(Part 2) If U Can Solve This U R King

I never EVER got above 655 while u and all others always hit 678 and 690. same applies to my physics scores and cpu scores in all types of 3dmark tests. I DO have an underpeforming CPU.

 

http://nucblog.net/2015/09/dual-channel-vs-single-channel-does-it-matter/

cinebench 15 and 3Dmark tests with single channel and dual channel ram. 3Dmark showed massive improvement but cinebench zero improvement in cpu score.



Another comprehensive testing showing little to no improvement in games:
https://www.gamersnexus.net/guides/1349-ram-how-dual-channel-works-vs-single-channel/Page-3

I'm trying to isolate the root cause behind cpu underperformance.

 

so preliminary conclusion is dual channel might not improve my cpu performance but could improve overall graphics performance. Hence the problem of consistently low cb r15 scores is isolated to my cpu and not the single channel ram mode. i will upgrade to dual chamnel ram anyway so Yreka and co plz stop being toxic about it.

 

Just how do u explain your 130-145fps with your i76700HQ and gtx980m while i only get 110-130fps with my i76700HQ and GTX1070 in the same map in BF3?

I'm curious how do u explain it. Calling bf3 a crappy mess of a game is not a proper technical explanation worthy of this thread.

 

Even more crippled than a huge numbes reviewed laptop with fully bloat. For the last time.... This ain't normal score for 6700hq. When will you finally understand this?

 

Just ignore it. Its normal performance on bloated OS and un-optimised ones. Avg cb scores after optimisation will be in range of 670-680.

 

Nope!! A huge reviewed laptops with 6700hq run with fully bloat as the MSI GS62VR in the review posted in pict above. All reviewed laptops run with bloat. MSI GT62VR can't be much worse than MSI GS62VR. As both is MSI, and all know what MSI put in their laptops...

 

Exactly. This is what is so disappointing in my MSI GT62VR. Multicore score is consistently at 644-655 even when fully tweaked, optimized, and bloat free.

Ironically my Single Core Cinebench R15 score is perfect and close to all reviews. It is 143.

What could be causing my multicore performance to suffer alone?

 

Single ram stick vs. dual will slower Cinebench scores on Skylake/Kaby cpu's with around 1.5-2%

 

yes but as you can see my multicore score is consistently 7.5%-10% lower than average. single channel ram mode cannot be the reason. plus, single core score is perfect.

 

Probably firmware - drivers. Not sure if you use exactly same as the tested MSI models. Have you tested with stock voltage, moderat or fully undervolt to see if you get an difference? And of course... No AV running in the background.

Edit. As well test all voltage scenarios vith wPRIME v1.55 (run the 1024M stress test)

 

In the clean install battlefield ran exacly the same as optimized OS, it's just that cinebench is a garbage way of testing your CPU. I get consistenly 640 - 650 score on clean installed windows, while having 690 - 695 on my non bloated windows, while maintaining the exact same performance in battlefield 3 on either windows platform. My optimized is snappier when browsing the web and stuff tho.

Also battlefield 3 and 4 don't care about dual or single channel ram, it's the same performance.

 

sidenote: I could probably make battlefield run even better if I wanted to, like 160 - 180 FPS. Since the game is a mess I think there is quite some potential in tinkering with it, to bad I don't really play it.

 

As you can see above... Some dont bother see the whole story!!

 

Very insightful, thank you Danishblunt. Can you please tell us what is your exact build number of windows 10 fpr each of the 2 scores? please be patient and share them with us. Thanks.

@Papusan i will perform the stress test on prime v1.55 later today and report back. thanks.

 

They are kinda on the screenshot....It's the same version!

 

You're welcome!! For comparison http://hwbot.org/benchmark/wprime_-...Id=processor_4428&cores=4#start=0#interval=20

 

danish please run "winver" command and tell me the full build number. it takes less than a second. Cinebench does not give accurate win build number.

 

Version 1607 Build: 14393.479

both

 

thanks.

 

so if it is not the sing vs dual channel, and not the bloated vs bloat-free windows, and not the low cinebench r15 multicore scores, then what is making your battlefield 3 (and everyone else's) play at higher framerates than mine? How do u explain it @Danishblunt? other than crappy and messy game.

 

Same performance in campaign, but test 64-player multiplayer and totally different story.

 

@Danishblunt you now are the only person on this forum, and maybe in the whole world, who has the right gear, knowledge, and dexterity, to perform the dual vs single channel mode test in multiplayer and end this argument once and for all. Can you replace your 2 ram sticks with one stick (same total capacity) and rerun bf3 multiplayer metro map? i dont see anyone else who can pull this milestone other than you over here.

