Clevo Overclocker's Lounge

So you mean by default once it finishes the 4 tests it will automatically stop running then?

 

What is the people here's favorite/best backup image program?

I'm about to start experimenting with tweaking the OSes and need a better backup program than what I currently have. Looking for something I can just redeploy if a tweak (or combo of tweaks) goes wrong.

Also, I know most here are not fans of XTU (myself included), but what can cause issues with perftune crashing when expiring settings/score?

Sent from my SM-G900P using Tapatalk

 

Macrium/Easeus/True image.

Intel XTU/Throttle stop over CCC XTU.

Also, I would not listen to anyone saying turn off system restore. Why? Because that will save you about 95 percent of the time from an image restore.

 

Please , can any give me the correct values for Clevo Control Center for my P775MDG3 with I7 6700K and GTX 1080 ?

I dont know what are the correct values , for optimal OC and stock . I have continously trottling

Thanks

 

First run at stock and post up your temps/clocks using XTU/MSI afterburner graphs over some gameplay. There is no point overclocking if something is up with the base.

 

So the culprit of throttling issue is the Clevo control center . I overclocked 4.5 with CCC installed and it's throttling around 1.2ghz 2.5ghz and when I uninstalled it. my clock speed runs now at 4.5ghz

 

Finally managed to get 3000mhz stable.
All I did was use XMP1 and then set to 1.25v



update: TM5 passed. Wonderful!

Goodnight guys!

 

More than likely it's the watts that did not get set when running CCC or they did not stick, thus causing it to throttle.

If that is G-Skill's then that is XMP2. CL16 is XMP1

 

Thanks. I've used Easeus before for its partition tools (which worked wonderfully to restore a deleted partition that was not changed/formatted after the accidental deletion).

For XTU, this is on my desktop (so I left the question general to apply to anyone). I'm using the Asus XTU tool, but cannot pin down the cause of the issue. I'm running a quad-boot scenario (Win 7 Pro SP1 unedited, Win 8.1 Pro edited wim before installation, 2xWin 10 Enterprise edited before install). Right now, I'm trying to track down the issues effecting each OS on the installed software (Win 7 is saying the driver isn't signed, I disabled the driver signature enforcement, it starts the program, but exits the benchmark process immediately after it is clicked. Win 8.1 Pro is taking WAY too long on the single core part of the test. Win 10 runs it fine, but then perftune error occurs when trying to export the profile for the score.).

I do keep system restore around, but I remove unneeded components from the OSes meant for pure benching that serve zero purpose (been trial and error, but I've gotten them stable and basically working with all benchmarks I use, but I have stayed away from XTU on my P770ZM, so I never paid attention to whether my tweaks or removing components affected XTU). I'm planning on removing a couple more items that cannot be removed until after installed and testing stability, which because I remove the partition for recovery image, I'm looking for a fast, quick redeploy option that mainly, if anything, involves troubleshooting a little on the boot side after deployed.

I realize with a custom image on all except Win 7 Pro, helping troubleshoot what might be causing the issue is much harder. More asking for general troubleshooting with XTU. I'll dig into those three programs and see what looks best. Thank you for the help!

In the coming weeks, with winter finally arriving here, I'll try to get back out in the garage to bench my ZM again (shooting for 52+ this winter, but we'll see where it goes)...

 

Your desktop board should have it's own over clocking software....

 

It does and I've played with DIPS on it. I'm solely using XTU for benchmarking and to submit scores to competitions (I know I'm not in the top contenders, but still like submitting to them anyways on esports) and hwbot. I prefer setting the OC in the UEFI for my everyday OC and play with the other if I'm trying to OC inside the OS and bench...

Sent from my SM-G900P using Tapatalk

 

*sigh* i wish i could set DIMM voltage in the bios, but it always resets to stock 1.20v

so looks like i got all timings figured out at 2800 mhz, only needed very slight adjustments coming from 2666 mhz (just command rate 2 instead of 1 and CL15 instead of 14). interesting tidbit: the tighter the ram timings and the higher the ram clocks, the higher my Vcore voltage needs to be for stability. example 4.5 ghz: at 2666 mhz stock timings, -80mV was stable. at 2666 mhz with tweaked timings i had to raise it to -75 mV, and now with 2800 mhz and tweaked timings its stable at -70 mV. curious, huh? might be one thing to look at when having troubles getting 3000 mhz stable, definitely best to start off with stock clocks and voltages on the cpu! that and healthy overvolts on cpu analog I/O (+300 mV) and system agent (+200 mV) in order to remove any potential bottlenecks.

