*** The Official MSI GS65 Stealth Owners and Discussions Lounge ***

I used MSIs because the latest official drivers are unsupported.


Sent from my iPhone using Tapatalk

 

I have no issues with the latest drivers at all.

 

Anything new in the Svet BIOS that isn't there in the unlocked stock version?

Sent from my BLA-L09 using Tapatalk

 

Honestly I could not tell you, I went straight to getting a bios from him.

 

GE75 my friend.

 

Too blingy for me with the gold hinges.

I’m interested in the Alienware m17, but I’m betting the cooling will suck. Might just buy a 2060 model and LM it...

Edit: misread that as GS75.

 

All good, ya the gold hinges on the GS75 are fugly. The GE75 is a beast!

 

Honestly I’m not a fan of the GE75 either though looking at it. It very much has that gamer aesthetic. I also use my laptop at university for CAD/Civil3D, don’t want to get laughed at, haha.

 

did you fresh install windows ver 1803 or 1809?

Also, nvidia standard drivers or dch?

 

Does the dch version differ in terms of performance when compared to the conventional nvidia driver?

 

I doubt so..

Just wanted to know what everyone is using in regards to updated drivers..

 

For anyone interested in my RMA results. I haven't opened it up since it was returned. I sent it in due to 3 of 6 cores overheating and temp difference on those 3 of 20-25C over other 3 cores, meaining heatsink was not making proper contact with cpu die. Anyway, right before I sent it in I did an AIDA64 stress test after undoing the undervolt. I had also re-pasted. After about 2 minutes the unit went thermal and shut down. I guess I did not do a good job of re-pasting since this did not happen when I ran the same test earlier with no undervolt. Again, not sure exactly what repair they did but pretty sure I got a new heatsink. Running AIDA when it was returned did throttle but it did not shutdown, so clearly was better than before. After the undervolt I was able to get max temp down to about 80C using Intel XTU stress test for 10 minutes. AIDA always seems to throttle. Another problem that occured prior to RMA was on startup the unit would throttle (I got the warnings in event viewer. Post RMA with no undervolt this did not happen. So again, clearly better. I imagine if I repasted and tweaked/replaced some thermal pads so the heatsink makes even better contact, I could get even better temps. For now I'm happy, I'm not running any games (yet) that would put the same stress as AIDA. I'm about to run the tests again because I did notice my temps have improved over the past week (When I first received it back it idled around 34C, now its at 29C). I did read that sometimes thermal paste takes a little while to reach its best performance so this may be why. FYI my undervolt is .1719 and has been stable. Lastly, I noticed a 10-12C difference in core temp vs 20-25C before RMA. I'll report new numbers later today @xLima thought you would be interested in these results

 

Just Ran Intel XTU stress test for 10 minutes with Great results!

 

Mickey, I saw a few pages back that your GS65 RMA was repaired and shipped on the same day. I just sent in my GS65 for an issue with the left fan making a loud noise and turning off sporadically.

They said they received it today, repaired it in about an hour and a half and shipped it. Based on your experience do you think they actually fixed it or was this some kind of error?

 

Ended up caving and ordering an RTX 2060 model. For only $300 more than I paid for my 1070... Meh, why not. Also, my 1070 is a US model, where as I am in Canada, makes me a bit nervous for warranty claims...

I'd have liked a 2070 or 2080, but Canadian prices are crap as is.

I'll use my 15 day return period to stress the crap out of it and see how it is vs the 1070.

WIll have to take both apart to swap my RAM/SSD, so maybe I'll take some cool pictures.

 

Well, they should sent a repair invoice in the box with the return telling you what they did. In my case I had to take their word for it because I didn't open up my unit to look but based on my testing they had to have made a repair, they just must be really good/fast at getting in and getting the repair done. I will eventually take a look inside to take a look at how well they put it back together, i.e. all the tape replaced, all wires routed neatly, etc. Good Luck

 

QUESTION. I've had a problem occur and am not sure if it is due to the undervolt I did. If I shut my units lid and let it sleep for a while, long enough that the hard drive also turns off based on my power settings and open the unit it does not wake from sleep. Hitting any buttons including the power button gets me nothing, just a black screen with the keyboard lit up and power button light on. I need to do a hard shutdown and restart to get it back. Any ideas why this would be happening? I usually shut down the unit when not in use but I do on occasion just shut the lid and go to bed and if I do that now the unit won't wake without a hard shutdown. I see no errors in event viewer. I did extend the amount of time before the hard drive shuts off but not sure yet if that is the cause or if my undervolt is causing this. Thanks for any help.