 

some links for you
https://translate.google.com/transl...t.com/portopc/dell/dell-alienware-15-r3.shtml
http://www.tomshardware.com/reviews/alienware-15-r3-gaming-laptop,4796-2.html
http://www.gaminglaptopsjunky.com/alienware-15-r3-review-gtx-1060-i7-6700hq/

 

My god this has to be the worst thread in quite some time. Just move to dual-channel RAM.

 

This.

 

Would anyone volunteer to test Battlefield 3 multiplayer metro map with Single channel mode and dual channel mode while keeping the same total memory capacity in both tests? You will contribute to demystifying a major question mark in PC gaming if you pull it right.

I will upgrade to dual channel eventually but not anytime soon. Shelling another $250 is not feasible in the short term, so I can't update my performance readings in this thread until then. Meanwhile, feel free to test it yourself and report the numbers here anytime.

Note: Most tests you find on google are done on Singleplayer games and not multiplayer which is irrelevant to this thread. We are looking for a properly validated, single channel vs dual-channel test in Multiplayer only. The presence of non-scripted human players on the map puts extra pressure on the CPU/GPU.

 

Here is a single vs. dual channel test of BF1 multiplayer (dual channel starts at 7:22):

 

Wow.. huge difference.

<sarcasm> So surprising.</sarcasm>

 

Hello, sorry for my "google translate" english. I watch this forum for a while.
I have а MSI GT62 6RD ( i7 6700 HQ + GTX 1060 \ 16 GB Dual chanel ). I had the same problem
with the fall of using GPU, which fell to 70-50% in games and the fall of FPS, but not in all games.
Аnd it was immediately with dual-channel mode. So the problem of single-channel mode not considered.
I had this problem in Crysis 3. But I found the gameplay of the same game on YouTube, with
GE72MVR 7fr GTX 1070, and Clevo P670RP-G with GTX 1060 i7 6700HQ
And they had 99% using GPU. ,
Besides during the stress test GPU+CPU AIDA64 GPU had 99%.
So this is not a frequency problem, this is not a dual-channel problem, this is not a power problem.
( The problem is that the GPU is not use at 99-100% in some games, but not in all ).

I think that the reason may be the OS, can be a driver, a game, an OS, and pascal problem too, and
laptop collector, just a little from each.
Many people are faced with this problem including desktop and on older generations 9xx, 7xx. But no one
could find a clear solution, except for a few people which reinstallation windows.
If you have tried everything, try different OS... Do not be afraid to install Win 7. This may not be easy,
you need a proper record on the USB flash drive, you need the correct installation of the OS itself. I found
the driver, you will need a "Driver Booster" it himself found, downloaded and installed all the necessary drivers
except one, but it does not cause problems. Some drivers and programs at Win10 compatible with Win 7, such as
Dragon Center, Synaptic, Nahimic they all work perfectly on my Win7.


And your 2-nd problem with temperatures. I set these fan speed settings, in the photo,
and changed the thermal paste to Gelid GC-Extreme. And got the following max temperatures
in hard games and stress tests Aida64 75-78 on the CPU and 70-75 on the GPU. (24 degrees in the room)

 

Thank you so much for your post Semmy.

 

on Windows 10 I had 80, 50% of the use
of the GPU.
After switching to Win 7 and editing some additional processor settings, I have become 92-98% the use of the GPU.
and the GPU became more stable.

What I tried and it did not help me:
1. Update BIOS, EC, vBIOS.
2. Using a different version Win 10.
3. Сhange all power plans
4. Сhange driver Nvidia
5. Disabled G-sync, V-sync

What I tried and it helped me:
1. installing windows 7.
2. editing some additional processor settings.
I also changed C-STATE parameter in BIOS, and now the processor works only 3.1 GHz without falling.
In order to include them you'll have to make changes to the processor management registry key in "Regedit".
( HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Power\PowerSettings\54533251-82be-4824-96c1-47b60b740d00\06cadf0e-64ed-448a-8927-ce7bf90eb35d )
there are 35, But you have to change only some. I Can not enumerate them by name, because I have them in Russian,
and it is difficult to translate technically correctly. I can only specify them in order recounting,
and show what values I chose. The main thing is to disable the parking of the cores and others.
You will need to go on the path indicated in the screenshot. And in each of the 35 parameters,
you need to change the function (Attribute) from 1 to 0. And then they appear in the power plan Windows.

all (from the network)
2 - 100
4 - 3
5 - 1
8 - maximum performance
9 - 100
10 - Reactive
11 - 99
15 - 5
18 - 2
21 - 100
23 - 99
25 - 1
27 - on
28 - 100
29 - 5000
30 - all cores
32 - 5
35 - 100

 

Sorry, I had some problems with editing the messages, go back, try to read again, now they should be in an English. I'm just new here and have not figured out the settings yet.

 

This is exactly my problem(s) too Semmy. Thank you for sharing your story with MSI GT62VR.

