Clevo Overclocker's Lounge

Nope

 

In America it's a slang for...

You are the s*yt.(poopoo)
or
You are the cats meow
or
You are the man!
or
You lucky dog!

All meant to be in an admiring way of someone else.

 

Nope. I do not glue them back. You could put a bead of RTV silicon around where the original seal was, but you need to keep pressure on it until it cures completely to avoid having the IHS lifted off the die. I leave it loose to make life simple for myself. If you glue it back then you have to delid it again later if you want to replace the paste. I just put it back in the socket and clamp it down. You have to hold your thumb against the IHS to keep it from sliding forward when you cinch down the latch, but that's not difficult.

 

Or, if you are the parent of a child in diapers, it's the stuff you have to clean up that smells really bad. You call it p00p00 because it's just not right to teach young children to say the "s" word.

 

For those looking to upgrade older Clevo and Alienware laptops to 1080, the GPU that @Eurocom Support is in the final stages of developing for the Tornado F5 may work, so this may be encouraging. It might breath new life into amazing machines like the P570WM, P870DM-G and M18xR2. Using a Tornado F5 1080 heat sink as a starting point, it may be possible to transplant the heat pipes and radiators from your original heat sinks. I know @Raidriar was interested at one point. @Eurocom Support has indicated that they are willing to offer a special price to these that would like to test - one person per platform.

Here are some photos of the new GPU (with others for perspective) they are about to release for the Tornado F5.

Spoiler: Eurocom Tornado F5 Heat Sink

 

I would be willing to test the 1080 out for the P870DM-G.
Just depends on how.much modding is required on the heatsink end. I do have an extra one, just don't have the tools at my disposal ATM.

 

Might email them to ask for details. I shared as much as I know at this point, but it would be great to get it working. If you could get the old heat pipes and radiator assembly off without destroying everything, a 200W 980 or 980M heat pipes could be attached to the Tornado F5 heat plate with a high quality thermal epoxy. Maybe using copper shims and thermal epoxy to glue together all of the heat pipes and keep them together as an assembly would allow desoldering the pipes from the heat plate with the rest of the heat sink submerged in water. It looks like no external power source is required with that 1080 variant.

 

Yah that shouldn't be that much of a problem.
The only thing that troubles me is the probable power limit. Specially if the GPU can pushed beyond stock.

 

That's definitely a potential limit, but if it overclocks some and soundly defeats a single 200W 980 with its max overclock it would still be a winner for those that are not able to buy a new machine. For those with 980M or 980M SLI, or something even less than that, it would be a compelling upgrade for a machine that would otherwise be headed for the recycle bin with no upgrade path of any kind.

 

Count me in for the P570WM!

 

Definitely agree. Any idea how much power draw the mxm slots are rated for?

 

I don't know the official answer and the technical spec for MXM is nowhere close to accurate any more.

I can say that I've pulled over 200W from MXM slots using 980M in the P750ZM. P570WM and P870DM3, so probably more than this GPU will be able to require.

 

Think from 150w up to +-200w for maximum peak power(175-180w maybe for the average). If wiki're right, it should not not be a problem with this Gtx1080 https://en.wikipedia.org/wiki/Mobile_PCI_Express_Module

 

I've seen the 86 x 105mm fit in my model with slight modding... but a new heatsink would be required for the 1080. I wonder how that would end up being? 1080 SLI with a 4930MX? I'm pretty sure entirely new heatsink redesigns would be required, but that would be an interesting scenario eh?

M18x R2 I don't believe will work due to eDP being required, but AW17 R1 should (the 120Hz model) and possibly M17x R4 as well.

 

The mounts for the core contact and the PCB to motherboard attachments have moved on that 1080. It looks like any heatsink would need significant modding to fit, if you are careful you could get away with a thermal pad acting as the support for the card to motherboard part.

If you could pull it off then even at somewhat lower clocks than the clevo 1080 it would still do pretty well.

The issue is that with the core moving even if you borrowed the contact plate from the 1080 native heatsink the heat pipes would not be the right shape from your old cooler so you could not do a transplant that way.

