The Official MSI GT73VR Owners and Discussions Lounge

I think that if you have suspicion, just don't buy it.
How many does it cost?
May be you could ask him if he used this laptop for mining

If it's a 6 month's minning old laptop, I shouldn't buy it (except for a few bucks)

 

My GPU Fan Speed is always @ 1200 RPM when just surfing the net and it doesn't do what you are describing.

I am on EC Firmware 17A1EMS1.108

Make sure you are using the EC Firmware for your laptop as I didn't have good luck when I flashed the EC Firmware .112 from a later model.

Also, you can try this as a last resort

1) Turn off your laptop

2) Press and hold the power button for 35 seconds straight and don't remove your hand off the power button, it will fully reset the BIOS/EC. Then configure your BIOS to your liking again and try

 

Offset 98 controls the fans.

 

IMHO if the laptop survived mining for 6 months it's pretty solid
May need a little compressed air through the fan-sink though!

 

Thank you for the fast opinions.
It should cost 2000€, but I think it's the 6820hk version.
Because on the sticker at the picture is 6RF, while the text says 7820hk.
But it's not a big difference right ?


Gesendet von iPhone mit Tapatalk

 

Find out if it's a 7820hk or 6820hk.
7820HK has +300 mhz boost speed increase over 6820HK.
7820HK may also require less voltage than 6820hk.

Some rare 7820hk's can do 4.5 ghz at 1.13v. (4.4 ghz AVX/OCCT rock stable; 4.5 ghz without AVX).
No 6820HK in existence can do 4.2 ghz @ 1.13v stable, at least I've never seen one.

 

Isn't it expensive for a 6820hk (6rf)? Mine is 6 month old, "only" 1070 (6re) but I paid 2000 euros with all taxes in Europe.

My gt73 run @ 4.2ghz daily but is quiet hot, 85c on heavy games, or 3dmark. 7820hk is of course better. But 6820hk is not bat at all ;-)

 

It's fine for GTX 1080, Iow9, but for GTX 1070, you have to be careful. Some of the bad revisions of the 1070 were on the 6RE, both Titan and GT72VR.
We already identified the problem was with one or more of the voltage steps or curves on the Boost 2 clocks, causing crashes, and it's impossible to know if you get one with a fixed revision or not, as this was a hardware fault.
Someone posted a few months ago that there were at least *SIX* revisions of the MSI GTX 1070 MXM card.

But I'm complaining because someone's selling a unit marking it with a 7820HK but it has 6RF on the sticker. I wouldn't trust a seller like that unless there was some sort of buyer protection or money back guarantee, and how would you enforce that from a private seller?

Unless you were able to do a battery of stress tests at the purchase site...

 

Yeah, I know this but it's seem's okay in my case. We always heard the one who have the problem but not the 99% that doesn't

Total agree with you on this point !

I made a few test with the skypro and hum... No screen at all now. Not even with full stock vbios. I tried the clear cmos etc

 

Thank you all for the fast help.
It's really the best forum.
I won't buy this laptop and the seller is not answering anymore since I'm pointed on the processor
I will search for another opportunity to get this beast.


Gesendet von iPhone mit Tapatalk

 

Guys,
please help me,
My GPU and CPU fan is changing in each 1 or 2 second in DC.
is it ok? or i have problem?
how your fan always is showing 1200 rpm?

for me:e.g GPU:
1200 - 1173 - 1206 - 1800 ...

please help

 

You don't have a screen at all?
Does the system boot to windows?
You can switch to the iGPU.

And clearing CMOS gives you a black screen for at least 3 minutes.
You will have a black screen for 60 seconds, then the power light will cycle (power reset), then after another 60 seconds, it will cycle again. Then the MSI logo should appear followed by more boot loops. If it does not boot loop at all (every boot loop, the power LED will cycle), you have to do the 45 second power button press again. This type of "Soft brick" usually only happens when messing with RAM timings. Someone said GT73VR has two Bios chips, and the boot loops are the backup bios copying failsafe settings onto the main Bios.

To switch to iGPU (if in windows):
GPU button (blind) -> alt tab ->tab->spacebar
If no other programs are running, this will switch to the iGPU and reboot.

