Clevo Overclocker's Lounge

Let me use this topic to ask a question!
Clevo W650sj with premamod, my gtx 850m still cant overclock more than 180mhz on memory, why? and i cant set power on msi afterburner...its already using premamov1, shouldnt i be able to control voltage and get more overclocking range?

I would like to go higher since with 180mhz + it goes only 65 celsius

 

Rapid does not reflect drive performance in any shape, form or fashion. It has a purpose, but using it with an SSD benchmark is not accurate. It will only measure the RAM, not the drive. The benchmark results won't have anything to do with whether the drive being measured is fast or slow.

Using Rapid or other RAM caching on an overclocked system that might experience crashing or lock-ups is hazardous in terms of file corruption and data loss. Any kind of system hang or BSOD that results in the system being shut down incorrectly can destroy files due to delayed write with caching. If you cache an OS drive, the chances of having corruption that cannot be fixed and leaves your OS unbootable are very real. I have had that happen to me a couple of times with Samsung Rapid and stopped using it because of that.

Below are the benchmarks on my molasses-slow spinner data HDD with RAM caching.

 

MSI Afterburner isn't a great choice for notebook GPU overclocking because it lacks voltage control. It makes no difference with a stock vBIOS that lacks voltage control, so using MSI Afterburner in that scenario is fine. It's limitations are the same as the crappy stock firmware. Use NVIDIA Inspector instead and you'll have all of the control @Prema intended. EVGA Precision X works better than Afterburner, but using it on a laptop with LVDS display (which yours might have) is dangerous unless you use an old version released before Q2 2015, like v4.21. Versions newer that that can corrupt the EDID (firmware) to your LCD panel.

If you want to try this old version, here is a link to it in my Google Drive. https://drive.google.com/open?id=0Bwdqi25LDwZyYjJTRHZVbTBXazA
I am not aware of any laptop having the LVDS LCD panel corrupted by this old version.

 

All right thank you so much, im with the nvidia inspector on my laptop also, uninstalled msi afterburner...will try your evga precision software later thank you so much for this!

i didnt knew about this problem with evga precision X, im surprised =o thanks alot for the info!

Whats your bet on this videocard? how many voltage and memory mhz do you think it usually can take on overclocking above normal limits?

 

You're welcome, friend.

I am not sure how much it can do because I have never had that one before. If you go to 3DMark.com and click search, then select your GPU and choose benchmark, you can see the highest recorded scores and look at the clock speeds that on the GPU when those tests were run.

I like MSI Afterburner. It's just a shame it lacks features for those using a modded vBIOS on a notebook. Otherwise, it's a nice program.

 

Nice! one more time thank you so much =DD

 

Honestly, I cannot imagine that anyone would not already own this game. That's like a rite of passage or requisite to achieving greatness in PC gaming. But, in case someone doesn't for some crazy reason, Doom3 is on sale for the next 10 hours for $1.15 USD.

https://www.bundlestars.com/en/game/doom-3

I'm surprised Brother @D2 Ultima hasn't already posted this. I was emptying my junk folders in Outlook and it happened to catch my attention when I saw 'First-Person Shooter Star Deal' (the only game genre I actually like a lot) in the subject line.

 

I actually have Doom 3 for the shield, it does not do too badly at it. Shows you how far things have come.

 

The latest Doom game is amazing!!! The doom 3 made history, thanks for sharing!

You guys can find me on steam to play something at: http://steamcommunity.com/id/edney_helene/

Ive been playing rising storm 2 vietnam, and doom its pretty sick!

 

Yeah I probably enabled it a long time ago when I installed the OS/drivers. Either way my system isn't overclocked for 24/7 usage, only for benches so should be fine.

That drive actually has nothing but games lol.

 

Doom 3 was like the red-headed stepchild of the Doom franchise, it was more of a survival horror game than conventional Doom. Still a good game though. But I was very happy about Doom (2016) returning to the series' roots.

 

Indeed, Doom 2016 is awesome.

 

I like how they prodded fun at Doom 3 by the doom guy throwing away the screen in disgust as he tries to monologue at the beginning.

 

Yes, I thought that was funny as well.

 

Back on topic...

http://www.3dmark.com/fs/13256664

 

Mwahaha i still lead in Graphics score.. *hides in a corner before John comes back with his beating stick*


In other OC Lounge defiling news, first 2800Mhz on an AW? 3000Mhz boots, but not very stable ...yet. -


And glorious 4.5Ghz -

http://valid.x86.fr/x6egzb

 

Will it actually hold 4.5GHz through a full run of Cinebench? If so, what are the temps like?

 

With the way it is right now. No way. Throttle hell galore.
In 2 weeks it will, and more.
Right now I'm using a 85W heatsink meant for the 7700hq. Once the 138 w heatsink comes in things will be much better.