 

If you glued together all of the heat pipes very securely with large copper shims and thermal epoxy (to keep everything from getting torn up after desoldering the heat plate) you might be able to tweak the pipes on the heat plate end. It looks like at least two might just about line up between the screws over the GPU core and the third one might be able to go on the outside of the spring arms instead of in between, without having to bend the heat pipes very much. It's hard to say without all of the parts laid out for examination to study things closely. Getting a heat plate that fits the GPU and radiators that are not destroyed would be the most important things. With those components one could buy new heat pipes and build their own heat sink setup.

The 30mm square shims I used to mod the 980M heat sinks for the 180W 980 are wide enough to span 3 to 4 heat pipes.

 

Yes if you had the ability to get your own heatpipes that would be the way to do it, take the contact plate from the original system and the heatsink fin array from the current one and then match them up with new heatpipes.

 

@Prema @Mr. Fox @bloodhawk @Papusan @infex @Donald@HIDevolution @Johnksss@iBUYPOWER @John@OBSIDIAN-PC @Shakeeb Anjum @Zoltan@HIDevolution

I lowered my overclock from 4.6 to 4.5 GHz but bumped up the Cache ratio to 45X as well and running XMP Profile 1 thanks to Mr. Prema
GPU = +150 MHz Core / +300 Mhz. Memory
GPU Fan offset = +10%

Cache voltage offset = +10mV
Extra Turbo voltage = 0
Extra PLL Voltage = 0
Uncore voltage offset = 0
Power limit 1/2 = 300,000
Power Limit Time Window = 1 sec
VR Current Limit = 8191
Native PCIe Support = enabled

Check the max temps

http://www.3dmark.com/3dm/17081227?

 

Do bear in mind that the GPU is not stressed for very long in 3dmark so you don't get anywhere near thermal equilibrium.

Nice physics score though

 

It has a more direct specific meaning in the context it was given:

You one lucky P00P00 playing with a "Hacked Client", as in the Prema hacked BIOS / EC / vbios.

 

Wonder what I am since I can do it with a "Stock Client" And still beat 98 percent of the people....

 

Something seems off. Even without the Prema bios mine scores higher, without throttling. I'm on the 04 EC I think it is, stock bios.

 

Unusual question, but what do you do to make sure that Windows doesn't cause any interference with the bench?

Could be the settings , as well as the fact that your temperatures might not be as high, resulting in sustained clock speeds on both the GPU's.

Seems decent, but im sure the CPU temps will be way higher under sustained load. The physics test only pushes for a brief amount of time.
If you are using Afterburner, can you take a screenshot if the Core Clock graph? That would help a lot in showing how much the clocks are sustained.

 

HWiNFO is in the background that's why. Without it I get like 2K higher score

 

I would think it will take time and a new perspective to get the most out of the Prema modified firmware, vs stock.

These results are early and unique, when everyone that's gonna get the Prema mod's starts playing around, we should see even better results.

@Johnksss@iBUYPOWER - aren't you getting the Prema mod's?

 

Now that the Skylake modded firmware is close, will we see a quicker turn-around for the Series 200 motherboard / Kabylake modded firmware?

 

Sure. I'm testing a Premamod, but the fact is still the same as I mentioned earlier.
A good based stock bios/ec/vbios are capable of great things while the rest of the community wait's on the better solution. The problem is...Having the right stock solution to start.
And when I opened my box and started benching it. It did not come with a Premamod. It came with a stock Sager Bios/EC/Vbios with no throttling whatsoever. And that is the main complaint here. So as long as people have that part 95 percent covered, then the waiting doesn't seem as long.

I do nothing. I just run it. You tend to gain more problems trying to tweak windows than if you leave it alone for the most part.

 

You will laugh, but that old shipping version with little throttle was an actual 'bug', which Kapok factory issued a patch with 'immediate forced system integration' for and Clevo pulled from their servers...we probably won't see that happening again with up-stream development...they had no official throttle-less version last gen, this gen nor...

 

I don't understand?

 

It didn't throttle because it was unintentionally broken...throttle values are all in place but didn't take due to a bug in another area. Wasted a lot of time trying to wrap my head around this...

The only thing we can expect in life is the unexpected. Learned that the hard way...

 

Hummm, interesting. Glad they gave me the unintentionally broken one. Twice.

 

Wait. Wait. So what you're trying to tell me is... the working stock sBIOS is working because it's bugged, and the others that throttle and are pretty much useless to a power user like me isn't bugged, and is so by design?