Note: there is a bug where the dGPU will always black screen due to a Bios bug, when switching from iGPU to dGPU. Several combinations of "warm power off resets and saved Bios User Profiles will trigger this bug.
Changing PEG Port 1 from Auto to Enabled will prevent this (requires unlocked Bios).

@powerofviva most of us don't use DC.
You can try the MSI "SIlent Option" program, or just use auto fans or cooler boost button.
You can also control fans manually by writing to EC register 98.

 

iGPU works
I tried the second MXM port but seems like it's the "second one" who doesn't work without the first one .

With the Nvidia Gpu, I think by the activity that it boot normal but no blacklight and no display. The gpu is working a little bit, listening to the fans
No output on hdmi

seems like it's working but I don't see anything

 

You probably ran into the PEG port bug. If you have unlocked Bios, change PEG Port 1 to Enabled. This will restore the dGPU and prevents the bug (it's a bug with the "Auto" setting).

This will prevent the "black screen" from happening.
Without unlocked Bios, this bug is almost impossible to overcome (it has caused some RMA's already).

Without unlocked Bios,
You have to switch to iGPU (first), then enter Bios, "load user defaults", save, exit, switch to dGPU, reboot. This "reverses" the bug. If you can't see the Bios because of the bug, you have to switch to iGPU "blind" in windows.

GPU button ->Alt Tab->Tab->Space bar.

If you flashed the stock Vbios and got a black screen, you almost 100% certainly ran into this bug.

Clear CMOS works but takes a very long time.

 

No success :-(

 

Thank you,
But is it ok that my DG is changing the value of CPU/GPU rpm every 1-2 second ?

 

Make sure pins are clean.
Try flashing stock 8A Vbios also. Check thermal repaste quality.
Clean pins clearly with Deoxit D5, leave a very faint coat remaining, then reinsert in socket and make sure MXM is inserted all the way evenly. Network card requires almost the same method but it goes in easier.

It sounds like a contact issue with the slot. A damaged vbios would not even POST @Coolane

 

Yes, it look like an hardware problem.

I Hope I didn't broke something. I will flash stock bios tomorrow. EC too. Clean everything. Weird

Edit : I will try teamviewer too.

 

I will try to reset the EC firmware, though I'm using 108 and I think it's DGC bug, Fan setting Auto speed not decrease to 1200RPM if temp between 41 to 44 but Fan setting Advance and it decrease the Fan speed to 1200RPM if temp between 41 to 44 (default settings)

Also your laptop GPU fan, on Cooler Boost, does it hit 5000+ or RPM fluctuate between 4800 to 4980 Max ?

 

on Cooler Boost it fluctuates between 4948 RPM and 5000 RPM

 

Next episode...

I try to flash the original vbios

 

That's going to happen if you're on the iGPU. You need to use the skypro.

 

No, I was on the Nvidia GPU with teamviewer. I clean the mxm port both side...

Seriously, I'm short of idea

 

Did you reflash with the skypro?
Usually that error with the 0mhz clocks is caused by a vbios error (from what I saw on the desktop gtx forums).
May I ask what part of the country you are in?

 

Hello guys, first post and new to joining the MSI family. I just got my GT73vr 7re-427 and love it. I wanted to know is there anything I should look out for in the early days besides the obvious, updating and installing drivers. What I'm most concerned about is what are the better bios to have running or is the latest bio in MSI website the latest and greatest??

Im not sure if the nvme drivers are up to date?

 

Yes, I flashed it again with success.
Do you know where I can download the 8A bios? I found it on a Chinese website but need an account.

When I read the bios from the chip, it works too.. May be I damaged a pin on the vbios but I don't see anything.

I'm in Belgium.
Too bad that my old MSI gt60 doesn't work anymore to try the card..

 

Hello, and welcome to the Owner's Lounge.

Please put your model and configuration in your signature so we can see what you have...it will help us, help you. No one wants to look up the specs for the GT73vr 7re-427, and if you got any upgrades, please indicate them.

Also please indicate from which vendor you purchased it.

 

I have done what you ask. Today is my first day getting the laptop but I see you offer a discount . I wish I had known that before buying it, im a college kid and every penny counts. Depending on the price I could RMA it, LOL. I got it from Amazon so their return policy is usually oriented towards customer service.

