***The Official MSI GT80 Titan Owner's Lounge***

They should. Specially current titan available. We are still on the same GPU generation and MSI promised GPU upgrades for two generations. At the very least, I should get pascal and I am not sure how much time pascal will last haha.

 

My order is in inventory now!

thanks @pat@XOTICPC!

 

Got my GT80 today finally, noticeably louder than the Alienware 18. CPU performance is incredible in WinRAR. L4 cache FTW.

2461.49 US Dollar Before tax

 

You can get it straight from MSI
http://www.msi.com/collections/pages/MSI-Keyboard-Wrist-Rest.html

They have all of their MSI swag for sale online:
http://www.msi.com/collections/pages

Update: It looks like the online purchase option is missing...

@MSIGeno where can we order items from the MSI Dragon Collection, there used to be an online order option, I don't see it any longer...

 

Congrats on your new GT80 SLI-263

That's a crazy low price, where did you find it?

 

NCIX

 

That can't be a "new" price, right? Or was it a special timed price?

Right now it is $4299.99 CAD = $3259.3164 USD
http://www.ncix.com/detail/msi-gt80-titan-sli-i7-cf-109502-1066.htm

Link?

 

Open Box with only 1 in quantity. I bought it the day my AW18 died. I've never added an item to my cart and purchased so fast in my life.

Newegg.ca has it for $3999.99 CAD so whatever that is in USD...

 

Please tell me you have some sort of 1yr-2yr warranty with that thing, or you at least purchased one.

That is a lot of money for an open box product, so I hope you got one .

 

Standard warranty + 3 month extension + 1 year credit card

total 3 years 3 months

 

I'm currently testing the same thing with the same results.

I don't like having to shutdown my computer, because I like the speed and saved state of sleep, but if it prevents crashes I'll live with it.

Sent from my Nexus 6 using Tapatalk

 

Hmm... Numpad activation is having problems, and it's only solved by reinstalling the driver each time. After reboot it goes back not responding (Win 10 + MSI guide to disable hidden updates). I'm 99% sure this is W10 being garbage.

CPU cooling is absolutely horrid. My 4710MQ could run 3.6Ghz @ 0.9863v and be cooled to 40c @ load. This CPU reaches 94c in 3DM physics 3.6Ghz 1.065v with stock fan control. Probably needs a repaste. I can't keep the wattage below 47W @ 3.7Ghz which seems about normal. I get roughly 48-49w with stability @ 3.7. This is around 1.08v. Yes, I have a cooling pad. CryoLX 3 fan.

I've even had a few 100c shutdowns. Because of the crap temps, my highest 3DM11 w/ windows 10 is just shy of 20k.

http://www.3dmark.com/3dm11/10581091

To make up for the horrid CPU cooling, the GPU cooling is absolutely overkill. It's roughly 10-20% better than the AW18 in GPU cooling. This is probably ideal for gamers anyway, but makes for a horrid benchmarking machine.

The CPU temps getting high means the fans operate loudly when they really shouldn't be. MSI needs to correct this in later builds. The AW18 can operate near silently.

The keyboard is much better than the 18's (obviously). Screens are the same panel I believe.

In regards to repasting the CPU, does this require taking off the GPU heatsinks to get to the CPU?

GPU1 ASIC 71.8%
GPU2 ASIC 73%

Also, does this notebook make use of dual PSU? I assume it does since there's no wattage limit.

 

I would certainly look under the heatsink with temperatures like that.

 

Definitely look into repasting. My temps with my CPU at 3.6ghz since it is a bit cold with ambient temps on the 77F range, I get max temps of 70s when gaming and 80s when really pushing it. I also seem to run at 1.024v at 3.6ghz.

In theory yeah the machine does use dual PSU as there is an user here that already tested it and worked.

 

I managed to get down to 1.05v on the CPU. The vdroop control is non-existent so it's hard to find a stable voltage.

How long have you been @ 100Hz for the display?

