build quality

WHAT IS GOING THE FUCK ON HERE

Gonna be a repeat of this, but this time there's no easy fix, some motherboards are just shit and can't OC. legitreviews.com/intel-x99-motherboard-goes-up-in-smoke-for-reasons-unknown_150008

Honestly Skylake-X CPUs are crazy enough even without OC. I don't think anyone should purchase these with the intent of OCing, just run at stock if you're nuts enough to buy it.

we're gonna need 1000 watt thermal coolers for this one.

I'll just throw this in here too, video exploring the inconsistent performance of Skylake-X due to the various architecture changes (mesh, L2/L3 cache sizes) youtube.com/watch?v=3VV_It9kdrs

Just get a chiller/compressor bro

How did you complete the captcha?

Blame Intel. Motherboard manufacturers are trying to cope with a CPU that draws 300W with a mild overclock.

>can't even do a stable clock over 4.5gHz

wow intel fucked up hard

You can, but you need a 600 dollar motherboard

Not Intel! Mothebroad guys..

You can go cheaper than that, just watercool the mosfets, didn't you watch the video?

Gonna be a tight fit getting the water blocks on alongside the LN2 pot on the CPU.

>just watercool the mosfets

I love the fact this is not told in jest.

GPUs draw 300W and above and manufacturers somehow manage to design adequate cooling for their VRMs, most of the time at least.

Comparing the two is utterly retarded. Graphics card VRMs are generally hooked up right to the main fin array, and cooled by two or three fans. When you try to cool them with a tiny plate, you get EVGA tier temperatures and exploding cards. You can't have a fucking graphics card-sized heatsink and several 100mm fans on a motherboard. There'd be no room for the goddamn socket, RAM or anything else for that matter.

>$1000

>watercooling the vrm's is no longer a meme

You don't need a graphics card sized heatsink. If you watch the video, he's saying that the VRMs are literally better off with a little bit of airflow over them without the heatsink at all. So the heatsink in question is so ineffective and shit that you're better off removing it entirely.

probably due to the VRM design.

more than likely they will be designed to allow for better heatsink cooling if intel doesnt get their shit together

Skylake 18 core plus vega pro duo. All this connected with a car engine rad and fan. That beast would consume 1200w+ at full load.

Why even turn on your heater/AC during winter? You're practically getting a 2 in 1 deal here

this isnt even the 18 core model.

how the fuck will the 18 core model work at all even without overclocking?

He's putting a 120mm fan over them. That makes a huge fucking difference, heatsink or no heatsink. You cannot have any sort of fan in that position on a motherboard. It has to be passively-cooled, which is a huge fucking ask when dealing with a CPU that draws THAT much power.

You apparently don't have any concept of just how much difference airflow makes to cooling. It's why passively-cooled 750 Tis for example have giant fucking heatsinks on them and STILL run way hotter than the ones with tiny heatsinks and a single fan.

By running sub 3.2GHz clocks

Either Xeon level clockspeeds or new motherboards that can handle 200+ watts at stock.

>pump more power

Oh yeah I wonder how that will turn out.

...

Just put a heatsink on VRM and problem solved!

You still got the melting mosfets, but hey you're off to a start!

>temps are so crazy they run the risk of melting the PSU cables
I realize the housefire meme is just a joke, but I think Intel mighty actually burn somebody's house down with this shit.

>Graphics card VRMs are generally hooked up right to the main fin array
This is completely incorrect, the main heatsink usually cools the GPU alone, not even the VRAM is included. In any case all the heatpipes start from the GPU. The VRAM and VRMs are often either cooled by a plate or by smaller heatsinks underneath the big one for the GPU. They get (warm) airflow from the fans which first blow through the main GPU heatsink.

The EVGA burning VRMs did not happen because they were cooled by a plate, but because they weren't fucking cooled at all. The VRMs on the back were covered by the backplate but WERE NOT touching it. The "mod" EVGA distributed simply had people add a thermal pad between the VRMs and the backplate, which did in fact provide completely adequate cooling.

The issue on X299 can only be caused by utter shit thermal pads but more likely than not the real problem is that there's no airflow and the blocks of metal they slapped on top the MOSFETs have no fins and a very small heat dissipation area. If they designed an actual heatsink instead of a useless piece of painted metal meant to look cool the problem would not exist. Hell, if the problem were so severe they could've even used a small, quiet temperature-controlled fan for the power delivery. Those used to be a thing back in the day, mobos had a little fan and the VRM/northbridge/southbridge had real radiators and some were even linked together with heatpipes. They must've forgotten what a heatsink looks like since they learned about RGB decorations.