*sigh of the gods*

 

A few things jump out at me... (1) not a noob; (2) not obsessed with the absurdity of under-volting; and, (3) willing to try.

It really begs the question of whether they are stupid, evil, or both. If the throttling actually was induced intentionally, I think the only logical conclusion is they are filthy and rotten to the core and stupidity follows as a natural consequence of the evilness leading to having a reprobate mind... an inability to discern a difference between right and wrong.

 

You know, I like, I have very strong logic-based opinions and I'm very fair and pro-consumer in my thought process. I usually don't go with this "evil" thing. But holy meowmix, I'm sick and fed up of 100% of laptop manufacturers screwing the consumer. This is ridiculous. Why does everybody go through the trouble of designing a good framework then decide to cripple it with firmware? It's the most counter-productive thing in the universe.

"Let's make this really awesome thing, pushing the boundaries of what's available in the world today!"
*7 months later*
"YES! SUCCESS! WE'VE DONE IT!" *celebrations*
"Okay, now let's code a BIOS where it doesn't do what it's supposed to, and pretend like it doesn't happen."

^ nVidrosoft, BGAtel, ASUS, AMD (polaris), Lenovo, Clevo, MSI

"Let's make portable stoves!"

^ Razer, Gigabyte, Alienware, ASUS

"Let's make utter garbage!"

^ Dell, HP, Acer, Fujitsu, Toshiba

 

That's exactly what Alienware did before they went completely brain dead with the BGA feces... released some phenomenal hardware that was emasculated with cancer firmware.

Maybe the rationale behind it is if the biggest fish in the pond (Alienware) can get away with these crimes, why not the little fish, too? The bar got lowered, bigtime. I guess as long as they can clear it without knocking it off the peg they'll take it as a win. I certainly don't consider it a win in light of the competition. I think we'd all prefer to see a bloodbath and annihilation over a simple win; which really isn't very special at all when you are competing against impotent toys made for people that have very low performance and functionality standards.

 

Getting there slowly.

http://www.3dmark.com/spy/984008

 

Derp :
http://www.3dmark.com/spy/983830

 

managed to get the notebook 1070 to perform like the desktop 1070.

 

98% even of the other enthusiasts is being a little generous to the other enthusiasts

 

With the original bios the register throttling can be avoided by not turning on TS.

Happy New Year to everyone too!

 

You mean if I don't use throttlestop it won't throttle?

*exploded brain*

 

happy new year to @Everyone, may your temps be low and your clocks high yadayada .

wanna share the secret with me maybe , my card even does what it want so far even with locked voltage in the curve editor, but at least i saw some 1706mhz thats better than nothing i guess

 

Nice! Got any benchmarks to share? How does it compare to the 200W 980? Did you sell the 980 to offset the cost, or holding onto it as a backup?

No, definitely still use ThrottleStop. You can use it for on the fly overclock profiles and other settings. You have it running but not turned on (green system tray icon instead of red).

 

I have also helped 100% of the enthusiasts as well.

 

I can change overclock profiles without turning throttlestop on? That's interesting... however I would be keeping it on anyway because I want my Windows to remain at 1ms timer resolution essentially 24/7

 

Run this as admin in command prompt and you'll have that without any need for software...

Code:

bcdedit /set useplatformclock true

If you change your mind later...

Code:

bcdedit /deletevalue useplatformclock 

 

Isnt this value related to HPET?

 

Yep.....

 

And, it's not clear whether HPET helps or hurts W8.X and W10 because Micro$loth has never fixed their buggered up OS RTC. HWBOT requires that it be enabled with pre-Skylake CPUs because the Windows 8.X/10 RTC is unreliable, but enabling HPET lowers some CPU benchmark scores with W8.X/10. I am guessing the "better" scores with HPET disabled are due inaccurate results, hence disqualified by HWBOT for those running an OS newer than Windows 7 (with a screwed up RTC).

HPET Effect WITHOUT ThrottleStop Timer Resolution
(Better score, but is it actually an accurate score with HPET disabled?)

HPET Disabled WITH ThrottleStop Timer Resolution Maxed Out
(Even better score, but is it actually an accurate score with HPET disabled?)

 

Finally got the machine booting at 3200 Dual Channel. Not 100% stable yet, but at least it's working.