 

Hmmm, hard to tell, maybe 9 months now? I have tried multiple refreshes too, but not higher than 100. I used 96hz most of the time, around 6 months, and just recently testing with 72, 75, 96 and 100.

 

It's easily the most EPIC feature, haha. What a difference @ 100Hz.

 

Even at 70hz already seems like night and day! I can't use my normal work laptop at 60hz without the feeling it runs slow haha. It is awesome to run so high fps

 

Well I have done a lot of testing with shutting down, using the latest drivers and general stability: It seems the culprit for nvidia drivers issues is related to fastboot/hybrid shutdown. After disabling this, I haven't had a single crash, not even after hours of use and shutting down, rebooting my unit.

Users that experienced issues with the latest nvidia drivers, try disabling fast boot/hybrid shutdown by disabling hibernation and following guides online for it.

Forgot to mention Zero989, I have no issues with the numpad on win 10. I even updated recently to the latest drivers on the MSI page and it works fine. I haven't had the need to reinstall ever, or maybe when I first got it back at the beginning of the year.

 

How exactly does one go about disabling fastboot/hybrid shutdown on Win8.1? Is there more than one way to do it, and is one better than the other?

I always shutdown my laptop from Win8.1 each evening, and that does not prevent crashes with the new drivers once I boot up again - so does that mean that this disabling fix won't help in my case?

 

In a Windows cmd window started As Administrator:

powercfg /h off

or, if you get an error, this will work instead:

powershell -Command "Start-Process 'powercfg.exe' -Verb runAs -ArgumentList '/h off'"

Then reboot.

That will disable Hibernation, which disables Fast Startup - the 2 disabled disconnects them from influencing Shutdown / Startup - and stops those issues.

 

Is this the same thing, but from the UI?

http://www.tomshardware.co.uk/faq/id-2082951/disable-hibernation-windows-windows.html

More specifically the "turn on fast startup" checkbox. Hibernate was already off for me, but the turn on fast start up was checked. is this what is causing the problem (which would explain things if it was, as fast startup only works after a shutdown, not a restart).

 

I always disable hibernation to save SSD space.

Liquid Ultra applied. I just went through a full syringe...

Lol, HUGE difference. Almost like it's shaved 25c off.

CPU Temp 86c/87c/86c/83c (Package 87c)
4.1Ghz (4 cores) 1.24v

GPU 1 55c 1.062v prema @ stock clocks
GPU 2 61c 1.062v prema @ stock clocks

That's after a 3DM 11 run

Once again, CLU does near nothing for GPUs. They already ran that cool. Same experience was met with the AW18, where CLU did very little for GPUs. 880Ms & 980Ms.

 

I disabled both fast start and hibernation and went through with what hmscott posted. Once you disable, try removing the drivers using DDU and install the latest drivers.

Since Windows 8 and 10 both have hybrid shutdown by default on. This makes the regular shutdown be actually somewhat a hibernation, thus not really a full shutdown. Once removed, we properly shutdown similar to restarting during usage. I have had no issues at all since I removed them, give it a try!

 

I agree, the GPU cooling is already good as it is, it is the CPU that stands the greatest benefit. Nice numbers you got there

 

It's the same, and not the same

The video is deceiving because it shows the Fast Startup option as still available, after disabling Hibernation. But, unchecking the box doesn't complete the task, you need to reboot - note my step for "reboot" after running "powercfg /h off" in a Windows cmd window Run As Administrator.

or, if you get an error, use this instead:

powershell -Command "Start-Process 'powercfg.exe' -Verb runAs -ArgumentList '/h off'"

After disabling Hibernation, through either method, and rebooting, the option for Fast Startup is no longer available. Hybrid shutdown / Fast Startup needs Hibernation enabled to function.

Spoiler

The video was incomplete, it didn't show you the result of unchecking Hibernation and rebooting.

 

Some other thoughts:

5950HQ has the Iris Pro which means the heatsink is cooling 2 chips. It will never run as cool as the other chips.

