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

Did anyone notice that sata port for hdd is only sata2 standard and gives max 260 mb read/write ?

And another problem :
https://forum-en.msi.com/index.php?topic=253189.0

 

Yep, such is the limitation of the chipset for intel machines right now. Most channels that are Sata III are already used by the m.2 drives, so the dvd and 2.5 inch drive work at lower speeds. I think this has been true for most MSI machines as of late.

As for the additional problem, the user needs to enter to the synaptics drives properties, re enable tap zones, and then try again. The same thing happened to me a while ago.

 

Weird everyday I need to tourn on tap zones ;/.

 

UPDATE: This story is one of those.. it's so outrageous it has to be real. I was told to back up my system before hand so I could reinstall with the new drive. So I bought a new USB to do so. Today, the day I get the new SSD, I get an email saying they sent DVDs to reinstall because it contains drivers for the raid so I didnt actually have to buy a USB at all.

Heres the kicker: Package contains the 5 DVDs. And NO drive. Now at this point.. as comical as this is.. I feel a bit off because I know I would assume I was lying. A missing drive on the original product? Replacement drive not shipped along with DVDs? Like.. You can't make these things up.

Your move MSI.

 

47w TDP is not an issue, since it is the design of the machine. What it means is that it will limit the maximum overclock of the CPUs, though it hardly affects 4720HQ, it certainly will limit the 498hq with its insane limit of 4.4ghz, which should be attainable in short bursts at a time. They still perform as they should and score tremendously well, and on stock with no OC, they rival desktop CPUs easily. Specially for gaming. It would be great to be able to squeeze more performance out of them, but we have to thank intel for this one haha

Despite the large size of the GT80, it is still considerable less heavy, as well as much more portable, than a desktop of equivalent performance. Carrying around your case, screen, etc around your house is a chore. I can still game in my room, living room, computer room, whatever without issues, or take it to a friend's house etc. I am just too used to laptops for main use, that my desktop is merely a server.

the included screen is of amazing quality. Sure 60hz is not "much" to the power of dual 980ms, but it helps in that you can max out everything at 60fps without issues so far. As time goes by, this will be increasingly harder to achieve. You can put 3 additional external monitors with the GT80s, and yes you can do 120hz 1080p, or even 3 monitors at 4k 60hz * (HDMI output might have to use chroma subsampling, I have yet to test this on my tv, otherwise it only does 30hz). Keep in mind things like uniengine heaven maxed out still run around 60-80fps. Who knows how demanding future games will treat us.

Compared to the competition, the GT80 sports a high quality screen from the get go, as well as a mechanical keyboard, and excellent sound system. It is also officially upgradable for at least two more GPU generations, so that's also nice (though competitors most likely will be unofficially supported as well). While the CPU is socketed and can't be upgraded, there is not much where to go in terms of CPU upgrade regarding gaming. Sure you could get a socket CPU and tweak it to get 4ghz or more, but that will hardly translate to earth shattering performance difference even compared to the 4720HQ. It would help when running at high fps, to maintain higher framerate with external displays, but it will depend on the game.

 

Wow. That was a lot of great information.

Linus said that they keyboard could be awkward to use without the pad behind the keyboard? Have you found that out to be true? Also he said that the screen sits a little farther away since the keyboard is so close. Have you had a problem with that? Or is 18.4 the perfect size for that as well for viewing distance?

So what you are saying is that 47TDP is perfect stock. That is nice to hear. Probably perfect enough for gaming in any case.

As far as the 57 TDP 4940MX though, could you not safety clock that to 5GHZ with two power supplies attached? The only major problem like in the Sager NP9377-S is the face that it does not have m.2. So if you did get a little more power out of the processor you would notice the bottleneck of the SATA connection.

 

Keep in touch with them. MSI is not a single big store where they ship things by the same individual. Chances are, they sent both things from different shipments, and you got the DVDs sooner. Definitely keep in touch so you can receive your missing drive and then perform the installation.

 

How odd, I only had to do it once. Have you reinstalled drivers? I reinstalled using the ones from the webpage, here http://download.msi.com/nb_drivers/...4__Win7_Win8_Win81__Signed__msi_18.1.29.7.zip

You can follow the steps by kranebrain here:

This fixed the button lag, as well as the tap zone thingy for me.