 

I decided to play around with my 4810mq clevo w65xsj ^^
https://valid.x86.fr/bench/x6asan

 

It corrupts eDP panels too (plus all pascal notebooks use eDP, unless an eDP to LVDS converter is used, like on the MSI GT83VR). If new versions don't corrupt, then the problem may be fixed, but there's no real way of knowing, and someone might just be the "unlucky" guy.

 

That's why I only shared the old version (based on MSI Afterburner, but with more useful features than Afterburner offered) that has never produced an example of EDID corruption. It's not worth the risk when the outcome is uncertain. And, the old version works flawlessly for everything prior to Pascal, so there is no reason to even consider using the newer versions that were associated with the EDID corruption.

Edit: The newest Pascal versions of EVGA Precision X OC seems to be fine. I have been using it on my DM3 and 16L13. The pixel OC utility is disabled and non-functional on laptops on the newest versions I have used. The voltage curve graph seems a lot more useful to me than the one provided by Afterburner. It's definitely more user friendly. The installation order is important. You have to install RTSS before Precision X OC and have RTSS running while installing or you will have C++ library conflicts. RTSS will crash when loading of you try to install it after Precision X OC.

 

@D2 Ultima


Also, see this post... http://forum.notebookreview.com/thr...94-findings-fixes.807276/page-5#post-10575421

 

Is it advisable to apply liquid metal paste on gpu if you always carry around your laptop almost everyday to work?

 

Probably not a great idea, unless you do something special to be certain it remains confined to the die.

I thought I would tax my system for a while (about a half hour) with Crysis 3 with all the settings for everything totally maxed out. Ran it with the CPU @ 4.7GHz and the GPU cores clocked to 1950. It is running pretty darned cool for how high everything is overclocked. Unlike benchmarks, that's 25 minutes of everything being under non-stop load.

This is after delid, liquid metal, lapped Bitspower IHS, thinner thermal pads on the GPUs, using Phobya NanoGrease Extreme on the GPUs, removing all of the plastic sights shield crap on the inside of the bottom cover so all of the grilles are open and using my modded U3 with Vardar Furious fans.

 

I skipped to 18 minutes and your CPU appeared to be at 4.2GHz and your GPUs are 1898MHz. It seems at 86c your CPU cuts turbo and falls to 4.2GHz. But you didn't get any stop errors this time I suppose, that's good?

I REALLY wonder how the CFL hexacore is going to fit in these things haha. Maybe if Clevo releases the next unit with better fans, a vapor chamber CPU portion of the heatsink and doesn't paint the VC for the GPUs around the fin area, then we might get more fight in them =O

 

Yeah, it stayed there to 23:30 then went back to 4.7GHz. I don't know why. I doesn't look as though a change in voltage or high temps caused it and there was not a big change in CPU load either. Weird. But, the GPU downclocking is normal Pascal stupidity. As soon as the master GPU hit 81° (16:53) the core clock went from 1911 to 1898. That's retarded, but by design... thanks, NVIDIA. There was no noticeable change in the game performance with either drop in clock speeds. Had I been using the stock vBIOS it would have been less than 1911 long before 80°C and it's only because I am running a @Prema beta vBIOS that it held GPU core clocks at 1911 for as long as it did.

We can only guess how CFL is going to behave. Based on the trends we see in PC tech today, I am expecting to be disappointed in CFL. I am fully expected it to behave like it was designed by retards.

 

Considering their "innovations" in power management recently I wouldn't be surprised. Unless it has an X at the end of it's name they will likely gimp it somehow.

 

Time to assemble a desktop and go full on LN2 ;D

 

O hay guise

Seems Coffee Lake needs a new motherboard and Z170/Z270 will not support them.

http://www.tomshardware.com/news/intel-coffee-lake-cpu-motherboard-asrock,35147.html

@Eurocom Support does this mean that users need a new motherboard for your Tornado F5 to use the new chips?

 

Sounds like the right move. Mainstream sucks. Always has, always will.

I kind of figured that would happen. In the end itmight not even matter.

 

Only higher OC possibilities save/hold up Intel now. + the OEM/ODM still will buying Intel's BGA chips for their JokeBook's. Intel have to run for their money now. Damn pity AMD's chips don't oc very well.

 

When coffee lake was being designed AMD was still not on the radar so I don't know why people are surprised they are not keeping it on the same socket

 

Considering AMD's continued support into 2020 for AM4 that could prove to be a mistake for intel. Between cost savings and performance improvements AMD has a chance to regain some lost ground if they play it right.
IF CFL can actually provide the theorized 30% performance boost they claim then they might get people to ditch the last chipset and socket. Not very likely though.