Sent from my Huawei Mate 8 NXT-AL10

 

Exactly... The computer won't shut down, but it'll just have the results screen sitting there waiting. I ran mine with max fans just to help keep the RAM temps down (if that's even an issue), no point in over-stressing the RAM due to heat. Not sure if the fan profiles monitors RAM temps or not to ramp fans up

Does it matter or not if TM5 is run with administrator priviledges? I've seen that message in mine as well, and it definitely goes away when run with admin priviledges

 

Trottling after 3 minutes playing BF1...

 

Cool thanks, ill give it a shot tonight and hopefully I will wake up with no errors!!

 

I am assuming you are running your fans in auto mode? Mine does the same after playing for a while as the fans never ramp up much at all. Try increasing the fan speeds, overlock or maximum and see what your temps look like and see if you see any throtling.

 

That's how it's suppose to be done. . Or a base set in bios and the os can be used to go in many directions from there.

I don't think that actually works on our (P870DM2/3) machines. I set that to -1000 on both and nothing happen.
So far this is stable, but does not seem to be optimal yet.

 

http://hwbot.org/user/44212_



http://www.3dmark.com/sd/4429183



http://www.3dmark.com/fs/11023961

 

@Mr. Fox Congratz on the new achievement!

 

Great scores! So you've gotten your G.Skill 3000 figured out, is it 4x8GB or 2x16GB?

 

Well I had some issues so did not get to try out memtest86+. Formatted a 3.72gb usb drive to fat 32 using the full format first then installed memtest86+ USB installer using the format in the program launcher as well and the USB would not boot when I switched the boot options in the bios. Then I tried with the .iso. Burned it to the USB and it still would not boot from the USB. The USB shows up in the bios but will not boot to memtest86+. Are there any additional settings that must be changed inside the bios.

 

Strange... It should work setting it from the BIOS.

You could try setting it up to boot from a USB drive in the "Advanced Startup Options" of Windows (depending on your windows version).... http://www.tenforums.com/tutorials/2294-advanced-startup-options-boot-windows-10-a.html

EDIT: Also make sure your USB drive is bootable: http://www.instructables.com/id/How-to-Create-a-Bootable-USB-Drive-Without-Using-A/

 

When you say it will not boot, do you get an error of some type, or it just skips over it and boots the OS, or what exactly? You are pressing F7 and selecting the UEFI USB Partition 1 from the menu? Then what happens? You're not using Secure Boot cancer poot, right?

 

Tried that and it did not work just booted to windows. Went back in and installed my windows recovery USB and it worked fine. Tried with the USB with memtest86+ and I got "repoot and select a proper boot device". It is showing up in the bios as a "Lexar" which is the name of this USB but wont boot. Must be doing something wrong; however, I tried installing it again and it did not work. I also tried it on a different USB before this one last night so don't think it is the USB drive.

 

F7 gives me options for windows boot manager or to go into the bios. Once in the bios I have options for windows boot manager and the usb. When setting usb first and windows boot manager 2nd it boots to windows. When setting the USB as the only choice to boot I get "reboot and select a proper device"

 

Try a different USB stick. If you have one, use an old slow USB 2.0 flash drive. Don't reformat it. Just let the tool that comes with Memtest86 do all the work. Some of the newest USB sticks are useless for anything but Windows-based flash storage due to proprietary partitioning schemes.

The USB drive is not bootable, or you would see it in the F7 menu. That's the problem right there. Refer to my comments above (in this same post).

 

Does the size of the USB matter? As long as it is formatted for Fat32 it should work correct?

 

I don't have a USB 2.0 maybe I will pick one up this weekend. How good is the windows memory test at finding errors?

 

Thanks, I will give it a shot with a new USB once I pick one up.

 

I don't think the size matters, and I don't think it even has to be FAT32... Mine are NTFS file system, I have a 2GB SanDisk USB2.0 one and a 128GB PNY USB3.0. But I do agree with @Mr. Fox I have had some USB devices that failed to work as bootable in the past, but work just fine for regular storage

 

Thanks, that must be the issue. I'm not going to sacrifice my bootable USB with windows recovery on it which seems to work but will pick up a 2.0 and give it a shot.

 

Nice scores bro Fox

---------------------------------------

BTW... Skylake-X Kaby Lake-X (112W TDP) + LGA 2066 could be a reasonable compromise for a new P870DM4/5 - similar Clevo model next fall!! And Skipped of the 91/95W version quad-core i7. And IGPU is not in use anyway. Wasted space on silicon

http://www.guru3d.com/news-story/skylake-x-will-be-launched-at-gamescom-2017.html



http://www.digitimes.com/news/a20161208PD201.html?mod=3&q=INTEL

 

Does anyone know if the GTX 1080 in the P870DM3-G has higher power limit than the GTX 1080 in the P775DM3-G? or they are 1:1 the same cards (vBIOS & HW) ?