 

Glad to be of help.

Using the keyboard without the palm rest is no different than using just about any desktop keyboard. They don't have. typically, palmrests. But lucklily you get one free! I have used with and without, and the only time I felt it kind of weird, was when I was typing on a very small table without it, and tried to game a bit on it. But that was one specific case of a very small table without space for my wrists.

For the 57w 4940mx, you would probably push around 4+ghz constant without surpassing the TDP limit, but you can change the tdp limit I believe, for those CPUs. The extra CPU performance won't be as noticeable, and 5Ghz for most cases is even unattainable without proper tweaking, modding and monitoring. I'd say 4ghz is the most typical scenario for those CPUs for the average user. I don't doubt the performance of those CPUs, and it would have been nice to have a socketed CPU that could push it, but honestly, for gaming, you won't see much performance increase. If the 4980HQ could push 57w, you would have a clear choice for performance. For now, if you need the extra CPU power, the 4980hq should not be far off from the top.

In the end, it is a choice. Some will value the extra CPU speed over the mechanical keyboard, others might not I personally am happy with current 4720hq performance in gaming and in general, so I don't mind it.

 

We will see tomorrow or maybe during my absence my wife is hacking something ... quick small litle fingers .

 

You should read this review[1].

BTW, have you guys experienced the "burned" background?

"However, the review unit had a problem with another bug. After a certain period of inactivity, the desktop background was temporarily "burned" into the display. We do not deduct any points for this because we could not verify the phenomenon without a second device. MSI told us that they could not reproduce the problem."

[1] http://www.notebookcheck.net/MSI-GT80-Notebook-Review.136724.0.html

 

Prior to my sending my notebook to MSI for evaluation, I never experienced this temporary 'burned', desktop burn-in into the display. With regard to system noise and CPU temp, seems NotebookCheck concurs whole heartedly with my findings.

In spite of anything negative reported about the GT80, I really like this machine. I see the GT80 as a prototype..... and the proving grounds, so to speak. Keep in mind this is MSI's first go into the SLI arena.

MSI has an outstanding record of listening to user feedback, and I can't wait for MSI's next flagship...... with next-gen Intel and NVidia. It is sure to be much improved, and AWESOME! A GT82 perhaps?

 

I haven't had any issues with a burned in screen so far. Everything has worked

 

So let me ask you this. If I get the 4980HQ will I be able to play BF4 for 4 hours and have Intel XTU running in the background and maintain 4GHZ turbo the entire time?

 

I don`t think so.

 

It's a lot better now, CPU maxed at 92C:

 

What did you find or do to improve his issue?

 

Thanks Ken!

Every little bit helps..... Prime95 and Furmark really ramps up the GPU temps. For grins, could you please send my a Fire Strike/HWMon screenshot? Also, Henry got with Taiwan regarding a fan profile tweak, right?



@GTVEVO, would you mind running a Prime95/Furmark test (stock Sport Mode clocks), posting a screenshot w/HWMon? Thanks in-advance!

 

What type of test are you wanting to see, I ran furmark at 1080p 8xaa and max cpu was like 69c and max cpu was 71c? Is this the right test just seems really low so might need the exact testing details to try and replicate without question.

I was hoping to hear that they went back to stock thermal paste so you can rule that out since its the only change on your system and when I tested the same ICD my results were not that great either.

 

I was asking you run Prime95 and Furmark, with HWMon in the background, and a screenshot just as Ken did...... but if this is a problem, no worries.

 

You can clearly see what is going on, thanks to HWmonitor. To give some perspective, my stock untweaked 4720HQ consumes around 55-58W under max load like when rendering, and after 1 minute or so, it switches to regulated 47W TDP. This results in performance between 3.4 and 3.6ghz for things that completely max the CPU. Anything that doesn't max the CPU, will at max 3.6ghz speed.
Initially, it seems the 4980HQ draws up to 82.5w, and then ramps down to 47w as normal. This means that normal max temps are around 87C for 4980HQ, while its short turbo boosts can reach temporarily 92 degrees. There seems to be no thermal throttling whatsoever, just power regulation. Anyone with a 009, can you check with CPU z and/or Intel XTU, what is your core clock speed AFTER the 47W power regulation starts? I am interested in how much CPU speed is there in the long run.

