*** The Official MSI GT63 TITAN Owners and Discussions Lounge ***

anyone tryed swap the original DDR4 with 2 new ones 2x 16Gb DDR4 3200Mhz like HyperX Impact? or this processor/chipset only get the 2666Mhz like it says on manual? (I'm talking without manual OC and tweaking BIOS)?

 

Processor's fastest memory speed is 2666MHz.
So you can insert any kind of faster memory but this just gives you the headroom to overclock. The rest is up to your processor's overclocking capability and requires BIOS tweaking.

 

No one has tested faster stock memory in these laptops.
Official support is 2666 mhz maximum on 8th gen. On the GT73VR, it's 2400 mhz yet I was able to run my Kingston valueram at 2666 mhz with the unlocked Bios, no problem at all, with terrible RAM timings.(18/19/43). However the Bios has multipliers internally up to 4133 mhz. But it's your guess whether the CPU IMC will work with that, and if the Bios will accept it. One person managed to get 2800 mhz RAM working on the GT73VR, but attempts with much older Bios versions a few years back failed miserably. @hackness (I think) got an official 2666 mhz GS65 to run at 3200 mhz with 2400(!) mhz Samsung RAM.

So yes it CAN work. Will it work? I suggest you unlock your Bios (if the DDR RAM multiplier isn't available at stock), take your laptop to a local shop that sells SODIMM memory (good luck with that) and test it right then and there. Otherwise make sure you have a good return/RMA service available if you intend to test Gskill or Corsair 3200 mhz sticks.

You will most likely have to unlock the Bios, change the DDR multiplier, and increase uncore offset by 200mv in the unlocked Bios (usually under Overclocking Performance Menu).

 

My settings since 10/1, no BSOD or Freezing yet.



@ldcosta
If XMP profile can't be selected in the memory profile you may have to use Custom Profile to reach.

 

Very Interesting... 3467Mhz. I have a GT63 8RG I want to get another so-dimm so I can OC the stock one to around 3200Mhz. Did you get high temps or no diference at all?!

@hackness you have 2x16gb ddr4 samsung 2666Mhz original so-dimms right? Or are you using diferent ones? (I'm having hard time trying to get an equal one)

 

I was using GS65 Stock RAM Samsung 2400MHz 16GB x 2,

Part number is M471A2K43CB1-CRC,





Edit: Regarding to the temperature, I'm not sure because these RAM don't have sensors, but my CPU and GPU temperature don't seem to be affected.

There are a few online sellers that sell them if you search using the part number.

 

I love when manufacturers are out here participating in the forums and can help and provide details like this!

 

I agree...since i had the original post about this issue i wanted to say thank you for the Killer reps (that sounded a little..."off" lol) for reaching out and clarifying this issue. Also big thanks to Donald @ HIDevolution on his quick reply and action as a reseller too....top notch!

 

I'm new here, but how does one go about getting a customer bios by Svet? I would really like to know what it feels like to win as well.

 

Update -180mV CPU and DDR4 hynix and not Samsung got 3200Mhz stable for now, probably I'll buy a HyperX Impact 16gb 3200Mhz and try too keep it stable dunno if I can get -190mV on 8750h.

Changed on intel XTU the Reference clock to 102,5Mhz still stable and a little OC on CPU/RAM ​

 

Well done! Maybe you can share the RAM timing screenshot for Hynix

 

Didn't save the print after OC but these are the timings for 3200Mhz, played GTA V and Forza Horizon 4 (around 15min each without any crash and surfing 1h on the web, videos etc):

Tighten to boost read speed:

Set DRAM Reference Clock (MHz) to 133
Set Memory Ratio to 12
Set tCL to 18
Set tRCD/tRP to 18
Set tRAS to 45
Set tCWL to 18
Set tFAW to 26
Set tREFI to 45000
Set tRFC to 524
Set tRRD_L to 6
Set tRRD_S to 4
Set tRTP to 9
Set tWR to [18]
Set WTR_L to 9
Set tWTR_S to 3
Set Command Rate (CR) to 2
Set Memory Voltage to 1.30 Volts

 

You can download testmem5 to see if you get any error, if there is an error you can raise the tRFC a little.

These are the rules:

tRAS = tCL + tRCD/tRP + tRTP
tCWL = tCL or tCL-1
tREFI = affecting read speed the most.
tRFC = Laptop won't be able to boot or brings error if set too low.
tWR = tRTP x 2, laptop won't be able to boot if set too low.