 

Yes, too steep of an undervolt. Back it off a bit.

 

Have you tried the stock setting and see if you still have the same issue or not?

 

Really? I've never had a direct correlation between sleep problems and undervolt, it's only the CPU not the whole laptop power, and it's only -100mV or so, shouldn't affect sleep.

That's a great idea. Undo the undervolt, then test sleep again, and then do a manual undervolt while in the same boot - don't change XTU/TS and reboot to test - to see if it's just a random sleep issue.

I'd disable the whole sleep / hibernation thing to get rid of any reliance on it, especially with the current Windows version (10), as drivers and Microsoft constant updates introduce changes that screw with sleep / hibernation enough to be annoying, so I just disable it.

Also disabling Hibernation gets rid of "Fast Startup" which rearranges files on disk to "optimize" boot time at boot time - which keeps being done because we are always installing / uninstalling new software / drivers / etc so it never gets to stop "optimizing", this also fixes "Slow Shutdown" issues, and saves 1x Memory sized C:\hiberfil.sys by removing it:

Start a cmd window As Administrator:

powercfg -h off <- disables hibernation which the above depends on

powercfg -h on <- enables hibernation, not worth it IMHO.

SSD based boot up is so fast these days to me it's not worth having hiberfil.sys taking up 32GB on C:\ - also I don't need pagefile.sys with that much memory (16GB or more), so I also disable pagefile, for me saving 64GB on C:\

 

It was a common problem when I had my XPS 15, you’d have a stable undervolt normally, but if the laptop went to sleep it wouldn’t be able to wake. Fixed by reducing the undervolt a bit.

If you check the 9560 folks out I’m sure it’s better documented.

 

Hey, I didn't say I didn't believe you, I'd just never experienced it myself nor seen it reported before.

@Kevin@GenTechPC has a good idea to check it by disabling and using normal non-undervolt settings to make sure the sleep issue is related to another sleep issue, not undervolt, which has been my experience, lots of sleep related issues brought about by driver updates and Windows updates - when that sleep related issue was fixed another sleep issue comes up later at another driver or Windows update - until I got tired enough of it to give up on using sleep / hibernate. YMMV.

 

Really odd that you’re only at 100mv though...

What’s your igpu undervolted to? My 8750h does some really wonky stuff if I push the igpu past 30mv, which really surprised me because I think I was holding something like 75mv on a 7700hq.

 

The exact undervolt value isn't important to the "sleep issue", I was giving a convenient example. -100mV is the average first undervolt attempt value I recommend - the exact value is unique to each CPU - and I have undervolted a lot of CPU's.

Giving an exact final undervolt value isn't relevant to the sleep discussion. It's just a handy quick example.

Remember I haven't seen an undervolt affect sleep, and when setting up a laptop I soon after disable hibernation and don't use sleep either, so I wouldn't have an example of a too low undervolt affecting sleep, or a reduced undervolt to "fix" sleep.

I look forward to his disabling undervolt and seeing if sleep starts working again, and if so how much reducing undervolt enables sleep again. It's new to me, and I've been doing undervolting for decades on a wide range of OS's... exciting to see something new after all that time.

I'm hoping it has nothing to do with sleep working, a coincidental type situation that can be debugged out of the equation, as there are so many other sleep issues I've seen I find it difficult to recommend tuning undervolt for sleep function on top of everything else.

If the undervolt is stable with prime95 small fft (AVX disabled), with threads not exiting from math errors, for say 10 minutes, I've found everything else works fine.

If you are seeing unstable sleep function from a too high undervolt, I think you can detect it better / quicker using prime95 thread stability quicker and easier - more reliably.

