MSI GT75 TITAN 8RG-094 Review by Ultra Male

I was able to get 1606 in Cinebench a couple of times just by manipulating the offset voltage. Of course it was throttling almost all the time, but for completely stock system it's not a bad result IMHO.
Not particularly stable though. Interesting thing was that most of the time crashes happened soon as the bench finished. I'm thinking it was caused by the infamous "voltage droopage" @Falkentyne was talking about?

 

same here with my GT75VR 7RF but after changing the ports where the ssds are connected i never had an raid fail after CPU OC. maybe you want give this a try the next time.

after doing an factory reset of my router i have now since 17minutes 1.733Gbps stable with my killer WLAN 1550. cross my fingers..

 

that is a very common symptom of overclocking. That's why in OC Forums (Overclocks.Net Forums) they always recommend to do all your overclocking and testing on AHCI and only when you find a stable overclock should you format and install Windows in RAID and copy all your data which I have no time to do.

 

From Silicon lottery bro...


And who would let @Danishblunt touch his computer @Phoenix ? See also his reply on your post below.

What? About time you start to read you up on the topic.... Overclocking!! What's needed, How to reach the goal and how to perform it. And what can go wrong with OS, raid array etc....

 

Messed up overclock settings can cause severe file corruption. That's nothing new... very common. Better to start off with higher voltage than what you want and work your way down to find the lowest stable voltage. Going the other direction takes longer and seems to cause more corruption.

Resetting the CMOS on an NVMe RAID array can also be a lot more risky that SATA RAID. Not sure why that is, but the NVMe RAID arrays have always been more fragile for me.

Yes, that is good advice, too.

Why you no use Macrium for this, bro? No need to fear the digital reaper with Macrium Reflect. I tear my OSes to shreds fairly often and it fixes things in a jiffy for me. Even stable overclocking can tear things up when a rogue driver decides to pitch a fit about it. Buggy cancer drivers are frequently the enemy of an otherwise stable overclock.

 

I appreciate the offer but firstly, I despise Discord or any other form of chatting platform and secondly, I am more than capable of overclocking and my CPU with its stock multipliers is already at its limits. One mistake and that could turn into a major disaster for me with my 10TB Data on RAID.

Finally, this is my main and only laptop I cannot be doing testing on it unless it was from Mr. Fox or PooPooSan or FalconEinstein who know exactly what they're doing.

 

Never happened to me ever. Kinda odd hearing it but I'll take your word for it since it does make sense considering how SSD's work and how sensitive they can be.

 

Because Macrium Reflect has an issue with seeing my drives in this NVMe setup.

so here's how I have thing setup

RAID Array = 6TB, partitioned into C: 200 GB (OS) and E: 5.5TB (Videos)
Samsung 850 EVO 4TB partitioned into D: 4TB (Data)

I stored my image on D:

When I hit restore, it would reboot and go into the Macrium Reflect environment and it won't see any of the drives. This started happening since Redstone 4 and they didn't release an update since ages.

so when I did try to overclock and it corrupted my OS, I plugged in my Macrium Reflect Rescue Disk and again, it doesn't detect a disk to restore to as if my drives were not there.

Mind you, when creating the Macrium Reflect PE Environment and the Rescue Disk, it does inject the RAID Driver but it's a no go

 

Man, that really sucks. Sorry to hear that is not working right.

 

the thing is bro, such an issue can be catastrophic for me even if I have all my data backed up onto external drives, it'd take more than a day to restore everything back. I can't deal with this BS

 

Oh god that guy is dumber than I thought.
He doesn't even know what VID Is.
@Phoenix , dont listen to Danishblunt unless you want a burnt laptop. Seriously.