 

I doubt it. Power regulation will kick in and drop you to 3.6ghz or perhaps 3.8ghz, I don't know because I don't have that CPU. But! It depends. Because BF4 doesn't load the CPU to 100%, which means it won't use full power, and perhaps, you could actually run it at 4Ghz, without using all its power? Who knows.
As I showed you, it won't really make a difference performancewise even if you drop clockspeed slightly. You will, however, enjoy BF4 maxed out, at 60+fps, on 64 players maps haha

 

Its not a problem at all, things just seemed lower so I wanted to be sure I was running the correct testing. I have put some screenshots below of running Furmark.

The first image is the Furmark test I was refering too earlier and the 2nd image I ran the CPU burner test and as you can see it hit a max of 80 degs c running for right at 5 min, after the turbo was done with its cycle it sat around 75c for the remaining 3-4 min of the test so I stopped.

If these tests are anything like his I would say there is a considerable difference between our machines and the only variable left seems to be the paste, just my 02

 

Hmmm, I don't think your CPU is being maxed at all there, which is why it shows lower temps Thats good as that means most apps will run at 3.8ghz easily at good temps, and only the very severe ones will drop slightly more, or get hotter

 

Its getting maxed out I am watching the monitor the entire time and its at 100% usage and the clocks at 3.8ghz until turbo runs out then sets around 3.4-3.5.
I will just post a 5min run from XTU

 

Here is my XTU 5 min run, You can see the CPU is getting maxed "as much as XTU does in its case" and you can see the speed and temp at each session.
The first pic shows my settings overall, the second shows 45 sec into the run at 76c @ 3.79ghz the third image shows around 4 min into the run [email protected] , max temp recorded on this run was 76c. This run was done at all high performance settings.

 

Stress test during test usage pic showing 100%.

 

Furmark is an Open-GL GPU benchmark, and Prime95 (download HERE) is a CPU stress test, and this is why I was asking for both. Also, try running Furmark for at least 15 min to for peak temps.

For Prime95:

1. Download the software and unzip the files to your desired location.
2. Run the Prime95 executable and select "Just Stress Testing" when asked.
3. The default options are sufficient to do a well balanced stress test on the system.

Don't forget to have HWMon in the background for the duration of both test, then screenshot as Ken demonstrated. In HWMon, please insure CPU, and both GPU temps are shown.

Here is a like example (Prime95 & Furmark) of my system done by MSI, before Ken worked his magic.... and adding the Dragon Gaming Center to your screenshot would be helpful too.

 

It was Furmark. GPU stays cool with only Prime95 was running. I'll post screenshot with Fire Strike /HWMon later.

Henry still working with HQ regarding fan profile tweak.

 

Just wanted to point out that 100% utilization does not necesarily mean maxed out There are different loads that a CPU can reach 100% utilization, and consume more or less power. That is all. The same happens with a GPU, that may seem 100% on two games, but one will consume much more power and get hotter, because it adds a different load on the system.

Still though, your power consumption seems on a similar "scale" to mine. On intel XTU, for example, I usually get around 50w of peak power consumption, but Cinebench got up to 58W. Both read 100% load but for a small second window, Cinebench drank more power haha.

All in all, those are magnificent temps for the system, not arguing over there, congrats!

 

Ok so the first run for about 8 min and I didn't run in windowed mode so I couldn't take a screenshot of the hw-mon together with the Futur program running. So another 15min later after that are my results, Peak CPU temp was 85c as you can see but it wanted to run after about 10min around 82c-83c and did pretty good staying there, clock speeds stayed at 3.8ghz the entire test, what that the GPU 1 maxed at 78c and GPU 2 maxed at 84c. Keep in mind this is after running several other tests for everyone on here so I think I have done my part in proving what temps can be. Theses were achieved in most cases with back to back testing so very little resting which proves the GT80 cooling is very good

 

Again, where is Prime95? Please read and look at the example in my previous post.

 

Just run Prime95 and then go to options and click Torture Test. On Furmark, first change the resoltuion to 1920x1080 and click GPU Stress test, let it run for 20 minutes and see what's the result on the HWMoitor.