 

Ahhh.

Well, at any rate, I’d just disable your undervolt and test to see if it still happens.

If it does, paw through event viewer and see if anything of note happened around when it entered sleep.

 

Thanks for the responses. I will troubleshoot today. I will run stock voltage and see if I can replicate the problem along with letting it sleep for a while and letting it sleep for a long time, etc then re-do undervolt and change level, etc. I'm pretty sure the freeze only happens if I its sleeping AND the hard disk turns off. I will report back when finished.

Which brings up another question, is there any need/advantage to having the "Hard Disk turn off" in power settings with a SSD? @AWholeHam @hmscott

 

I disable "HDD" power savings, as there isn't a particular advantage for power savings given SSD's handle that in their firmware, and resetting the SSD power outside SSD control could cause issues with delay starting up. Same goes for HDD's really. I also disable USB power control as it can screw with USB devices / USB HDD's. Same goes for ethernet / network, the devices themselves have power control internally. I'd try disabling all the Windows power control of devices individually and see how your laptop responds with all disabled.

Usually I do this for the High Performance and Balanced modes (I don't use Windows 10), and save the power savings mode for "power savings", but I rarely use battery anyway, so disabling power savings in devices is also a performance / desktop mode for me.

If you are having wake / sleep issues it's a good place to disable all those Windows power savings options and see how that helps, and then enable one at a time if you really need to get more battery time - keep them disabled for your AC power use.

 

Well that was fast, my 2060 GS65 has shipped and should be here tomorrow...

I'm pretty busy during the week, but on the weekend I'll do some comparisons vs my 1070 GS65 for any of you interested in the upgrade/sidegrade. I won't be opening it up right away as to preserve my ability to return it, should it not bench how I expect.

 

When you have time would you please run the 3dmark Port Royal benchmark and post the results?

Nvidia reduced gigaray's throughput for laptop mobile RTX GPU's, with a High/Low value now for mobile 2080 / mobile 2070 and the mobile 2060 is now reduced from 5 Gigaray's to 3.5 Gigaray's - both High/Low, I made a couple of posts about this in the Nvidia thread, here's the chart + benchmark graph with links to the posts:


See that the Port Royal RTX 2080 laptop GPU is down around RTX 2060 desktop performance, so the 2060 would be much lower than that.

RTX arrives in laptops
http://forum.notebookreview.com/threads/nvidia-thread.806608/page-173#post-10856786

Given the already poor RT RT performance in game on desktop RTX GPU's, the laptop RTX GPU's give 50% of FPS with RTX ON with an RTX 2080 laptop GPU; with an RTX 2070 / RTX 2060 laptop it's going to be even worse.

Given these greatly reduced RTX laptop Gigaray specifications, and confirming benchmark results I'd venture to conclude RTX is useless in RTX laptops.

 

Can do.

I just purchased the full version of 3dmark too, it's on sale for $5 CAD on steam.

I honestly am not even interested in raytracing, I went to a 2060 because I am interested in what DLSS holds for the future. If anything I may use raytracing in a few eye-candy single player games (I'm usually an Xbox One X gamer, I can stomach 30FPS just fine for good looks).

 

DLSS would be included in that performance reduction I would think, although so far there isn't even a benchmark or game to measure / compare DLSS performance...only demo's on prescripted rails - to look enticing?

DLSS was paired in the RTX release to make up for the performance loss with RT RT...not sure if Nvidia would have kept DLSS performance metrics in line to go with the reduction of RT RT performance?

 

It seems that the lower end cards benefit more from DLSS vs the top end (from what little I can find).

I am very interested to see how this will run Metro Exodus... My hopes are not so great, considering a 1070/2060 are the minimum spec for high 1080p.

The problem with being a student is there is no budget for a desktop on the side, and mobility is paramount. So few compromises exist for high end laptops. If the 2060 isn't adequate I may return it and get a GE75 with a 2070, just not a huge fan of the aesthetic of it (seriously, why are there so few mid-size gaming laptops that don't scream gamer?).

 

What seems to work best as reported from students that are gamers, is to build a nice gaming desktop and use a 2in1 carry laptop - non-gaming - for class.