The CPU heatsink actually cools 3 motherboard chokes as well as Broadwell, AND the Iris 6200 Pro/L4 cache EDRAM. That's 5 things... LOL. WHY MSI????

MSI GT80 does not play nice with Throttlestop 8 Beta 5. XTU doesn't allow static voltage, so TS is the only solution, however, after some playing around, launching TS starts to cause system freezing even at a once before stable voltage. Perhaps BGA and static voltage don't mix. The Alienware Area-51 R2 (MSI motherboard) doesn't like static voltage either so this isn't new to me.

TS also caused some IRQL BSODs. Hopefully all is well now.

 

"powercfg /h off" then reboot fixed me as well.

If you get these:
EventID = 42 Message = The system is entering sleep. Sleep Reason: Application API
in EventViewer, then you did it wrong.

I actually just realized this last Tuesday thanks to hmscott.

I am also noticing a pattern in my event viewer history such that when I thought I was shutting down, upon the next boot, if I didn't get:
EventID = 2 Message = Intel(R) Management Engine Interface driver has started successfully.

Then I would get the crash.

I wonder if there is some sort of API disconnect between IME and NVidia drivers.

 

On this notebook, I believe so.

 

I think mine is not so good as yours, I am starting to worry about it,
firestrike score
http://www.3dmark.com/3dm/9457752
my temps attached,

what i can do to reduce the temps?

 

Undervolt
Repaste
Cooling pad

 

Why use prema vbios at stock clocks?

 

Well, I wasn't going to keep them overclocked. I'm actually already back on the stock vbios.

 

Went ahead and bought myself one hell of an expensive Christmas Present from GenTech PC.
Ken has been very helpful and prompt answering questions. Cant wait to get this bad boy with the custom setup!

MSI GT80S Titan-072 SLI GTX980 Desktop GPU

6th Gen Intel Skylake Core i7-6920HQ, 2.9~3.6GHz, 14nm, 8MB 45W

32GB DDR4 2400Mhz Kingston HyperXImpact CL14 Dual-Channel 8G X 4

SLI Dual nVidia GeForce GTX 980 Desktop GPU 8GB X2 GDDR5 DX12

IC Diamond 24 Carat Thermal Compound on both CPU/GPU

Raid-0 2x 512GB 950 Pro NVMe Gen 3

M.2 Slot - 1x 250GB Samsung 850EVO SATA III 6Gb/s

1TB Samsung 850 EVO Series Solid State Drive 2.5" 7mm SATA III 6Gb/s

6X Blu-Ray Re-Writer/8X DVDRW Super Multi Combo Drive (Standard)

Intel Wireless-AC 8260 802.11 ac/a/b/g/n 2.4Ghz/5.0Ghz Dual-Band 300/867 Mbps + Bluetooth 4.0 / WiDi

 

Fingers crossed this has fixed it. Made the change, installed the most recent drivers. 4hrs in and no TDR's. Oh, and Battlefront is spectacular. 1080p, ultra details, rock solid 60FPS and only hitting the 2nd GPU for 60%.

 

Try lowering the processor cache as well. I loose 0% performance dropping it from 3.6ghz to 3.0, and I do gain a bit of power headroom of about 1-2w (less temps, less heat, more core performance).

As zero989 mentioned, I would recommend a repaste and check for a clean system. Also, definitely try undervolting. I am currently running all multipliers at 36x, processor cache at 30, and undervolted dynamic voltage to -75mv. It works great for me. Try to undervolt little by lilttle (negative, not positive voltage!) with intel XTU.

Prema's have better power regulation on stock compared to the stock vbios, as well as higher temperature to throttle (92C, compared to my stock 85 or something like that)

Congrats on the monster purchase! I hope we hear from you once you get your machine to get some numbers of performance!

Believe it my friend, it is fixed. I have ran this setup the whole week with no issues, crashes, anything Now we need to know why it happens!

 

Its so stupid that we've all spent weeks trying to fix this issue, and it turns out to be something so simple, yet so hard to figure out! (as its not logical at all).