 

8A Vbios.
Keep in mind that your maximum benchmark speeds will be slower with this, however your frame times will be more consistent (less stuttering) because the GPU clock will not change due to automatic overclocking from boost 2 being removed.

You DO need to use the "preset" button to allow higher than 150W TDP to work. The preset button should allow 200W TDP maximum, then you can set the slider from 100% to 130 or so%.
The preset button changes the max TDP limits to max the GTX 1080. I do not know how much power draw the MXM slot on the GT73VR can handle without the auxiliary power connector that the 1080 has.

The Clevo slot can handle up to 258W (this was tested by Khenglish). No idea about the GT73VR. Going to 170W TDP is safe. Any higher is at your own risk.
If you choose to use 170W or higher TDP, please unlock the 330W AC power limit by using RW Everything and changing offset E3 to either 11 or 91 (both work on the kaby lake GT73, only 11 will work on the skylake, as 90 (230W) and 91 (330W) were added to EC 17A1EMS1.108 for kaby lake support. Bypassing 230W if you don't unlock the 330W AC limit will cause CPU PL2 throttling. The 230W PSU is capable of 250W but I would not go above that (if you need more than 250W sustained power draw, please buy the 330W AC adapter).

 

what arethe benefits of having 1070 at desktop TDP level? How much performance did you gain?

 

About 15% improvement. You can exceed desktop 1070 performance by using 3A Bios and going crazy on the TDP, which is really not recommended going above 170W unless you have money to replace hardware. Look at Coolane's scores in the Pascal tweaker thread in the Sager Clevo section when he was pushing 258W through the slot.

 

If you didn't use a 1.8v adapter/programmer you may have burned the chip. You may have been unfortunate enough to actually destroy something with static. May have knocked off a little smd resistor somewhere, who knows.
Good luck with the RMA though!

If you don't have any problems with your 1070 on bios version 3A I recommend sticking with the official MSI 3A firmware.
8C is a debug bios made to disable boost clocks on dodgy cards.
The 1070 won't see any gains modded past >200w. For a single card using version 3A modded to 200w you can expect performance to be better than overclocked stock.
Whilst they've had a myriad of problems the MXM1070 in MSI's GT series has an overkill VRM so it will handle more power.

 

mine GPU fan fluctuates between 4800 to 4948 and barely touch 5000 where early days it crosses 5000 and now it sometime goes below 4800 also whistle sound is coming from GPU fan while on full speed, I clean the fan but same result, my guess it is failing....too much gaming.....

1 month left in first year warranty, time to call Tom Ho...get a new fan before it totally fails....

 

Ok found out how to reactivate full AC power (no CPU throttling) when the battery is disconnected.

Offset 31: change 00 to 09. This will reactivate the battery charging circuit (LED will turn on), even though the battery is not present.

Offset 42: change 00 to 5C. This will remove CPU throttling and artificial power limit from the battery being disconnected

Changing offset 42 by itself is not sufficient.

If offset F4 is changed to 8C, change it back to 0C to avoid system critical limit power cutoff.
Change offset E4 to 91 (11 if you have a skylake system) from 90 (or 10), to change 230W AC power limit to 330W.

 

Well guys, I have bad news. When I got the laptop today i turned it on and installed anti virus and steam and chrome. Then it shut down automatically after windows finished updating. I went to class and came back around 10 pm. Computer off pretty much since morning. I turned on and there was a green pixel on the screen, I didn't get mad hoping it would get unstuck but it didn't. I'm RMA it becuase I refuse to try to work with a laptop that I have used for less than 4 hours and it getting a dead/ stuck green pixel for that kind of cash.

My question, do you think it is possible the windows update had anything to do with it?

I didn't install any other drivers just whats mentioned above.

 

My stupid GSM replace all "vbios" by "bios" but I have "advanced" knowledge, I modded bios and vbios in past years (but not good @ English).

Now, I don't want to overclock, just a working laptop should be good ;-)

Thank you! I used the 1.8v and the chip is working as I read it without any problem... I thought about static. Don't know if they have protection against that

 

EDIT: Working on Skylake EC108.