For taking notes and doing google / chrome a nice AMD APU laptop has better GPU performance than an iGPU Intel device, and you could do some light gaming on that in a pinch - but battery life and student work functionality with light weight and cost is far more important than including gaming weight performance in your carry device.

You can change out the 2in1 yearly as it wears out, and keep the desktop upgraded with new "pieces" over time - AMD Ryzen CPU's and PCIE GPU's can be swapped out easily.

A full performance thin gaming laptop that's fragile, using it as an every day carry, isn't sustainable unless you have a lot of money to have a spare 2in1 to take over when the gaming laptop is in the shop.

Either way, I'd budget to have 2 computers in case your only computing device crap's out and needs repairs.

When you are in class every day you don't have time to put your only laptop in RMA turnaround, which can take weeks.

 

The problem is that I am in engineering, so I do actually need some horsepower on the go, mostly for Civil3D and structural analysis. I also spend very little time actually at home, I tend to live in the study areas.

I'd be tempted to go for a quattro based workstation laptop for this, but I do also like gaming now and again and I've read those cards are just not that capable in that area.

I have a ratty Toshiba Tecra (6700HQ/930m) that I use as a backup, but it bogs heavily even trying to render a simple cut/fill, that is if it doesn't just randomly BSOD when you boot C3D. My girlfriend also has a newer inspiron kicking around that never gets used, so its the backup backup.

 

Depends on how much you really do on the road vs at home, and how often that added power is needed, everyone has different needs, but most don't have need for a discrete GPU level performance while mobile in class.

Even so 90% of your time is likely spent doing mundane note taking and browsing, which can happen on a much wider range of inexpensive light laptops / 2in1's / tablets.

The workstation models are usually overpriced, you can game on the Quadro GPU's, it depends on the app as to whether Quadro is better performance or the same as similar GTX (RTX) GPU performance in modeling.

I hope the GS65 works well for you.

 

Sofar the 1070 model has been great to me!

I only ended up ordering the RTX as I paid very little for my 1070, I stand to make a good profit selling it and that really bridges the gap to the RTX.

I had initially just planned to wait for the RTX, but when I saw that 1070 come up so cheap (and brand new) I ended up buying it just to see if the computer was suitable for me.

As far as note taking goes, I'm super old school. Fountain pen and paper, haha.

 

I recommend you not giving much tought to what others think of your machine. You bought it for a purpose, not for the looks. I have used gaming laptops in both meetings, at work generally and even while I was in college. And back then, several of the available gaming laptops had even worse gaming aesthetics (my GX660R was like a nightclub).

I work in the power generation industry and I have seen all sorts of laptops, from generic Dell, Lenovo, to even random engineers using MSI's GT72 or HP Omens. And none of them use them for gaming, they use them for work and the GPU they need. Trust me, nobody cares, its such small thing for someone to give you flak about.

Anyways, if it's too much, you can always put a skin on them and go undercover haha.

 

It's not even me caring about others, I just really really hate the aesthetic of most gaming laptops. Very sharp/angular/colorful. I have always been a very subdued style type of person. I did have one of those Asus stealth bomber looking 17" laptops back in the day that I rather liked though...

I think I will just eat my distaste and buy a GE75 if the new GS65 doesn't perform as well though.

 

The GE75 is fantastic and remember, there are good looking skins that do change the overal aesthetics. We have come a long way to have very powerful thin machines that don't overheat or explode haha.

The Gs65 should be more than fine, I just hope the RTX2060 lives up to the expectations of everyone.

 

I’ve got 15 days to return it no questions asked open or not, so I plan to hammer on it for 10 of those and see if it really is any discernible improvement over the 1070.

The more I think, the more I really want the GE75 though. I’ve just never seen one in person, I have no idea how it actually looks/feels.

Sorry to crap up this thread with my rambling folks.

 

We are here to have our ramblings and help each other. If you want my advise, I would suggest the GE75. I have no enough data for the RTX2060, but I often see it being compared with the 1070MQ and that is never good. I suspect it can be both faster in some ways and yet slower in others compared to the regular 1070.

