Investigating 6820HK performance and overclock!

I'm pretty sure this asus has a limit of 75w on the CPU

Good headroom, considering that if you don't repaste, 65w is already too much heat to handle (this is with the bottom fan cover OFF)

 

No, actually it doesn't. I have had it as high as 91W so far, but it was not running optimal either.

 

I'll PM you the bios dump shortly

 

I am going to PM a link to something you might find useful. Not sure it will work on that machine, but it may let you write some values to SMI that are higher that the power limits available in the BIOS right now.

 

It's definitely hard limited by the bios to 85w 8sec and 75w indefinite / 4096sec

I think the alienware machines don't have this issue due to the partially unlocked bios -> http://forum.notebookreview.com/thr...-settings-for-4-0ghz-to-4-3ghz-stable.787791/

 

Yeah, but limits are often artificial. Just because that's as high as the numbers go in the BIOS menu doesn't mean it cannot be overridden using a bigger hammer.

 

Maybe is put as a protection, the cpu vrm does not come with heatsink. Only black sticker with copper plate on top.

But you're right though, for majority of user using this machine without liquid metal paste, even 68w is hard to keep cool. Especially without removing the bottom panel and allowing fresh air into fan.


NBFC doesn't work with G752-VS btw, the older maxwell one can be fan control with Zenbook UX32 profile.







Bga cpu uncork tool(s)

 

The underlying problem is frequently a situation of mistaken identity. Many people buy things based on assumptions that because something is sold by brand A or brand B it is automatically going to be awesome. That happens often enough that it is not even necessary for products to be good to sell like hotcakes. If it were properly engineered, there would be no basis for concerns about thermal management.

 

But not many people have an agnostic attitude towards branding

 

That's really unfortunate and it is probably one of the reasons we have to tolerate so much pathetic garbage from so many directions. Results matter. Brand does not.

 

Well half of the issue is the fact that people prefer form over functionality. Like 80% form and 20% functionality. If it even came down to 50/50, things would change drastically.

 

That too, but the brand thing is more prevalent than you'd think. Look at Linus' forums under laptops for example. Guaranteed, if somebody asks for a laptop (*ANY* KIND OF LAPTOP) and I, Dackzy, don_svetlio or Pendragon don't reply first, there *WILL* be somebody recommending a:
- Razer Blade
- Dell 7559
- Acer Nitro
- ASUS

and ignoring basically everything else. It's because they get good reviews all around and people don't look. Brand is almost everything to anybody who doesn't specifically dig in laptops. Even if they're the biggest tech nerd of the desktops and live in custom loop watercooling who'll grab a BIOS machine and dump a Titan X Pascal vBIOS, mod it, flash to new chip, then replace them, they'll STILL know nothing about laptops in general.

Then when on top of this you add the 80% form 20% functionality to the equation, and the fact that reviewers don't ever actually point out throttling etc and state that the machines work fine & acceptable for laptops, and then people get this wrong idea that those ill-suited form factors are actually acceptable (and worth their ridiculous prices).

 

Agreed. I lurk there every now and then, but never dared to post. Some of the suggestions are just too obnoxious .

 

Yeah. And their reasoning is "oh, Linus reviewed it and said it was good" or "Oh, JayZTwoCents said it was good" or "look Notebookcheck rated it high" etc when NBC is only useful for raw data (like temps or TDP limits) because they rate like, outward feel and looks high and stuff.

It's dumb.

 

I like how you just say Asus. And people really recommend that for non gaming?

Edit: why has Linus never clearly talked about clevo? Like he reviewed a P650SE but never said that and just said Venom Blackbook 15 even though he loved it, he made it sound like some 3000 dollar exclusive thing, but Venom just has terrible pricing.

 

Because he never deals directly with Clevo, hes provided systems by Sager/Metabox/Venom etc.

 

Still the only Sager I can think of him addressing was a P870DM, never any other models like the P650 P750 P670 P775 etc...

 

Because he really doesn't care. He's said more than once that DTRs or high performance laptops aren't his thing, and he likes his razer because it's small and handles the light loads he uses on it.

When I called out the Venom Blackbook review being a Clevo and I was sad he didn't point it out, he claimed how "everybody knows that already" and said it's just like how everybody knows power supply X was really a company Y rebrand (which I didn't know actually, driving home my point). He basically won't say "this is a Clevo" and is insistent on leaving it on whatever the default is (Sager is Sager, etc). He's reviewing the P870DM3 soon because Sager sent him a review unit, but... I don't expect too much from it, really.

 

Get ready for the fan noise and high temps bashing.