 

Got a modded bottom panel coming my way

BTW if any of the portuguese GT63 owners around here want a samsung 2666 DIMM PM me!

 

@hackness can you send me a link to download memtest5? (Never use it)

@Shark00n nice, much difference on temps?

I'll be off this weekend, probably I won't test anything this weekend. I'm getting another hynix module next week to get DualChannel.

 

I want sent you a pm in zwame forum.

 

Download testMem5 here:

http://testmem.tz.ru/testmem5.htm

Please use google's web page translate, it is in russian, but the program is in English.

Extract it, right click on the TM5.exe and select property, go to Compatibilityand tick allow to run as Administrator.

Now run the program and wait for it to finish the test.

 

Hi everyone,

Would just like to get some opinions regarding voiding warranty to re-paste.

In a summary, my CPU currently has a core temperature differential of 13 degrees (Celsius). That being the difference in temperature between the coolest and hottest CPU cores. So under load I have two cores nearing 95 degrees, and the remaining four run at a decent 80-85 degrees.

Being an IT professional, it would be easy for me to correct the problem, but in the process of doing so void the warranty (Heatsink removal). I know most would say to send it in to get it looked at, but I have a fear of doing so given my past experience with the local MSI service center and their chosen delivery service provider (Startrack).

Does MSI sell replacement parts for reasonable prices in the event something were to fail? Does anyone have any experience to share that may put my mind at ease.

Thanks

 

Warranty is NOT voided anymore (it's illegal to void the warranty in the USA now anyway, them doing that will get them sued) as long as you do NO DAMAGE from the repasting.
If you repaste with liquid metal (hope you did your research and study and applied foam dam barriers and super 33+ tape), make sure all LM is removed, the heatsink is polished with 3000 grit sandpaper back to the copper, and normal paste reapplied before even considering thinking of a future RMA).

 

Thanks for the response Being that I am in Australia, I dont believe there is any law to cover me on the repasting. I would only be using Arctic MX-4 and the appropriately sized Arctic Thermal Pads so I have no worries there.

Prior to taking the heatsink off (assuming I decide to proceed), does anyone have an image showing thermal pad placement and thickness on the GT63 8RG?

Would be of great help in preparing for the job

Sent from my HTC 2PS6200 using Tapatalk

 

MSI uses 1.0mm thermal pads, make sure when you screw the plate back, you apply force to each screw evenly and lightly in circular motion without tightening it all the way and tighten each screw gradually until all of them are secured.

 

I highly recommend 0.5mm pads (CPU heatsink only) to deal with the annoying coretemp issues on cores 0, 2 and 4 (0 and 2 on 4 core systems).
This convex primary heatsink block has been used ever since GT72s and MSI is too lazy to change it (at least the LGA WT75 / F7 or whatever 6 core consumer version will be called has a new heat block).

 

Intel XTU isn't saving the undervolt after reboot, can you please tell me if there's a way on doing that on BIOS? (too many options didn't find the undervolt CPU cache/voltage option) I need to apply changes every time I restart my machine.

I tested the Dragon Center with the OC +120MhzGPU +200Mhz Mem and got better values with this one, the Curve on MSI afterburner for somehow didn't go over 1898Mhz, with Dragon Center 1911Mhz sometimes 1923,5Mhz​

 

Unlock bios menus first.
Then go to Overclocking Performance Menu.
You can set the negative/positive offsets there.
i assume you are using adaptive voltage here. For Static you just key in the voltage and select override voltage.

To be honest, if you unlock the bios, you are better off NOT undervolting and just going to CPU VR Settings->Core I/A Domain and set the IA AC and IA DC loadlines to 1.
This basically does the SAME thing as undervolting without risking idle BSOD crashes. This also makes override voltage actually work without overvolting it by over 100mv compared to what CPU VID shows and what you entered as an override voltage.

This is not guaranteed stable in all MSIbooks though. Works fine on the higher end Titan laptops. Not sure about GS65 or GT63.

 

Don't even bother with Intel XTU. I, and many others use Throttlestop to undervolt. Just set Throttlestop to launch at login via Task Scheduler and you are done (with undervolting on reboot). Plenty of guides out there to explain how to undervolt in Throttlestop. A quick Google Search should help you find what you are looking for. It can be downloaded via the link below:

https://www.techpowerup.com/download/techpowerup-throttlestop/

Sent from my HTC 2PS6200 using Tapatalk

 

How do you unlock the advanced BIOS menu on an MSI GT? Haven't tried on my GT63. Do you mean flash a modded BIOS (I.e. Svet)

Sent from my HTC 2PS6200 using Tapatalk

 

That or the combo keys. Check the other threads.