Power throttling is really annoying with my GTX 1080...I just need to load up Time Spy test, even in the first minute it will power throttle, which makes the core clock to drop for a millisecond. Core clock is mostly above 1800 Mhz, however because it is jumping around this also causes occasionally microstuttering in some games.

Anyway to "fix" this ?

My GTX 1080 vBIOS in the P775DM3-G is version 86.04.2A.00.2B, however GTX 1080 in P870DM3-G is version 86.04.2A.00.2C meaning it is slightly newer. Is there any difference ?

 

You can use any size USB stick. There is no need to format it because the tool provided by Memtest86 will reformat it. Be sure you are using the included USB formatting/configuration tool that comes with Memtest86. If you are not using that tool, or using it incorrectly, that may also be why it is not bootable.

 

Hey Brother @DRevan as far as I know they are exactly the same 1080 GPUs between the two machines. vBIOS is probably the same as long as you are not comparing G-Sync against non-G vBIOS. Those are different. My vBIOS version is 86.04.2A.00.2D (with G-sync).

If you are having CPU throttling that same underlying issue could be affecting the GPU behavior, in which case an EC that does not have messed up power limits could resolve both issues at once. Another other thing that could be affecting your system is thermals. 1080 begins dropping core clock speeds between 45°C and 50°C and beyond that continues dropping core clock speeds. Unfortunately, it doesn't have to be "hot" or even "too hot" for 1080 to throttle.

 

Hey @ Mr. Fox
Will try to set back the CPU to factory speed (4.0 Ghz). However if I remember correctly I had the same problem with 4.0 Ghz too. Could you please save your vBIOS and send it to me? I would try updating my 2B to your 2D version. Hopefully it has a better power table.

 

Here is my stock 1080-G (HWID 1BE0) vBIOS. https://drive.google.com/file/d/0Bwdqi25LDwZyYkt5YU9jTHZMS0U/view?usp=sharing

 

Here is my stock 1080 GTX (GSYNC) VBIOS:

http://www.mediafire.com/file/wvrvz6tlijul8rh/GP104.rom

 

Thanks guys! Will try out a new vBIOS.

As you can see, with the default one I get PWR throttle immediately when Time Spy loads.

 

Successfully updated to 2D:


Now let's see if it means any difference ...

 

hmm, will need some more testing, but PWR throttle seems less aggressive. Still throttles, but core clock did not go under 1800 Mhz like before.

update: nvm, its as bad as the factory one.

 

I am relatively certain the CPU and GPU misbehavior you are seeing is controlled by the EC to a great degree. If they have power limits set in the EC that are not higher than the machine is capable of consuming it will be broken for everyone. Power caps/limits need to be set higher than what is possible even with the most extreme ludicrous amount of overclocking. Better yet would be that they have no limits whatsoever. There is no reason or legitimate need for setting power limits.

Or, it could be something else is screwed up in the stock Clevo firmware and I am totally full of crap.

 

I think it would be better if someone teamviewered into your machine to make sure.

 

It is a bug (feature?) in the EC or vBIOS, similar to the PWR Throttling with the Founders Edition cards...however here when I use GPU-Z render test there is no throttling, but when I load Time Spy immediately throttling occurs ...

 

Then you get people overclocking and stress testing the machine to the point it goes bang and returning it for a refund and bad things happen to prices or shops.

 

Can't say I have ever seen it happen yet, but I suppose anything is possible. At any rate, using factory settings that cause every product sold to malfunction is probably worse than one machine among thousands going bang, simply because it curses everyone that buys one with problems, as opposed to the one-off example of a dead machine that was killed by user ignorance. Using a Procrustean process based on the lowest common denominator is an Alienware approach to business.

 

There's safe limits, and then there's absurdly low limits that aren't remotely close to what should be considered potentially dangerous.

 

True, I would make a physical check of the pad on the GPU VRMs to make sure they are making good contact.

 

The CPU, GPU and AC adapter will be the limiters. There is no reason to synthetically cap things off at the knees using cancer firmware. If the CPU gets too hot it will shut down. If one uses wrong settings it will be unstable and never reach full power because of user-induced throttling, Windows crashing and freezing, etc.. If the limits of the AC adapter are exceeded the AC adapter will shut off, or the machine will shut off from lack of sufficient power, or both. I still think there is no legitimate excuse to capping system power limits. Voltage, perhaps, but not power limits. You can only go as far as you can go. It should go as far as it can go without throttling to attempt to stay below an arbitrary power draw threshold that has not completely exhausted the capacity of the system.

 

Im pretty much inclined to agree with half of this. Mainly because I'm sure the components used aren't 100% to spec and certain batches will have lower tolerances. Now if something like the processor is stable at a high frequency , some other component might not have the same tolerance / sustainability.