 

https://en.wikipedia.org/wiki/List_...lake_microarchitecture_.286th_generation.29_2
http://ark.intel.com/products/88200/Intel-Core-i7-6700T-Processor-8M-Cache-up-to-3_60-GHz
https://www.amazon.com/Eluktronics-Business-i7-6700T-Professional-SuperMulti/dp/B01C2N2C7G
There's even a locked 6820HQ using more power than the above processor.

6820HK ( unlocked unless locked in bios? / uses more power than LGA? )
TDP 45W at 2.7ghz
igpu at 1050mhz
Turbo 5/7/7/9 or 4/3/2/1 core modes at 3.2ghz / 3.4ghz / 3.4ghz / 3.6ghz

6700T ( locked ) May be slightly thicker due to IHS/Socket
TDP 35W at 2.8ghz
igpu at 1100mhz
Turbo 6/6/7/8 or 4/3/2/1 core modes at 3.4ghz / 3.4ghz / 3.5ghz / 3.6ghz

 

Interesting. My 6820HK is a gas hog, but I'm able to undervolt quite a bit.

Powered by: Octa Core Exynos + 6820HK

 

Voltage is just high on the mobile.

 

It should have been the opposite with a mobile . But Intel doesn't care

 

The 6700T is a 35W TDP OEM-variant chip. It's going to be for business-class machines in general, and needs to stay within that 35W spec. So they'd make sure voltages were low.

Laptops just get screwed, as we know.

 

That chip should've been the 6700HQ lol...

 

The 6820HQ should've been the 6700HQ.

And in fact, they all should've had a minimum of 3.3GHz 4-core boost with most of them being 3.5 or 3.6 at base. This is ridonkulous. *rage seeping through*

 

all of them should not exist except the 6820HK

 

At the prices they charge... yup

 

For some sleek laptop, 6820hq should be fine.

 

The base specs (excepting overclockability and intel technology support) have the same specs.

 

6700T is 38% more power efficient clock for clock vs the 6700HQ.

303 USD vs 378 USD and you could upgrade to a 2.9ghz 7700T at 35W TDP

 

7700t is out already?

 

Not yet but that's what they will release by end of this year supposedly.

 

http://www.coolaler.com/content.php/6793

The new desktop Xeons out in a few months

The E3 1280 v5 Skylake is 80W TDP

http://ark.intel.com/products/88171/Intel-Xeon-Processor-E3-1280-v5-8M-Cache-3_70-GHz
3.70ghz base clock

The E3 1280 v6 Kabylake is 74W TDP
3.90ghz base clock

http://ranker.sisoftware.net/show_s...e8d5e5c3ab96a680f8c5f5d3b6d3eedef88bb68e&l=en
Z270 board in testing

 

Confirmed that the 75w limitation is BIOS issue, the Alienware machines can go up to unlimited CPU power usage through bios override.

 

I update the first post on how to run the 6820HK properly


Please note that you cannot have XTU and ROG Gaming/Command center installed at the same time

 

Told ya man. ASUS wants you to use their stuff ONLY. It's part of their proprietary crap. Everything has to be ASUS with them.

 

It's not uncommon for tools that use the same sensor access to conflict, that's why I always remove the MSI/Asus control panels that affect the CPU/GPU before tuning with XTU/TS/AB. Clevo has their control panel too, which most tuners remove as well.

 

The problem is when both start automatically, they conflict.

Clevo command center uses xtu to interface with bios iirc, unsure on the asus cancer center / msi dragon gaming utility

 

Except XTU doesn't work properly by itself for ASUS. I rather just use Throttlestop.

 

Weird, I used XTU for years on Asus laptops, what's changed?

On what CPU's is there a problem, what is that problem?

 

Actually there's no problem, XTU actually works in changinig PL1/PL2 whereas throttlestop doesn't

 

I was more talking about the amount of options you actually had rather than the functionality. I guess I worded that badly. Sorry.

 

Talk with @unclewebb

 

The problem is PL1/PL2 is hardcoded to bios, johnksss tried to uncork it a few days ago without success.

 

Then that's normal... throttlestop doesn't bypass the BIOS. Just like XTU won't, etc.

Unless throttlestop can't get the TDP to go above 45W/56W?

 

XTU and ROG cancer center can up to 75W/85W

Tstop cannot... even for negative adjustment

 

Mmm... That one is odd. You should talk to unclewebb and see; anything XTU has access to, TS should have access to.

 

Anyone with a 6700K try running 2.8ghz and 0.900 volts? Can you check if the TDP is ~33-36?

 

2.7ghz 6820HK @0.9v = 30-32w