 

I'm using the combo keys alt+F2 (left side) and ctrl+shift (right side)

@Falkentyne tks for the tips, I'll check it later

(the next step is trying to get better results on GTX1080 and trying to disable/enable in windows but I think its not possible without changing lots of stuff)

@Pureownuge69 On my old laptop with i7 4710HQ Intel XTU worked fine after restarts but not on this one...

 

I wish the GS series also has a tougher heatsink design like this, squishy heatsink is the biggest problem on the GS series
Hopefully there is a slight chance that MSI release the improved version of the GS heatsink and still usable in the GS65, I'd grab one straight away.

 

The fact is that. At this point, the highlighted red thermal pad is the reason that - the heat sink does not reach the processor. No matter how thin the thermal pads are, they still give me uneven temperatures across the cores of 10+ degrees. And only when I remove the thermal pads from these places the difference in the cores reaches the norm of 3-5 degrees, but this is still not ideal, because the heat sink is uneven.
So if you want to reduce the uneven contact of the heat sink, you will have to remove the thermal pads from these places.

 

So MSI. This problem of uneven heat sinks continues from GT62 to GT63. When do you stop making uneven heat sinks ? May be enough ?
Maybe your factory workers have problems with their hands ? then heal their hands please MSI

 

@Pureownuge69

Finally have time to post some basic temp readings after running some stress tests. This is with fans on "Auto" and no overclock on anything. Full 1080 Nvidia GPU. This was screen sniped right after the test was done.


This is with the LM repaste by HIDevolution on both the GPU and CPU


Right now my HWinfo temps are reading this after it cooled back down and these are my normal readings when just surfing the net or doing something basic.

 

GT63 owners, please share your AIDA64 tests (CPU + FPU + Cache + GPU) 5 min. as in the photo. My fan settings are 2nd photo.

 

@Shark00n

Also have you received your modded bottom panel yet? if so, can you provide some photos and and updates on temps, etc?

Thanks!

 

Anyone with GT 63 tried this?
going to CPU VR Settings->Core I/A Domain and set the IA AC and IA DC loadlines to 1.
This basically does the SAME thing as undervolting without risking idle BSOD crashes. This also makes override voltage actually work without overvolting it by over 100mv compared to what CPU VID shows and what you entered as an override voltage.

I'm not having too much time to play/tweak the GT63 (too much work on Portuguese country... )

BTW is there an option to get CPU integrated graphics instead of full-time GTX 1080 with the advanced menu options?! Too much options on CPU and GT stuff...

EDIT: New tests -200mV under test, -190mV already stable and DDR4 had to raise tRFC from 524 to 550 (I don't know if I can lower but as I read somewhere for 3200Mhz I think it's a good value for 2x16gb modules)
Datasheet:
https://uploaddeimagens.com.br/imagens/mem_skhynix-jpg


EDIT2: GTX 1080 +100/200Mhz on dragon center testing GTA V crash after 15min I think its too much on this MXM card. Guess I'll forget that and back to the curve option without OC
The think is with Dragon Center I get better core values over curve in MSI afterburner... (wondering why...)

 

Yeah I got it. Will post some pics later.

Works great. Got a solid decrease in temps, specialy on the GPU side. Tops at 83C after a good gaming session, before was always glued on 90C and slight downclock. Now it's great!

But I don't understand my CPU's behaviour. In BF5 it hover around 3.6Ghz or 3.9Ghz, about 50/50. I don't understand why it doesn't hold the boost longer. Temps are around 86C and it doesn't get close to 90 like before.
A buddy of mine has a brand new GE63 8RF and hid 8750H holds 3.9Ghz for much longer.

Dragon Center is set to Sport mode and I have it on High-Performance.

 

Power or thermal issue, how much are your PL limits?

 

Doesn't seem like a thermal issue, temps are under check. My PL limits would be stock as I don't have the unlocked BIOS nor do I have XTU installed. Only Throttlestop for undervolting.

 

Try undervolt the gpu at 0,900V with a max clock of 1911Mhz with msi afterburner and test. With TS, you not checked any error in limit reasons?

 

@Shark00n

Talk with @JNogueira , he also have the same issue and i think that he solve the problem with the tip of @Falkentyne about tdp.