Shut up, goy!

Dumb Frenchman.

DELETE THIS VIDEO NOW

ISRAELS INTEL PLANT MUST STAY OPEN

SHREEEEEEEEEEEEEEEEEEEEEEEEEE

He's German you fucking retard

>The EVGA burning VRMs did not happen because they were cooled by a plate, but because they weren't fucking cooled at all.
Wrong, it was a bad batch of capasitors. Gamers Nexus tested this quite accurate.

...

>Wrong, it was a bad batch of capasitors.
>tfw, bad capacitors make your motherboard pull 300 W through an 8 pin

...

How is that even comparable?

EVGA literally distributes thermal pads for the back side of the PCB where VRMs are located and provides instructions on how to install them and have also released a VBIOS version which adjusts fan control in order to keep the MOSFETs cooler. The manufacturer quite clearly explains the issues they found and how they've been addressed on their own, official site.
evga.com/thermalmod/

...

Yeah, it's funny how copper heat sinks for increased cooling performance used to be a selling point and now it's just some abstract aluminium extrusions with a smooth painted surface because who needs heat transfer anyway.

Hahaha, this guy.

...

>90c cables potentially catching fire
>downclock to 1.2 GHz due to VRM temps
>gaymur bling instead of properly designed heatsinks

X299 just keeps getting better and better.

Prime95 is not a valid benchmark. No real world program will ever stress a CPU that hard.

Prime95 wasn't a problem until Skylake started melting faces.

>Multi-core doesn't matter!
>Price/performance doesn't matter!
>Power usage doesn't matter!
>Temperatures don't matter!
>Soldered dies don't matter!
>*NEW* Locked platform features don't matter!
>*NEW* Synthetic loads don't matter!

>Intel earlier: Don't overclock i7-7700K
>Intel soon: Don't overclock i9-7900X

They just had to show that they acknowledged the problem. The temps were fine. Look it up.
gamersnexus.net/guides/2691-final-evga-vrm-thermal-torture-test-and-analysis

Those temps look like exactly what EVGA said, "marginally within spec" in a bad thermal environment, like a hot case. The original cards with the original VBIOS had the absolute minimal yet still adequate cooling for the VRMs, which the update and thermal pads greatly improved upon. Hitting 110C+ PCB temperature because of 40C air isn't great at all since a case with shitty airflow could very well get as toasty as that, even with a single card.

>fan control
>to keep mosfets cooler

I assume they tweaked the cpu fan profile then because the mosfets dont have fans. also goodluck using an aircooler, you need an aio which has its own fan curves makes that change irrelevant

The discussion is about EVGA GTX 1080 graphics cards.

oh, nvm then

God forbid people use demanding software

I've been out of the loop for a while. What other problems has been going on with X299?

EVERYTHING IS FINE

They aren't soldered anymore, this is leading to serious temperature issues even at stock.

Everything
Will
Be
Fire

So how hard would it be to make a waterblock on the caps, Vrm, and Mosfets?
i have a VRM block for looks basically.

Hold up
What's stopping us from making a giant motherboard water block and place it with thermal pads behind the motherboard?

I have a 120mm fan blowing on the back of my CPU socket that reduced temps by about 3C alone.

All you'd need to do is blow a fan on them. But the point is the cooling is not properly designed in the first place.

You'd still have to make sure to get a board with more than an 8 pin CPU power connector.

Pretty sure prime 95 is bullshit, I had a FX-4100 that ran stable at 4.8Ghz but if i tested it on prime95 always got bad results compared to other bench mark tools that actually did something instead of just run at 100%

>pretty sure my anecdotal evidence isn't bullshit

Let's just ignore that practically every review site, every decent overclocker and most stability testing guides tells you to use P95

...

That won't do shit, heat is in a small area the size of a few fingernails.

It need a water jet cutter

Aim a fire hydrant at it?

>he uses hardware with the ME
HAHAHHAHAHHAHAHHAHHAHAHAHHAHAHAHHAHAHHAHA

Make sure not to turn up the PSI too high. youtube.com/watch?v=VfgbYSuZC2U

It'll explode before I connect the hoses, also you need more than 1 person for that job.

Isn't it common knowledge that prime95 shouldnt be used on anything post haswell, causes overvolting or something?

No.

...

lul.

Yeah I got a $10 iTunes GC which is literally useless for me. Intel is shitting themselves HARD. Don't know anyone in the district that uses Intel's iPOS or RXT for selling. AMD doesn't even need a shilling site to get increase their sales.