 

@hmscott @Phoenix @Papusan @Mobius 1

you guys have any idea what actually happened to CPU HS, I applied LM Conductonaut, few days ago, may be 5 days ago and not noticing any mind blowing temp drop so today I thought to look under the hood and pulled off the HS and that's how the HS is after cleaning LM, nothing happened to CPU die, it's in perfect shape.


after cleaning, i re-pasted using Kryo but not sure the temps look good any more....

 

I'm on my small phone and can't see the picture very well... But The worse the fit is between HS and Die, the more will Liquid metal react with copper(You by yourself said you don't get good results with LM aka seems like bad fit). Use a powerful cleaner similar Indigo Extreme cleaner + sand carefully with very fine sandpaper. Minimum 2000 grits.

 

Kryonaut is the worst Thermal paste I ever used. Period. Garbage. Absolute rubbish.

 

Kryo is the only paste which gives me good temps, LM failed in my case so as ICD, look at pic what happened to HS after LM....

 

I cannot get my hands on on Indigo, not available in my country, yeah I can lap the HS and clean it, I think I have 2000 or plus grits sand paper lying around somewhere.

 

For me Kryo only gave me good temps for 2 days then it started to get worse and worse

 

Dont sand/shave of all the copper You can use the strongest quality isopropyl. You find it up to 99%
For thermal paste. Try Phobya NanoGrease Extreme

 

Kryonaut requires perfect mounting to avoid dryout. My GPU has Kryonaut and temps have been the same for 2 months since I applied it after doing vbios mod, and stable..
For CPU's it's all over the place. My first 3 Kryonaut CPU attempts had degradation, but the very last one went much better. Maybe it rose by 4C at WORST then remained perfectly stable for a month (I was satisfied with this) until I switched to Conductonaut.

Conductonaut dropped my CPU temps by 10-15C(!), made 4.5 ghz Pentium 4/type 0 FFT 8 thread prime testing possible to stay under 80C in an A/C'd room (Great), and
made 4.4 ghz AVX/FMA3 prime 8 thread possible without reaching prochot/99C. This was impossible before.

Try Nanogrease Extreme. @Mr. Fox had great success with it. He said it was better than Kryonaut for him. Unfortunately, in MY case, it was worse than Kryonaut.
With Kryonaut, I had a cpu temp delta between coldest/hotting cores of about 4C. With nanogrease, it was about 7C. Multiple attempts. Also I had worse temp degradation with nanogrease than Kryonaut. That's why I went back to Kryonaut before going to LM, and tried redoing the screw method for partial turns, which seemed to work better.

For now, I have Conductonaut on CPU, Kryonaut on GPU, both great and stable.

 

I checked Nanogrease Extreme spec sheet and the viscosity is lower than Kryonaut.

 

cannot get my hand any of your suggest stuff, where I live only option is to import so lapping was the only option, though I'm thinking to buy CPU HS from MSI, since day one I'm struggling with CPU HS...

after lapping, it also shows that it's surface still uneven, though i left that LM on HS, it's now part of the HS copper.....temps back to normal like it was with Kryo before applying LM...i will stay away from LM till I get new CPU HS....

 

Yeah, people think MSI don't deliver uneven/warped HS for their laptops. The more expensive the model is, the better quality and QC. People/biased owners take always wrong They should stop fool themselves!! @Phoenix
And exactly same temp as before? You could take of HS again, clean of and lapping a bit more, then remove the 4 washers before you put on new paste and screw down. You could probably shave of a few degrees with this step even with Kryonaut or other top thermal paste. Worth a try, bruh.

Edit. You should take a extra look on how the fit is. See heatsink as the pict. Lapping slightly or see if you can use thinner pads if needed there as well. The whole HS can be uneven. Test out.

 

I saw this posted on baidu, as well as some discussions about this in the clevo OCer thread.
Will this work with 16L1 as well? i have a friend who currently owns a 16L1, and his machine cant run pass 230W. He has a programmer, but i have heard that the EC chip is not easy to program.
He will probably try using RW first, i will post here if he succeeds

 

temps nearly the same, may be bit better, last kryo re-paste (before LM), 3 to 5 degree difference between cores and this time after lap, re-pasted with kryo and 2 to 4 degree difference between cores where core-0 and core-1 still high as compare to core-2 and 3 also core-2 and 3 temps always same....I will give one more lap to CPU HS

there are no pads here....