Clevo Overclocker's Lounge

It even works on my desktop. Not sure what's going on there.
Also for higher refresh rates you can't just use auto, you would need to start making changes to get anywhere from 5 to 30 or 40 mhz higher. Based on screen of course.

 

It's got me scratching my head because I haven't even gotten to the point of needing to make manual adjustments. Dialing in 120hz and testing passes, stays for a 5 seconds after I choose to keep it. Then blinks back to 100hz and doesn't save the setting either.

 

I think that means it's really not stable, but not 100 percent sure.

 

You can get a power brick with a molex on the end delivering 5v and 12v.

 

You sir need to start using quotes.

 

Between what exactly? 6700k vs 6700hq? Probably around that at stock. 3ghz vs 4ghz would give you 33%.

And the second question is, is a single 6700k enough?

 

I was tinkering with an 880M earlier today in the P750ZM. Here's a pretty decent 3DMark 11 for 880M. Second place worldwide, right behind Brother @johnksss top score with the same GPU. I stole second place from svl7.

http://www.3dmark.com/3dm11/11378132

 

how do you browse scores by just gpu type?

 

Use the search function and the Advanced filter and you can search by both CPU and GPU. With single GPU search there are often too many scores and you have to set the minimum and maximum, or select a CPU to narrow it down or you get the "Ooops" message with no results when you try to search

 

Finally delided my CPU. The short term tests had a decent drop, the long term test did not...

Stock everything, 74F ambient temps for each test. Previous was my stock 6700K with CLU and secondary was delided 6700K with CLU.

I've also added some pics of the delid tool.



Procedure:

Spoiler: Step 1

Spoiler: Step 2

Spoiler: Step 3

 

Umm, whats the clock speed?

EDIT: Nvm, its stock. Dont you think its a bit too hot for stock speeds?

 

Nice! Looks like the tool works very well.

I agree with @blookhawk, we need to know what clock speed.

If that is 4.7GHz or higher the temps are very good. You're running on hot summer air, right... so the wPrime 1024M test is going to be too much unless you're ins a chilly air conditioned room.

 

Stock clocks (4.2Ghz on all four cores).

This is stock cooling, fan speed 100% nothing else.

 

Well, they're still really good temps and a massive improvement. It may be that wPrime 1024M takes so long that the heat sink is getting saturated and cannot expel the heat fast enough to keep up. What is your ambient room temperature?

Nice job, bro.

 

Here is a crude run i just did @ 4.5Ghz (my daily driver for summers) (Ambient @ 78F)

CLU on DIE -> CLU on IHS -> Copper SHIM -> CLU

You might want to double check if the IHS to Heat Sink contact is even, after de-lidding, the height of the IHS goes down and that offsets the VRM end, causing a gap.



Dont mind the score, i have a lot of stuff open in the background.

 

@bloodhawk - Your temps are really nice. Are using your modded cooler? I think @ssj92 is at work and it might just be sitting flat on a desk.

@ssj92 may benefit from adding a shim, but the temps other than the wPrime 1024M are really good.

 

Nope, sitting in bed watching overwatch

The temps are even better when i turn on the central AC at night and keep the ambient around 60F.

Working on a Saturday! :O

 

@ssj92 I think 9 degrees difference between coldest vs hottest cores is in the upper level. Have you checked your heatsink for uneven/warped problem?

 

There is a good chance it might be uneven because of the VRM side being a bit higher.

 

Probably. The difference has been slightly larger after the delid. 9 degrees difference is too much.

Edit. @ssj92 Should take a look on @Mr. Fox testing https://biosmods.wordpress.com/2015/10/21/sky-x9/5/

 

If someone is still confused, something like this happens :





The figures are approximate and for illustration only.

 

I'll check later but it probably is. Even before the delid there was a decent difference between cores. My P570WM had a warped heatsink as well.

But the thing is, I didn't remove the glue at all, I left it exactly the way it was and placed it back where it was as well. Guess I'll mess with it later when I have more time.

 

Awesome diagram. Something that hadn't occurred to me.
I haven't delided yet. I plan on doing it very soon, just waiting on some a new HS and a deliding tool.

The HS I had was shocking.

 

Nowhere near as bad as my first heatsink on my zm!!! That was bad!

Sent from my SM-G900P using Tapatalk

 

Yeah seem to be raised a bit towards the VRM's. Will be more raised after de lidding.

 

You may want to consider ditching the pads and using a very small amount of thermal paste to help compensate.

 

Was actually thinking about that, might give it a shot tomorrow. Only problem is that one of the components is slightly lower in height.

 

The other option is to re-shape the heat sink a little but you may need proper tools for that. 0.5mm/1.5mm pads is also an option, it all depends how big the adjustment was.

When doing something similar I will have to get calipers to figure out how much things are moving.

 

Im not gonna mess with the heatsink for now, since im working on getting a cold plate attached.

 

Fair enough, it's always better to do things in stages if you have plans

 

Here's a new record in 3DMark 11 for single 980 MXM. No AC cooling... in my lap in the recliner this lazy Independence Day morning. Happy July 4th to all. To our Veterans, thank you all for your sacrifices to ensure our freedom and safety.

http://www.3dmark.com/3dm11/11384452

 

Nice to see some bench again... I sitting here in Canary Islands outside Africa with my phone and not my beautiful beast. I REALLY MISS BENCHING!!! Good 4th July to all my friends And happy benching.

