Clevo Overclocker's Lounge

i didn't even notice an other power connector on the mainboard yet ^^ but then again i was never looking for one yet haha will check this out as soon as i get home!
thanks for your always so super usefull infos which i have no clue where you take all of this knowledge from haha ;-)

 

1. Using 362.00

2. Gotcha! This games pushes the whole system quite a bit.

3. The power cable is already plugged in.

4. Always using the Max fan profile when overclocking and gaming.
I figured that something else was over heating on the board, do I put in new thermal pads (Fujipoly/Phobya) and made sure that every component makes even contact with the heatsink.
After doing I actually had core temperatures drop by about 8-10C. But inconsistent black screen still persistent with anything over 110Mhz on the core.
Odd thing is that the black screens sometimes happen even when the core is running rather normal 65-70C and sometimes only after they exceed 85C.
I do have a feeling it's probably a problem with a MOSFET or Power phase , because a friend of mine had a similar issue with his TITAN X in a desktop. He would get very inconsistent overlocks and the system would bakck screen sometimes at stock speeds when doing CUDA based heavy compute work.

 

Have you the same opportunity as other... Set up the machine on top of an AC and test with better/lower temp on hardware?

 

Hold on...a second power port. Is it to possibly power two high power cards(like 980SLI)?


Sent from my Micromax AQ4501 using Tapatalk

 

Sure I can give that a shot tomorrow once I'm back home.

 

I added in the missing circuitry in my 980M rather than changing the chips themselves.

 

Wew. That must need some serious equipment .
But what was the circuitry added for ?

 

To add stability at higher voltages I added in the missing regulator chips. I got a small shop in the UK to do it as the soldering needs to be perfect. It will increase power draw at stock but it stopped the cards shutting off at a high voltages/clocks.

 

Ahhaa. Well let's see if I can get mine replaced. If not, then I'm gonna eat your brains.

 

Hopefully not literally, I do use my brain occasionally.

 

Well Ted worked his magic and Vivian sent me a return label. Gonna send the GPU back tomorrow.
In the end if you explain your situation to them, they always make it work.

Btw . @Phoenix ,zoltan linked me to your driver thread with the warnings lol, asking if I had installed the latest drivers. XD

 

Maybe they should sent an harsh email to Nvidia and told them that they create garbage drivers who kill graphics?

 

Looks like HIDEvolution premier support being reported again, I am taking note trust me.

 

Talking about using the other power port, not the one it is using right now...unlike the ones on the GPU these 2x2 additional power phases on the motherboard are NOT cooled as they should and swapping over to the other 2 two may safe you a lot of time as it may not be GPU related at all.

 

i noticed that i have only problems in the division and firestrike. i can run any benchmark with like +100 / +190 / +0 but in firstrike i get screen flickering (as if you quickly shut your eyes) after the gpu reaches some hotness. and if i remember right (i was drunk last night lol) even adding like +12.5 doens't help with it. i guess the already mentiond crappy quality vram applies here xD

 

hmm, mine is plugged in underneath the pipe connecting the heatsink of the gpu to the heatsink next to the fans. are you talking about this one or an other one? sorry man, a bit confused.. ;-)

 

Is their anyway to monitor the VRM temps I keep looking for it in HWInfo but I just cant find it.

 

i had this option once on the 670m's in my alienware m18x-r2. never seen on any card anymore after that. but i really am now too curious anymore about this since...the vrams on those cards always went up to like 100 deg. c.. so i was mostly playing games just staring at the vram temps hoping the whole machine won't just melt.. ;-)

 

Oh my bad, i couldn't find any other power port like the one which sits right above the heat sink. Ill take a closer look today. But since they are not cooled appropriately, wont it cause issues in the long run ?

 

There are two connectors (blue) and 4 power phases (purple):

Spoiler

@bloodhawk
Swap the connector to the other port and see if that stabilizes the system. Ideally the 2 power phases in use should get a copper block on them for cooling.

 

As I said earlier. Remove the black covers and use a mod U3 with good 12v fans. Cools the entire motherboard much more than without a coolerpad.

 

I have the fans, but haven't been able to find a decent 12V DC fan controller.
And the black covers were removed long time back broski.
That is one of the reasons i dont want to send the whole system back to HID, cuz they will freak out xD (i kinda went Freddy Kruger on the foil cover thingy as well, looks neat though)

LOL no idea why i never noticed the other one. Ill give it a shot today afternoon.

 

I have no fancy 12V DC fan controller, but a very old power supply+diff resistors I use for the fans. Thought to buy one but things take time here at home, LOL

 

Cool! Im actually thinking of turning one of the swiftech 8Fan splitter into a controller.

 

1) The built in GPU Overclocking tool of my Clevo Control Center runs automatically at startup, I haven't overclocked with it but how to disable it and not disable the actual Control Center from running at startup?

2) When I am applying the below overclocking in nVIDIA Inspector, are they saved permanently or do I need to run nVIDIA Inspector everytime I restart? Also, what if I close nVIDIA Inspector after applying the overclocking settings, does it still need to be running in the background?