VID: VID is a pre-programmed voltage identification that the CPU requests from the voltage regulators on the motherboard (note: some chips had built in voltage regulators), where the CPU will request a certain amount of voltage at each mhz step, precalibrated based on CPU binning silicon quality, starting at 800 mhz, and going all the way up to the maximum turbo boost speed (1 core). This VID is then recalibrated by the current resistances specified by IA AC DC Loadline values (IA AC specifies VID boost at greater current draw, IA DC specifies VID droop at greater current draw). Note that IA DC does not directly affect the actual voltage the CPU is getting, only IA AC does, but IA DC will show the amount of vdroop based on Intel's design characteristics. This is all based on using adaptive voltages with or without offsets, and is designed WITHOUT motherboard specific "Loadline Calibration" in mind (LLC), meaning LLC must be disabled for all specifications to be followed up to the maximum 1 core turbo boost speed. For Skylake era tech CPU's (skylake, Kabylake and 8700K all share the same basic hardware design), this all applies the same way. This still begs the bizarre question if why doe IA AC Loadline raise the voltage by resistance of 1.80 mOhms, if IA DC loadline shows vdroop of resistance of 1.80 mOhms? I have no answer for this. But keep in mind that vdroop is not coming from the CPU itself (on cpu's without the voltage regulators on the CPU itself) but rather what the CPU -thinks- should be happening from the mainboard itself.

Using static (override) voltages will override the VID given to the CPU. However the CPU VID and the CPU Vcore are not the same thing and cannot be used interchangably. In some cases, it's possible for (on boards with true vcore sensors) vcore to read 1.30v, while the VID reads 1.45v, but this is getting way off topic.

The IA AC DC loadline values, especially on "Auto", are not supposed to apply when static voltages are used. The VID may still show completely inaccurate high values compared to vcore, but the desktop mainboards are supposed to supply the target voltage requested, then compensated by vdroop via the "Loadline Calibration level" setting. IA AC DC should not be skewing these results. At least this is how it works on desktop boards (8700K), and SHOULD on 7700K desktop boards. But it seems that the laptop's vcore signal is being directly derived from VID, which is why IA AC DC is affecting laptops with static voltage overrides.

off my soapbox now.

 

@Phoenix maybe you wanna give Paragon a chance?
I use it and have good experience so far with this backup software.
the only thing what i don't get working is to restore it directly at the moment i must first install the os with raid 0 and driver and then the software acknowledge the raid.
i don't have so much data like you on my raid but maybe it fit to your highly request of solution.

 

Yeah, I get it. No need to explain. That's a huge pain in the butt.

You know me well and I was a RAID0 addict for years. Now I no longer use it because all of the overclocking I do it just became too much extra work having to deal with frequently trashed OSes. Dual booting has also saved my bacon many times. When one OS gets shredded, the other is usually still good. So, I keep everything on discrete drives now and suffer with a slightly slower data transfer rate in favor or high CPU and RAM clock speeds.

Speaking of RAM, buggy memory overclocking tears things up even faster than a buggy CPU overclock. It can cause corruption, cross-linking, orphaned files, etc. Also buggers up file and folder permissions sometimes, where everything is read-only due to corruption. But, when you're pushing for the edge of usability, you've got to take the good with the bad. Windows 10 does not handle this stuff as gracefully as Windows 7 does. I also went back to using Legacy BIOS and MBR drives instead of GPT for the same reason. I only use GPT on volumes that are too large for MBR.

 

Guys, c'mon. It's OK to call BS on bogus information. No need to do the eye for an eye thing. Just let it go. Extend an olive branch. Life is too short to hold a grudge.

 

Yes this happened to me too not long ago after overclocking/crashing etc etc.... Like you said, Windows didn't recognize them. Windows wouldn't even recognize it's own back up!?? Lol!

 

This is a bit over my head. All I know is that changing IA AC/DC Loadline to 0.01 mOhm causes a boot loop using override voltage, and an instant BSOD after Windows logo (WHEA uncorrectable error) on adaptive voltage.

 

it should be static voltage not adaptive when changing the IA/DC Loadline

 

Especially since overclocking such a high powered laptop is pretty much good for benchmarking only to show who has a bigger dick. I'm with you on that one.

 

Override is static voltage.

 

What CPU speed did you test this at?

Instant BSOD is obvious. Adaptive voltage is based on default VID + offset, and the default VID stops scaling at the highest turbo multiplier. And that highest turbo multiplier (highest) is, afaik, a 1 core default speed for whatever SKU.

If you want to test IA AC DC loadline, please set your CPU to the highest *stock* 6 core or 4 core multiplier speed for that SKU. Example: for 7700K, I think that's x39 or x42 and then start there at something like 1.2v.

You also have to remember that MSI has built in loadline calibration (which cannot be disabled) which 'removes' vdroop. It's been PROVEN that eVGA does NOT have this loadline calibration, so the static voltages have to be set over 100mv higher on those, because of vdroop.