Go higher plus the GE75 is also thin (compared to my monster GT73) while still offering good thermals and all.

 

Perhaps I will swing by the local memory express on the way home today and see what it’s like if they have one.

Would you bother with the 2080 over the 2070? It’s like a $800 price jump for me. Pretty nasty for what I imagine is very minimal performance gains that OCing could make up for.

I’d probably repaste the GE75 with LM too. I didn’t want to do it on my 65 because of the awkward layout.

 

I personally find the 2070 useless, but that's because I have a GTX1080. I suspect the highest OCd 2070 will reach a stock 2080, while the 2080 can still go further, near a 1080ti.

but for an 800dlrs premium over the 2070, yikes. I would probably stomach it if it were 400dlrs tops.

The 2070 is just as capable as my GTX1080, so its still fantastic, and should be a bit better in higher res.

 

Ah hell, I’m gonna go pick one up tonight to test. They also have a very open ended return policy.

My credit card provider will think I’ve gone crazy...

 

Well @hmscott We think alike, read on. I tested my system with no undervolt and with undervolt by closing the lid and letting it sleep for 30 minutes. In both cases, opening the lid woke the computer. Next I turned off undervolt, shut the lid and left it for about 2 hours. I had the "turn off HDD..." in power settings set to 1 hour. When I got back, opened it up, wiggled the touchpad and it woke up. Next I turned the undervolt back on. Set the "Turn off HDD..." to 13 minutes, shut the lid, had lunch, came back in about a half hour and opened it and it woke up. Ugh, could not replicate the problem. I do know the not waking without doing a hard shutdown has only happened twice and both times after I left it idle for longer than 30 minutes. So this tells me that sleep and turning of HDD in power settings alone are not the cause. Something else must be happening when I let it sit in sleep mode for a long time. I do believe the undervolt is a contributing factor since I did leave it for 2 hours and it woke woke without an undervolt. So up next I am disabling all "turn off ...to save power settings" in device manager. I just did this before I read your post @hmscott, hence we think alike. I also set turn of HDD in power settings to 1 hour. Leaving my undervolt active, I'm going to close the lid and wait 2 hours. If it freezes again, I'm going to change turn off hdd in power settings to never. However, I don't believe that is the problem but I need to be sure. Leaving it sleeping for a long period, more than 1 hour is causing something, I may have fixed it by the device manager settings. If it does not freeze after this test, I'm done testing, no need to beat a dead horse, if it does freeze, I'll try the hdd to never option.

 

I'm trying to replace my keyboard on my GS65, but I can't seem to find the right part. All of the parts on eBay like this one have two ribbon cables, including a smaller one in the middle. Mine only has one ribbon cable and then a little FPC connector in the bottom left, does anyone know what's going on and if there's just two different models of keyboard for the GS65?

Visit my reddit post for more info: https://www.reddit.com/r/MSILaptops/comments/anhkls/gs65_stealth_keyboard_replacementparts_question/

 

After trying different setups to try and get my computer to freeze after sleep as it has done twice, the results are in. The undervolt was NOT the cause. After disabling ALL options in device manager that say, "allow this device to shut off to save power" the freeze after a lengthy sleep went away. Don't have the time or desire to narrow it down to exactly which device was causing it. I do believe the undervolt contributes meaning a the combination of a device shutting down and the undervolt was the cause.I'm no expert in the electronics of these systems but my guess is a device such as the touchpad was shutting down to save power during sleep and due to the undervolt when I opened the system not enough voltage was available to re-power that device or something like that. Someday I may narrow it down to exactly which setting was causing it but it takes me 2 hours each time (to keep testing consistent) and there are a lot of devices to check. Anyway, if anyone has this problem, now you know how to stop it

 

This is the correct cable:

 

I am aware of what the cable is supposed to be, but the connector that the cable goes to on my keyboard broke, so I need a new keyboard that has that specific connector.

 

@hmscott https://steamcommunity.com/games/223850/announcements/detail/1751240527163977726

DLSS Test on 3DMark

Sent from my BLA-L09 using Tapatalk