I am curious, did switching to windows 7 improve the GPU usage before changing any processor settings? Can I improve my GPU usage in old games just by switching to windows 7 without changing any registry or BIOS settings? or do i have to do both?

 

it is better to do both. Сhanging the bios parameter is not at all difficult, you need to change the only one option and save

 

@Semmy just sealed the deal and proved my point I made some posts prior. Windows 10 is in a state of ****fulness right now and for actual gaming I think you'd be better off on Windows 7 for the time being. I have been running Windows 7 Extreme Edition or something (I paid for the 64bit iso from Amit back when his site was selling it and the 32bit version was free... that was yearrrssssss ago.) for years now and it's been rock solid. I can post up my iso if you can't get it for yourself I guess. It should be available on torrents or elsewhere though. I will be running a dual boot system myself based on my Windows 10 experience thus far and the findings I have been coming across daily from others like yourself having trouble at random. When installing windows 7 however, I would warn you to stay away from certain updates. I will attach my DO NOT INSTALL list for Win7 I keep on hand for reformats. There may be a few additional DO NOT INSTALLs added since I made the list in June or August of 2017 though, but the important ones are all in this list and mostly older updates. Ymmv, this was for my late 2012 HP laptop that originally shipped with Win8.

Iirc the most important ones NOT to install are:

KB2670838 (breaks Aero Win7 - do not install IE!)
and
KB3133977 (buggy update)

KB26700838 completely ****ed my system and corrupted all sorts of stuff irreversibly. Avoid it at all costs. It also degrades graphics performance by replacing the Win7 graphics stack with the Win8 graphics stack which is the cause of the ****ery.

 

I also had one problem with the curve plate of the radiator, which is visible in the photo, so I had throttling. And drops in the core up to 10 degrees. But now I've fixed it partially.
Hello MSI ) if you see me

 

Are you happy with your framerates now in older games like crysis 3 and BF3? By how much did they increase?

 

Damn that is a hell of a warp. That was also mentioned by multiple people earlier on too I believe. How did you end up fixing that gap specifically?

 

It's hard for me to answer in English. And it is difficult to understand a lot of technical information, so I can not always respond quickly. But thanks to everyone for replies

 

crysis 3 и BF3 this is an old game. And with them there are no problems in terms of the number of FPS. But they are not as old as Aliens: Colonial Marines, and in such games is not more than 60 fps, but probably this is due to the limitation of the game itself.
but first of all you need to pay attention not to the number of FPS, but to the percentage of using GPU, it should always be close to 99% and will not change, then your game and FPS will be stable.

 

On my windows 10 and MSI GT62VR, The Witcher 3 runs at 99% gpu usage with normal framerates.

Battlefield 3 and crysis 3 and Ryse Son Of Rome run at 50%-80% gpu usage with lower than average framerates compared to other laptops with same hardware specs.

Do you have a guide to change the processor registry settings in English?

I did not uderstand this part which you posted:

all (from the network)
2 - 100
4 - 3
5 - 1
8 - maximum performance
9 - 100
10 - Rective
11 - 99
15 - 5
18 - 2
21 - 100
23 - 99
25 - 1
27 - on
28 - 100
29 - 5000
30 - all cores
32 - 5
35 - 100

 

it will be easier to show. This part marked red did not allow the plate to lie flat. But the plate itself is still not even. The result is a common design leveled, the differential on the core has decreased from 10 to 2-4 degrees under load, the throttling has disappeared.

 

( This is Part 2 )
all (from the network)
2 - 100
4 - 3
5 - 1
8 - maximum performance
9 - 100
10 - Rective
11 - 99
15 - 5
18 - 2
21 - 100
23 - 99
25 - 1
27 - on
28 - 100
29 - 5000
30 - all cores
32 - 5
35 - 100
First First you need to do this
In order to include them you'll have to make changes to the processor management registry key in "Regedit".
( HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Power\PowerSettings\54533251-82be-4824-96c1-47b60b740d00\06cadf0e-64ed-448a-8927-ce7bf90eb35d )
You will need to go on the path indicated in the screenshot. And in each of the 35 parameters,
you need to change the function (Attribute) from 1 to 0

 

i got that thanks.

Do you really have to keep the CPU running at maximum frequency all the time for the GPU to reach 99% in games?

 

I do not claim that everyone who has win10 games will always go bad, and do not claim that Win7 will help you, but that's what helped me. And you need to try to change the variables to find a solution to the problem

 

if everything goes well then they should appear here. This is Part 2

 

There is a slightly easier way to get those extra power management settings as well by using Bitsum Park Control:

 

Can't I just apply those registry settings in windows 10? Did you try it?

 

50%-80% - this is the reason for the fall of FPS compared with others in which 99-100% of the use