 

Hi all,

Well, not really
I am currently with UV of -0.145 because the tdp hack made my CPU temps climb as hell with temperature throttle kicking in quite heavily under stress test.

Even with UV I have the tdp throttle, but not as aggressive as with stock voltage.

So I have the same problem as Shark00n.
This is not really a problem. This is all about the PL limits.

So, if the CPU has the TDP reading above 45 watts for a sustained period larger than 45 seconds (I think this is the period), the TDP throttle will start in order to force the CPU to stay at 45W. The UV helps to bring the CPU to higher speeds when tdp throttle is on as it lowers the TDP usage as well, but does not solve the problem at all.

The TDP hack just tricks the MB readings dividing them by 2. By this way the system never sees the correct reading and "thinks" that the threshold is respected all the time.
The dangerous part is about the temps, but worked well indeed.

 

I just find it weird that a GE63 holds it for longer. Even though the fan noise required for it is much higher.

Anyone know of a BIOS update for the GT63? MSI Support site has the same drivers for months now

 

Hi Shark00n. Do you know how long does the GE63 holds the PL above 45 seconds?

Nothing by my knowledge that GT63 has a new bios since past May

 

So with that tdp hack, the problem is solved? If you read the review in this website(notebookreview), i think that is mentioned about throttling...with my GT75, i never got throttling even with long sessions of gaming...3,9Ghz always.

You checked limit reasons on TS?

@JNogueira

This is the same issue like GT72VR that we have in the past about BD-Prochot?

I can check the values for for the tdp in GE63, just link me the bios that your friend is using.

 

Hi Pedro,

No, this is working as per design. Nothing wrong.

The long gaming sessions with 3.9Ghz I also have them successfully but if you check XTU monitoring during gaming sessions you will see that the CPU load is not enough to go beyond 45Watts on the TDP, meaning that there is no power throttle. I observe the same on my gaming sessions where CPU does not go above 50% of usage.

Now try to run a CPU burn test (100% usage on the 6 cores) for 10 minutes with XTU opened to monitor and you will see the power throttle kicking in after 45/50 seconds. This is what happens when intel adds 2 more cores to the previous architecture. Power consumption will increase and of course the tdp limit will kick in earlier and heavier.

This is the CPU specification working as per design. It is not the most pleasant thing for sure, but this is the specification working.

 

But that is the same issue like GT72VR, if we have a pc gaming like GT63, you need the 8750H always with 3,9Ghz...a GT63 is much better than GE63 omfg...GT series...

 

The MSI GT72VR had BD prochot enabled due to insufficient energy (and not temperature as it should be) and when disabled, CPU would throttle but without any throttle flag as well as the GPU.

Here, this is the CPU specification working. Contrarily to the MSI GT72VR, here there is no such problem.
You can find all these details on the intel page of 8750H as all over the internet. There are brands that are limiting it to 35W and even to 25W due to bad cooling solution. This is done by firmware.

On the other hand Acer made an option available to allow 45W+ TDP without time limitation (based on the same principle of your assumption), but this is a specification override made by the manufacturer. This is not the way it works as per design.

Hope that MSI does the same, but on the GT63 there is nothing wrong with this behavior.

But here is some info:
https://ark.intel.com/products/134906/Intel-Core-i7-8750H-Processor-9M-Cache-up-to-4-10-GHz-
It is the way it has to work. The CPU must work on 45W TDP on a sustained way. It allows override for a certain amount of time (Turbo Boost), and then gets down to 45W again, showing power limit throttling.

 

You can do the same as acer in Bios or TS.

 

JNogueira that is wierd did you ajust the timer on intel XTU to 128sec? I never get power throttle and when I get on stress test after around 3min it goes from 3,9 to ~3,7Ghz, but stock without the -190mV undervolt it goes down to 3,1Ghz on stress test.

 

Hi @ldcosta

I am not able to do such aggressive UV on my machine unfortunately. tbh I've been lazy about doing UV with different value under load and on idle. I know that on full load I am able to have even higher UV numbers, but still didn't do that.
I know that when UV is as good as near -0,200v you will stop to see Power Throttle. But on stock, I am having exactly the same behavior as yours. 3.1Ghz when Power throttle kicks in wothout UV.
Currently with -0.145v I am able to stay at 3.4Ghz / 3.5Ghz when the TDP is forced to 45W.

I really have to stop using XTU and stick with TS.
I have set as 47 seconds on XTU. Let me try with 90 seconds (max allowed on my situation).