Here are the screenshots of FireStrike:

 

Thanks Ken,

I like those temps, but the Fire Strike scores look like they were done in 'Comfort Mode'..... scores are low. Can you check the MSI Dragon Gaming Center, and make sure the system is in Sport Mode?

 

How long would you like for it to run?

 

This was done my results were earlier, 15min as instructed.

 

It was at Sports Mode, but now I switch PhysX to SLI:

 

Did you run Furmark and Prime95 both at the same time for 20 minutes?

 

Ah ha, now that's more like it!

 

No seems like there might be a few directions here I think I have earned credit to enjoy my evening and try later on.

I ran prime 95 for around 15 min and it hit 81c.

92c on the short run test really seem high don't they. Run skydiver and see as it will be higher. All in all someone else with stock TIM should run these tests for a real final nail to prove ICD isn't helping but yet hurting alittle.

 

In the end, you got the best temps so far. You have CLU installed right? It seems like a very good compound.

 

Right which is why I would like to see another stock player. Mainly I just don't want MSI to get dinged if things are being changed on new machines that doesnt help performance. I am sure everyone realizes that and I don't want it to seem or be hurtful to vendors either, I just want the best for everyone that owns this product.

 

I thought my post HERE was clear, but I can understand the confusion..... and yes, you have earned credit to enjoy your evening, LOL!

When you have time, just merge my request with Ken's info..... post the screenshot, and all will be good in the world again.

As to the TIM application, I think way too much credit is given to one thermal compound, over another. Generally comparisons are very subjective, and outcomes are usually within only a few degrees of thermal variance. Conclusions drawn by many hardware sites, generally agree, that it is the 'proper application' of the thermal compound, which provide the greatest benefit.

Example: Thermal Paste Comparison, Part Two: 39 Products Get Tested..... excerpts.....

Though I am doubtful of any significant improvement over ICD, for grins and to put this to rest, I am entertaining CLU..... we'll see.

Looking at many test results for the GT80, your reported CPU temps are extraordinarily low. I suspect this is due to differences in testing methodology, rather than the thermal compound. To narrow things down, and for more accurate comparisons, I think it most important we are running the same test, same intervals, with the same settings. My guess is if you run both Prime95 and Furmark, as Ken and I have outlined (and screenshot plz!), your CPU will be a bit warmer.

 

Attached is the result of both programs running for more than 16min and temps weren't changing so your requested result is below.

As to the comments in test methods there isn't much difference to make, you set it up and click a button I have posted several shots over and over again of the exact same tests. If I was unsure of the tests I spoke up to ensure I did it in the same way and always wanted to help.

As to the comments with specific thermal applications it seems a bit ironic when people are paying for different compound on one hand yet on the other state there is very little difference between them when they can be measured in thermal connectivity scientifically as fact and I find it useful to stack things to your advantage if you can. Can these results show up as drastically in every application no because there are variables as you state such as application I agree but you can't expect one to work the same way in every application, CLU worked great in another gaming notebook application with a very similar cpu mounting setup and it seems to work very well here. I think if you give CLU a fair try you could find something very beneficial.

Either way I have provided a great deal of data and hopefully some find it useful.

 

I've never seen Intel XTU CPU Stress Test causing thermal throttling. If you want to test for thermal throttling you need to run a different benchmark, such as Prime95.

The best way to test for thermal throttling is to run Prime95 Torture Test with Small FFTs option selected (this will overheat the CPU) and FurMark at max resolution and 8X MSAA (this will overheat the GPUs).
You need to have them both running simultaneously for at least 30min to measure the max temps.
Note that you need to publish also the Fan Control setting and room temperature.
Good luck!

 

I also did some rendering test. I built a simulation and tested with CPU, single GPU and SLI rendering just to see how it performs, the rendering Engine is Cycles which still does not support fire and smoke with GPU but only CPU so in my simulation it was tested with water dropped on the objects:


CPU Rendered, took 32 minutes and 42 seconds:



Single GTX980M rendered, took 26 minutes and 26 seconds:





GTX980M SLI rendered, took 13 minutes and 16 seconds to render:



Fully rendered frame 38 at 1920x1080:




Fast GPU really helps rendering if the rendering Engine supports it.

 

That's so cool. I mean it low temperature by that.
I think you can safely overclock those 980Ms. Have you tried it?