 

It'll be 6 cores vs 4; the boosts will be there

Sent from my OnePlus 1 using a coconut

 

Even with or without AMD being relevant, I would have been very surprised if CFL would have used the same socket. And, even if it did, I would expect them to intentionally do some something to not allow an in-place CPU upgrade. They have no financial incentive to allow SKL > KBL > CFL upgrades and doing the right thing for "end users" hasn't been one of their guiding principles for a long time. I put end users in quotation marks because we are not their customers. Their customers are OEMs and ODMs and their customers want to sell motherboards and pre-built systems. So, none of them give a damn about customers being able to benefit from upgrades. If we can do that they make less money, and making money is all they care about.

 

There are other ways to motivate sales... You know maybe, innovate? They have Optane technology for the 270 boards that isn't present on earlier versions that could actually be beneficial. Get partners to start taking advantage there. I hope AMD starts absorbing some of Intel's unchallenged share forcing them to move forward with innovation instead of the anti-consumer practices they have gotten used to. Give us a reason to upgrade instead of forcing us to.

Sent from my SM-G935T using Tapatalk

 

I truly believe that they could have used the same socket/chipset for coffee lake. I remember in the X58 days we only had quad core extremes (bloomfield?) then later it supported the 6 core extremes(gulftown?).

 

About time we start having socketed chipsets as well IMO.

 

That would be interesting. Not likely to happen soon or even at all, but interesting.

 

If they operated on that basis it would be awesome. Extortion and fraud is a much easier business model for them.

I think there is no question that they could if they wanted to. I think wanting to is probably the problem. They won't make as much money doing it the way that is best for end users.

That's a great idea. That would allow for standardization of motherboards as well, kind of how we used to have MXM standards before the creeps at NVIDIA got brain cancer and decided to stop supporting MXM.

 

This turned out pretty good...

Spoiler: GPU Cooling

I cut some nifty foam barriers to keep Conductonaut from leaking out onto other areas of my GPUs. I have Kapton tape covering the GPU components around the die, then put one drop of super glue on each side of the die to hold the foam in place on the Kapton tape. The foam is 1/8th inch and easily compressible so as to not interfere with heat sink to die contact. Temps turned out pretty nice. Never dropped below 1911 on core with the @Prema beta vBIOS. Nice compliment to the Bitspower IHS lap job.





@Prema @Johnksss@iBUYPOWER @Donald@HIDevolution @Ted@HIDevolution @thattechgirl_viv @Zoltan@HIDevolution

 

Nice! Glad you took some good pics. I was just reading about this in an Alienware thread about LM. A pilot who travelled with his rig a lot was worried about the LM leaking out and shorting his MB due to vibration and turbulence. He did the same thing but I believe he did his processor too. Unfortunately the pics he posted did not register on the forums. So it's nice to see some pics of what this would look like.
My 940XM and Conductonaut just arrived today. I think my 10 year old son is more excited about building it than I am. We will be doing something similar as soon as I find a way to get my 765M's to work in my rig.

 

A hexacore at 4ghz is a lot faster than a quad at 5ghz.

Now we have to wait for a new laptop for CFL support. I am not as annoyed over the fact that I cant swap as I am about waiting another couple months. If I must be entirrly honest, I am not at all happy with notebook options lately. P870DM is too big and I dont need SLI at all. The worst of all, its a quad core on a 17inch 45mm chassis.... Seriously??? I like my F5 but I want a hexacore option or a ryzen octacore option.

On the bright side, I could pull the old 7700k and build a LN2 bench rig. Now that would be fun!

 

Wow this looks fantastic ! these are very very impressive temps for an laptop !

Very creative stuff !

 

With the foam I would just make sure it's very light foam that compresses easily or maybe have it JUST slightly higher than the core to make sure it does not interfere with core contact.

 

Would this thermal pad work, instead of the soft foam? as the foam will only handle 85c before it dry's out.

Thermal pad surround (thermal conductive too):
http://www.aquatuning.co.uk/thermal...eitpad-fuer-chipsatzkuehlung-35x35x1mm?c=2868

Soft foam (The max temperature states 85c max):
http://www.ebay.co.uk/itm/BLACK-NEO...hash=item20ecba04aa:m:m0ft2U6TPx3LelsYYVsHwOg

 

The height would be super critical to make sure it does not push the heatsink up.

 

Got you. I have just found this Low Density foam . It is chemical resistance and can handle above 100c:
http://www.ebay.co.uk/itm/BLUE-PLAS...hash=item3d433021e7:m:mInJ9Vq98ivgCVAfNK-hhxw

#Sorted

 

That's a safer bet.

 

Is it possible just to use a thermal pad with a .5 mm or a 1 mm thickness? What is the height of the gpu core with respect to the board?