I do not know what your Clevo has, however. If your clevo does NOT have any sort of built in loadline calibration, then if you set IA AC DC loadline to 0.01, you are going to probably be getting 150mv of voltage droop at load.

 

sorry, brain fart, I was thinking Adaptive

This is me right now if you ever wondered how I look like...

 

Stock 1-core turbo is 4.6 and all-core turbo is 4.1 on the 8600K.

I tried it at 4.7 with adaptive at 0mV and -40mV offsets, BSOD both times. So maybe I should try a positive offset?

Using my normal 4.7 at override 1.175V, changing IA AC/DC Loadline results in boot loop, so I guess it only works with adaptive.

How do I check if I have built-in LLC?

 

Your clevo probably doesn't have built in loadline calibration. MSI has a nice whopping of it.

You said 1.175v?

if you set IA AC DC Loadline to "1" (0.01 mOhms), you probably are putting 1.075v into that chip at 4.7 ghz.
That's not going to work unless you are papusan.

Try 1.325v with IA AC DC loadline = 1 and static voltage.

100% chance it will work

Try it

Should take you just a few minutes to test. Post back please.

 

Yes, Clevo it has it in the Prema BIOS. But the number implementation can be different based on vendor implementation. I can and do use 1 with 8700K in my P870DM-G. Doing that helps greatly with eliminating vdroop and overshoot (erratic voltage swings). Using 1 is the same as 0.01 on ASUS (1=1/100th) but entering 0.01 is not not valid on Clevo. If I entered 1 with 7700K installed in the P870DM-G it would not POST and I would have to clear CMOS to recover. No idea why it would not work with 7700K. Weirdly enough, entering 100 did work with 7700K. It would boot fine with 100, but that seemed to be exactly the same as 0 (Auto, default).

 

by a clevo barebone?

 

@Mr. Fox

Macrium's reply to my concern that I told j00 about today:

 

Hello i have a Problem with my GT75 Titan. When i will boot it up the Screen is Black and it goes into a boot Loop. I think i made a mistake when i change from UEFI to legacy in BIOS. Since then it won't boot anymore. What can i do. Any help is welcome.

 

Hold power button down for 60 seconds.

Then release and wait 2 full minutess
Bios will be completely reset.

This includes any IA AC DC Loadline settings, unless HIDevolution made the "1" value default in your bios, in which case that will still default to 1, ONLY if HIDevolution did it for you (the MSI default which overvolts at load is 179 or 1.79 mOhms).

 

No luck still in boot Loop and no MSI Logo on the Screen

 

that can't be man, if you keep holding the power button after you turn off your system, just wait for a long while and you will eventually see the MSI logo and the BIOS would be reset so you can set it up to what it was like RAID or AHCI

Where did you buy it from?

 

I buy it in Germany. It is the new one with AHCI or Intel Optane. Set to ahci and boot to legacy not uefi.

 

Can you remove the CMOS battery after unplugging the main battery and power for 30 seconds then plug it back in?

 

Where is the CMOS Battery on this Notebook GT75 TITAN 4K-071?

 

You WILL have FIVE OR SIX boot loops!

Be patient.
This WILL Fix the problem.

Try this right now.

Turn off the laptop.
Press the power button and do NOT RELEASE IT
It will turn on then off.
Leave the power button pressed for SIXTY SECONDS.
Don't skip seconds.

Then power it back on.
wait 60 seconds, the power LED will cycle off and on.
Wait another 60 seconds.
The MSI logo will appear and it will boot loop 5 or 6 times.

Then it will go to windows.
(Are you using a RAID setup or a single SSD?).

 

OK I tried a few different settings at 4.7 and IA AC/DC Loadline 1.

Override 1.3V = Locked 1.3V at all times, Prime95 Small FFTs (no AVX) throttle to 4.5 due to 'EDP Other' on Core and Ring

Adaptive + 80mV offset = Same as above

Adaptive + Extra Turbo Voltage 1.3V = 1.3V idle and CPU-Z Stress (no throttle), Prime throttle and Vdroop to 1.26V and BSOD

Adaptive + Extra Turbo Voltage 1.3V + 40mV offset = 1.34V idle and CPU-Z Stress (no throttle), Prime throttle and Vdroop to 1.26V and BSOD

Compare that to my old settings:

AC/DC Loadline 0 + override 1.175V = Unstable idle/low load VID, CPU-Z Stress 1.23V (no throttle), Prime 1.22V (no throttle)

Looks like I'll be sticking to my old settings.