 

Most problems seem simple once a solution is found

The thing is, nvidia definitely broke something with hybrid shutdown/hibernation that was not present before. I already updated the issue with them hoping they can fix it.

Sent from my SM-G925I using Tapatalk

 

My concern is not the cpu, but my gpu2 temps, is there a way to switch bitween gpu1 and gpu2?
Clean system? What do you mean?
Repaste ... Witch one i have to choise? What about ARTIC COOLING MX-4 ?

 

My gt80 died. Time to cry.

 

:'(
Sorry for your lost,
What s happened?

 

Guinea pigs tend to do that. ..sorry for your loss ...how long did she live

 

2 days... lol

 

Well your under warranty at least

 

Correction, the GT80 uses a MUX to switch on / off the Intel / dGPU so when in dGPU mode - the mode most people will run most of the time - Nvidia 2 x GPU mode, the Intel GPU is off, not generating heat - or stealing TDP from the CPU

As far as cooling the parts around the CPU, it's most likely to help transfer heat generated by the CPU and transfered by proximity to it - the heat transfer plate is reaching out around the CPU to get as much heat out of the CPU affected area as possible.

It's not 100% efficient as I see heat transfered under heavy CPU load to the M.2 drives sitting on the opposite side of the motherboard.

In comparison with temperature reports of the Haswell models, my 5950HQ runs 5c-10c cooler; it doesn't run hotter - but as with any high end high clocked CPU you can reach the limits of the cooling / power system.

Please share your TS settings once you settle on them. If you can get stable operation at 41x and above, please share that too.

Using XTU I have to add positive voltage offset at 41x and above. My 41x x4 settings are stable at +50mv (1104 xtu score), and 42x x4 at +115mv (1131 XTU score), both at 36x cache +0mv offset.

It's nice to have another 5950HQ owner here to share testing

 

Wow, that sucks. Sorry to hear it, but good thing you exercised it through OC'ing while in the initial return period, so you don't need to go through Warranty delays.

That's why I recommend testing all ports, function, and OC'ing using XTU / Afterburner in the first 7-14 days - the usual return period - so you can uncover any flakiness quickly.

The XTU / Afterburner OC limits are all within spec and perfectly good for long term use.

Testing with them early triggers infant mortality that might take weeks or months otherwise - and then you are in the warranty period which takes a lot longer to resolve.

What are the symptoms? I am sure you tried all the tricks to revive it, what did you try?

 

Oh no, what happened?
What's the symptom?

 

I'm aware it's off during dGPU use with SLI. It's never been supported in tandem with SLI on any notebook.

I believe the eDRAM is integrated into that chip, but I'm not sure how much heat its giving off.

I was able to benchmark several times at 1.047v static voltage. Then it just stopped working so I installed XTU and deleted TS. I never bothered beyond 41 because of the 47w limit. a 28 second benchmark score doesn't excite me. The highest XTU score I received was around 1109 @ 4.1.

Turns on for a few seconds then turns off.

 

This might be worth trying:

Topic: My Titan died!!! Does the GT80 Titan have a reset switch?
https://forum-en.msi.com/index.php?topic=255596.0

Also, you re-pasted with CLU. That stuff is conductive, and might create a short that can cause a power down immediately at power on. You could look around the area under the heatsink's treated with CLU to see if there are any drips - sometimes heat / cool expansion will open up gaps and allow the hot paste to migrate - drip.

It's better if you find the drips than MSI, as they would then likely charge for the service / RMA.

If there is a chance of a short, you might want to disconnect the battery cable, as noted in that above link, to reduce the issue getting worse, until you find all the drips and clean them.

 

Hey all, i'm having another problem. Every few hours, the tiles on the Windows 10 start menu will either disappear completely or turn black and keep the logos there. I am on the 359.00 drivers and it is really weird. Rebooting usually fixes the problem but it would be nice if I didn't always have to reboot just to fix the tiles.. Any help would be strongly appreciated.