 

I have been studying a variety of 6700K XTU profiles on HWBOT and found some interesting differences. My XTU benchmark score is amazing, but I wanted to see how my setting compared to those that had a little bit better results. What I find is that it seems Skylake is extremely inconsistent. The variance between settings that work well among all 6700K processors is nothing short of ridiculous, even among some of the most expensive desktop enthusiast motherboards. Often those settings not even close enough to call them similar. Anyhow, tinkering in the BIOS based on what I found to experiment with, I am able to replicate my previous Cinebench R11.5 results with entirely different settings and slightly lower temps. This is using adaptive core and ring voltage, core 25, turbo 25, +25 offset, with PLL set at 5, and 45x on ring set with the same voltage. Again, this is just fans blowing, sitting in my recliner this morning. No AC cooling.

 

The smaller processes can be quite fickle, note how no LN2 scores are with the smaller nvidia chips at the moment, they have had a hard time getting them to scale.

 

'The Independence Day Bench'!

 

AMEN! Can I get an "ah ha" people? FREEDOM! Freedom from crippled firmware and BGA filth. Hallelujah!

 

Just a thought, @Mr. Fox , Has anyone beta testing for @Prema ever run BITS as part of their testing?
https://downloadcenter.intel.com/download/19763/BIOS-Implementation-Test-Suite-BITS-
https://biosbits.org/

 

Never heard of it before. I'll look at it. Thanks.

 

great score, updated on leader's board.

why does your GPU-Z say your GPU is a 1080M??????

 

More and more impossible. OEM's throw out Mainstream laptops marketed as high end machines today. And buyers of this garbage do not understand that they are being cheated. We who still appreciate proper hardware / firmware is soon only a minority and will be gone..., Tablets / phones and APPLE similar laptops is the new high end and performance standard. And stupid buyers of the newest junk have not yet understood that they have been duped!! Whenever they will understand... It's to late.

 

I'm not sure. It shows up as 1080M randomly. It showed up as 1080M on something else recently, but I don't remember what application that was. GPU-Z knows better. Maybe because the specs are close to the same and it's just confused. It will be interesting to see if 1080M is actually a better performer or not. It will certainly be better than 980M, but maybe not anything to get too excited about if you already have 200W GTX 980 MXM.

 

My GPU-Z 0.8.9 says GTX 980

Maybe your GPU-Z took a little bit of white horse

On a side note, I contacted the developer of GPU-Z telling him to stop with the 0.XX version numbers already! it's been ages since this great little program is out and it's time it go out of the 0.XX range makes it sound like a premature program that was just developed last month. like a beta

 

Well, that's under Windows 7. Under Windows 10 it's showing 980. Let me check. Maybe I still have the modded driver installed in Windows 7 from when I was playing a practical joke on everyone a while back. I modded the INF and Prema matched the core and memory clocks to 1080M specs. I'm not using that vBIOS any more, but I may not have reinstalled the driver without the INF tweaked to spoof 1080M.

 

speaking of VBIOS, are we gonna get any new VBIOS for the desktop class 980 GTX?

I am still on this:

 

I cannot think of any reason we would need one at this point. What we have is amazing.

 

4.6Ghz benching at work lol

 

There should have been some information about 1080 in laptops a while back in the big ol' Pascal thread.

Am unsure how much it dictated, and am really curious on how good would 1080 compare to 200W 980 performance wise.

EDIT :: Found some info. 1080 for laptops reaches 980 ti levels...

 

Sorry...No laptop benching at the moment....
http://www.3dmark.com/hall-of-fame-2/fire+strike+3dmark+score+extreme+preset/version+1.1/2+gpu
http://www.3dmark.com/hall-of-fame-2/fire+strike+3dmark+score+ultra+preset/version+1.1/2+gpu
http://www.3dmark.com/hall-of-fame-2/3dmark+11+3dmark+score+performance+preset/version+1.0.132/2+gpu
http://www.3dmark.com/hall-of-fame-...+score+performance+preset/version+1.1.2/2+gpu


Happy Independence day!!

 

Made the Linux USB disk, worked OK and I ran all the tests, but none of them really meant anything to me. I think the way I have my system configured (ACPI disabled, ASPM disabled, LPM disabled and C-states disabled) may be throwing things off.

Was there something you had it mind that I should look for using this? Maybe I just don't know what I am doing.

 

I noticed skylake was not officially listed on the webpage. What my thought was to run the tests, including using the boot to os through the BITS thumb drive option so it gathers info under stress testing and with use of TS and, if you are willing to mess with it, XTU, as well as the standard bench portfolio, then only run a report of failed tests over to Prema. He'd likely know more on the error data than us (obviously not belittling your knowledge on the matter). Just figured, if it worked, could give new type of feedback to him. I do know 4790K is supported, but that means some things might not show up with skylake if the program isn't there yet. I'll see if I can find more info on how to use it for certain things later today or tomorrow. ..

Sent from my SM-G900P using Tapatalk