 

If you are from Germany and need some help to translate something let me know it or you can send me an pm with open topics. I will answer it after I wake up later this day. Good night gents and lads.

 

To access the CMOS battery you will need to pull the MB so before you attempt to do that press the EC reset button on the bottom of the laptop using a paper clip.

 

EDP Other is because you reached current limit throttling point.
Did you change CPU VR Settings->Core I/A Domain->VR Current Limit to 200 amps (depending on the divider shown in the help to the right,this may be a value of 800, 1000, or 6000, check the divider)
If you are NOT using a prema bios, you may have issues here.
If you are, you will have no probliems.

 

Tried everything - it is the BIOS and what i understand do the EC reset Button not clear the Bios to Factory.

 

Yes upped current limit in Prema, no difference.

The other thing I've noticed is that using IA AC/DC Loadline 1 makes TS Bench go crazy with errors. Not sure how accurate TS Bench is because even my months stable 4.7 1.175V is getting a few errors now. The algorithm seemed to have changed in TS 8.60 because it uses more power than 8.50, and I had no errors in 8.50. Or maybe my silicon degraded already and I need to up the voltage.

 

T

Thank you for your feedback and testing. Appreciate it. Since you have a laptop which works as it was designed to work as, just leave that setting at default then.
MSI users need to set IA AC DC to 1 to stop the horrid overvolting (mainly because MSI also combines this with internal, impossible to disable Loadline Calibration (LLC). The amount of LLC seems to be equal to a "medium" level used on a desktop board, but without a digital multimeter or knowing where the vcore read points are, impossible to know for sure. No schematics are available....

 

I do not understand you.
can you GET into the Bios after the automatic boot loops finish?
Or is the laptop crashing and rebooting before you load windows?

Tell me--CAN you enter the Bios?
Make this clear, please.

if you CAN enter the Bios, then set it to UEFI. That's how these laptops are supposed to be used at, NOT LEGACY. Period.
If you can NOT enter the Bios, RMA it.

 

And thank you for the settings advice, learned a lot from it. It's interesting how the same BIOS settings have different results on Clevo vs. MSI. On my Clevo, IA AC/DC Loadline 1 definitely hurts more than it helps. When I change them back to 0, I'm Prime stable at lower VID, with no more 'EDP Other" throttling when running non-AVX Small FFTs:



I also figured out my TS Bench errors. The newest version of TS Bench is more power hungry and runs at lower VID, plus my NVRAM was corrupted after tweaking BIOS settings too many times. After a CMOS reset and bumping up from 1.175V to 1.2V, no more errors.

PS: Sorry for the OT, MSI owners. I thought my BIOS settings exploration might be applicable to you guys/gals as well, but turns out our laptops are quite different.

 

Nah MSI owners need to see this.
Clevo makes a non cancer laptop with Prema Bios, which performs exactly like a "Mini ITX" desktop would perform.

MSI makes cancer. MSI owners NEED to see this.

 

If only the P870TM1 had a mechanical keyboard and 120hz 3ms screen, it would be an automatic choice for a lot more people. Hoping Clevo somehow sees this post.


Sent from my iPad using Tapatalk

 

Hello sorry for the Confusion.

1. Notebook is powered off
2. Press Power Button
3. Fan are spinning
4. Power Button glow red
5. no MSI Logo , Screen is black
6. Power Button goes black
7. then it stuck in the loop with Power Button glow red -> goes black and so on

 

Did you or did you not hold down the power button DOWN PRESSED DOWN for 60 seconds?

If you did, RMA the laptop. Either it's broken or the RAM is broken.

 

What are exatly ram modules in stock notebook with 16gb ram? I want to plug in another 16gb but i dont know what i need to buy.

 

You should install HWinfo64 and it will show you the exact part number of the RAM installed in your laptop. https://www.hwinfo.com/files/hwi_584.exe

 

or cpu-z both are good for needed information for memory etc.