 

Hi guys, I thought some explanation of the how and why to run prime95/furmark simultaneously would be helpful.

It has helped in the past to explain the value of setting up and doing this test, and why using the same parameters for all laptops is important. It helps compare results, and confirm or correct expectations for a cooling system, power delivery, and systemic performance.

Running a stress test that stress both CPU/GPU(s) helps find problems early in the ownership of a laptop, so you can return it, or assure yourself that the laptop is running 100% acceptably, and it is good to keep

Simultaneous. This means both programs running simultaneously - at the same time - both generating heat - both drawing power - all running at the 100% utilization level. One test focuses on the GPU - Furmark. and the other test focuses on the CPU - prime95 small FFT.

It would be nice to know if test / parameter changes need to be made on an SLI laptop

Individually both programs are nice to run to find out the single program hit against the system - and it is valuable to run both individually to get a baseline reading - if you are Thermal Throttling the CPU or Crashing the Display driver when running one at a time, then there is no sense in running both simultaneously - at the same time.

Here is a sample of a run on my G750JH, OC'd on both CPU/GPU, Prime95 small fft, and Furmark Burn-in Benchmark 1080pm 15 minutes. I actually only run it for 5 minutes and then take a reading - the results are the same at 10 minutes, and 15 minutes - but since you are running this for the first time, take a screen shot at 5, 10, 14.9 minutes with Windows Snipping Tool.

We have found that laptop cooling systems interact, GPU and CPU, and while gaming you are at times running both near their limits, this simultaneous test using Prime95 and Furmark simulates that highest stress situation. On the G751, it has been found to show Thermal Throttling on the CPU as the heat from the GPU mixes with the CPU cooling system.

Once you know this situation exists, you can tune for 3/4 core loads by de-tuning the Turbo setting for core 3/4 just enough to stop Thermal Throttling - which kills performance - but not so much as to affect gaming/app performance.

It's a good test to run during the 7 - 30 day acceptance period with the seller - you can more easily return it if you find problems - instead of finding out weeks or months later, requiring a lengthy RMA for repair from the vendor. It's good to stress test all the components, within reason, during the time you have for evaluation toward acceptance.

I like to arrange the display's for the programs I am running, and scroll them to the place of interest, to make a screen shot, or just an area grab, of the information of interest.

 

Actually, if you read the Tom's Hardware conclusion I linked above, it's not uncommon for enthusiast to choose one thermal compound, over another, for 'little difference' or the mere reduction of just a few degrees.

Secondly, yes the properties (thermal conductivity and impedance) of thermal compounds can be 'scientifically' measured, mechanical properties (contact pressures) that can be measured..... and other subjective factors to consider, such as, viscosity and useability..... and still, in all the wading through the myriad of data and consideration, it only amounts to a few degrees. Again, your reported CPU temps seem extraordinary, and I was simply trying to narrow this down, because I do not believe CLU is a magic bullet. If it were, it would be conclusive, it would be the clear choice, and everyone would be using it exclusively. Have you ever Google'd, "Best Thermal Paste or Compounds"? I bet you have, and if so, you know it is a minefield of data, and there is NO CLEAR WINNER!

Thirdly, when I (or the Tom's Hardware quote) was speaking to ' application', it is referring to 'how' the thermal compound is physically applied to the CPU or GPU.

Anyhow, I did not intend to offend sensibilities, and your efforts are appreciated. Thanks for your help.



@Cormogram,

Fully aware, and I am not testing for thermal throttling..... but I appreciate your input. Also, in your reply to Ken, you do realize he is playing with MY GT80, don't you? Quit giving him ideas.... he is already going crazy with it, LOL!



@Ken, GenTech,

Very nice rendering test, good data..... but aren't you getting too happy with my GT80...... you brake it, you fix it! Have fun! (you are soooo lucky I still enjoy my GT72) LOL!

 

I didn 't realize that. lol Overclock it but keep it cool!
According to Coollaboratory CLU will lower 5°C when correctly applied.
That's a lot better then most of the other thermal interface materials that I've tried.
If you are getting even lower temperature drops by applying CLU that's probably due to other factors preventing correct cooling, like incorrect TIM application, bad cooler/heatsink/pipes assembly, etc.