@Papusan @Mr. Fox

 

Make shortcuts of your different stable Oc profiles from Nvidia Inspector(you don't need to start Nvidia Inspector every time you Want to OC). One click and you have the preferred OC. It's very good thing that you start up your machine every time with stock clockspeed on graphics And forget the Oc bloatware in CC

 

Right, so I need to double click on the shortcut every reboot and then the OC sticks until next reboot? nVIDIA Inspector doesn't need to run in the background?

Also, you didn't tell me how to disable the built in GPU overclocking tool

 

Double click your preferred clock shortcuts if you want to OC. I haven't bothered to look for solutions to disable the bloatware. I nearly never touch that CC Bloatware

 

Thank you doctor

 

You're welcome!! Happy benching

 

It's even easier than that. And, it doesn't need to run in the background.

Instead of left-clicking the "Create Clocks Shortcut" try right-clicking on it, then create a Task. If you change your mind later, delete it in Windows Task Manager.

 

Tried +200 Core / 400 Mem with +25mV and even +37mV but Fire Strike crashes with a grey screen right before the first benchmark ends. I guess I got a suck GPU (ASIC = 70%)

 

ohhhh so the offset needs to be way higher than what I tried...... gonna try higher and see and thanks for the tip about the task, I missed that

 

@Phoenix the jigsaw gskill that i asked where you got it from and you told me hidevolution i got those too.. are those at 28000 mhz? do i have to set the meory at the profile 1 setting?

 

That screen shot does not show the voltage. Also, be careful trying to copy those setttings. The offset calculation for 980M is not the same as 980 and it can change by vBIOS mod as well. You might end up crashing or freezing if you try the core overclock setting shown.

 

which ones did you get? 2400 MHz or 2666 MHz or 2800 MHz?

You need to go to Overclocking Menu / Memory / set the profile from Default XMP Profile to XMP Profile 1, change nothing else, then save and reboot

now launch HWiNFO and see what clock speed is the memory running then multiply that by 2 to get the actual value since it's DDR

 

so +37mV wasn't enough for +200/400MHz overclock, any suggestions what I should try?

 

hey guys, noob question here, im a bit confused about setting the undervolt in the bios. i see where it says core voltage offset, but not sure how exactly to enter it in. the description was something like 0x50 to 0x10.... basically, my brain hurt. can anyone here give me a brief rundown if i want to set it for -150mv?

 

this is an example for a -70mV offset. Just type using your keyboard (not the numpad, the top numbers on the top). enter 150 then make sure to change the offset prefix from + to - by hitting your ENTER key once on it, then press the down arrow to change it to -

If -150mV wasn't stable, try -140mV, if not stable, try -130mV

 

thanks phoenix. ya, i am going to have to go that route because XTU just isn't cutting it.

 

i got these ones
G.SKILL Ripjaws Series 64GB (4 x 16G) 260-Pin DDR4 SO-DIMM DDR4 2800 (PC4 22400) Laptop Memory F4-2800C18Q-64GRS

this is a cinebench result

you think its safe for me to go for 3000mhz on the ram?

 

1) That's not the screen I want from HWiNFO bro.

Please close it and run it again, then uncheck the RUN SENSORS only box to run the summary, then you should see something like this:




2) To properly benchmark your RAM, run AIDA64 Extreme, then right click on its icon in the taskbar and choose cache and memory benchmark, here is my score for you to compare:

 

perfect! your RAM is running @ 2800 MHz indeed and your score is slightly higher than mine as my score was taken while the CPU was @ 4.4 GHz but now I'm @ 4.5 GHz

Thank you for buying the same RAM like me.

PS: Please go to BIOS > CPU Configuration, then enable Intel Speed Shift Technology. Then that SST in red in HWiNFO would go away

 

i will do so.... quick question... the sillicon i bought was the 4.8 since i want to try to go to 4.7 at least. what should i change?

 

I have 4.8 Silicon Lottery CPU. Mind you, that was tested in a desktop so on a laptop our score will be around 200-300 MHz less which is still awesome for a laptop

for 4.6 GHz stable, I used a -30mV for Core / Ring.

Leave the Cache Max OC Ratio / Min to 0 (which means default which means 41X), but make the voltage offset for both -30mV

Set your uncore to -30mV:



Also increase all power and turbo limits like this:















The CPU will only draw the power it needs so no harm maxing those settings out to give it enough juice when it needs it. This way the only thing you have to worry about when overclocking is the voltage and multipliers

 

eeeeeek i think ill stay at 45. as soon as i started to encode a movie temp went to a max of 86 so if i OC it more those temps are going to go above the 90's and i suppose thats not a good thing.

 

true, I can hit 4.6 GHz stable as well but temps reach to 90s sometimes so I think 4.5 GHz is the sweet / stable spot on the long run for a 24/7 overclock.

Remember the days when we couldn't even get 4.2 GHz on all cores stable? this is like a new CPU man! love it

here are my memory benchmarks with my current 4.5 GHz overclock:

 

Yup similar. yours are a few points higher

 

so i applied the -150mv underclock in the bios but it doesn't seem like its sticking. i checked it in hwinfo and it shows 0mv offset under the cpu portion.