Upgrading my Ranger

Well, before I tore down the laptop earlier, I did a CB run and it DID thermal shutdown. I'll try again, but won't do much troubleshooting after as I'm quite tired, hence why I'm probably grumpy I'll edit this when I get my results.

EDIT: Got a thermal shutdown, dunno what is going on, hopefully the CPU isn’t screwed and it’s just the LM being weird. I’ll see what everyone has to say tommorow morning.

EDIT 2: Took all of the LM off of the CPU this morning, and then re-applied it, and the same ****ing issue. I'm just stuck, I guess its back to the Asus laptop for now until I figure this out.

Edit 3: Now that I think about it, I wonder if the iGPU is causing this issue. The package temp is of the entire CPU, and the package temp and the iGPU temps are always the same. So I'm now thinking that the cores are fine, but the iGPU is hot, which causes the package temp to be hot, which is what the motherboard reads which leads to a thermal shutdown

 

The iGPU shouldn't even be active in your setup so that's very strange. Running the laptop in dGPU only mode disables the iGPU.

Did you make sure the LM is distributed evenly across the CPU die and the heatsink?

 

Yeah, I did the best I could to make sure it was spread evenly and didn't look like pooled up water, here are some pics I got of my last re-application. Sorry about the flash refelcation, was in a room with sub-par lighting when I took the pictures.

 

That looks like an excellent LM application to me, so we can rule that out.

Another thing to check is if there is anything around the CPU that can cause a short. I don't see anything though.

The only thing left I can think of is if you shorted the contacts around the CPU die. I need someone else to chime in with more ideas.

 

Did you try and compare with stock heatsink?
For me, the stock one is better
With lm, heatsink contact is everything, thats my experience. I guess one would not need any special protections with the right contact and just a small amount of lm.
That paste around cpu, you sure it doesnt go inside below cpu?
That tape looks thick, i used varnish, foam barrier and some kapton tape around cpu edges

 

The thing is with my stock heatsink is that it is physically bent. I recently looked at it and noticed that the left portion of it is slightly bent upwards, so its basically useless and I would probably risk more damage to my components.

The paste was used as a barrier as I could not cut the foam I bought thin enough to be able to be used as a barrier. I'm thinking about just removing the leftover paste entirely to further troubleshoot.

The tape I used was some Super 33+ recommended by Falkentyne, so the tape shouldn't be causing issues.

The thing about the heatsink contact is that everything was fine on Friday and Saturday morning, but all of the sudden temps just started skyrocketing, and my applications of LM seem fine. The only idea I would have to suggest that the heatsink is at fault is that I originally put too much LM, but because the contact was bad, the LM substituted, but then later moved around causing weird temperature issues on the iGPU. But even that idea doesn't fully explain why my core temps are still normal and have not changed. I have my 4710mq that I may use as a last resort to see if the heatsink or 4930mx is at fault, but I'd like to try a few other things before that.

 

Just re-seated the CPU, blew over the pins and CPU socket holes with compressed air, re-applied my electrical tape, and re-applied LM, and its the same ****ing issue, so I'm just going to assume the CPU is broken or something. Is there any way to disable any sensors? What's annoying is that it seems the motherboard reads the damn Package temperature, and the iGPU and Package temperature go hand and hand, so if the iGPU temp is hotter than the cores, then the Package temp is going to be whatever the hottest sensor is. If I can manage to disable the iGPU sensor, then I can probably solve this issue as I'm 99% sure I have good contact, but I don't really know, I'd rather not have to buy another 4930mx.

 

Do you a favour. Put in the 4710mq for excessive testing. First with good thermal paste, then test with Liquid metal.

What bench software you test with? The iGPU shouldn't even be active in your setup. Then fire up Heaven who don't tax the Cpu and should put the dGPU in work. Then fire up CPU bench as CBR15 at same time. Maybe try first with lower clock speed on the Cpu before you push clock speed. Do the same testing both for the 4930mx and the 4710mq.

 

I'll do that in a bit, I'd like to keep troubleshooting this thing. I just need to go in with a calm mind as I tend to get too entangled in this kind of stuff, which makes me quite grumpy.

I usually bench with CBR15. Well, I don't really know if it is the iGPU, but HWINFO lists what is getting hot as the GT Cores, which from what I read, are the graphics cores. Will do your suggestion tomorrow when I have a bit of time.

EDIT: I had just gotten home from the weekend earlier, so when I was playing a few games, my CPU package was around 85-90C. I had to leave and do some things, and I decided to leave my door closed so my room would heat up as it usually does. I just got back, and am playing the same games, and despite the core temps being around 6-7C hotter, the package temperature is still in the 85-90C range. I think this basically proves that there is something wrong with the temperature sensor as if it were working normally, it would have gone up a few degrees, and it didn't.

 

Could be.
You bought an used 4930 that is hot by definition. Now look at all that you have done to make it cooler.
Very few people do these things as they have no idea about it. It could have burned heavy in the previous life and now it just kind of works.
I spent money on one of those that did not work at all, burned beyond repair probably and due to me not returning it in time i lost money and i dont think ill have another chance to get one.
Mind u, at least i have a lovely 3d domed 4930mx sticker

 

Alright, just got the 4710mq in with some NT-H1, and everything works fine. This either means that there is something wrong with the LM, or there's something wrong with the 4930mx, and I believe its the latter . I'll apply LM to the 4710mq tommorow as I will be busy and don't have too much time to crack this thing open again. I also apologize for posting in the wrong thread lol, I could of swore I had this one pulled up when I shut off the computer, but I guess I didn't.

 

Installed the 4930mx with some paste like an hour ago and booted into windows fine, but the clock speeds were stuck at 800mhz. I shut down to get into the BIOS and it took me a minute, but I got in and noticed everything was incredibly slow. When I went to the overclocking performance menu, the BIOS froze, so I shut off the computer and restarted. I then waited literally 15 minutes because the laptop was being weird, but the BIOS finally loaded, and as soon as I moved to another menu it froze (picture attached is what it looked like.) I then restarted again and I couldn’t even get into the BIOS as the laptop would just turn the backlight on and do nothing. I’m currently resetting the BIOS by taking out the CMOS battery to see if that’s the issue, but I just find it odd that everything was fine when first booting up, and then everything started degrading after each restart.

EDIT:
Everything works now after resetting the BIOS, the only thing that changed is the weird artifact-esk thing that pops up right when I turn on the laptop, it now flashes on the right side of the display instead of all of the display.

The GT Cores sensor is damaged though , so I guess its time to get another 4930MX unless someone can tell me how to disable the GT cores sensor.

 

Hello there I'm new to this thread but been reading it All the way through . It's odd your having issues like this ( indicates a bad CPU ) I have a 4930mx and I'm able to hit 4.1ghz at 1.190 volts the highest I'll hit is 87 c at full tilt . I was considering. LM but the thermal take stuff I'm using seems fine I even have the 2 pipe heatsink . C0 stepping cpu though

 

I'm debating about going 2133 ram though . Since the bios allows up to 2666 lol not even sure if you can get ddr3l so dimms at that speed . One question I do have though is what is the part number for the OEM 3 pipe heatsink ?

 

If you are talking about a 3 pipe CPU heatsink, I'm not 100% sure if those ever shipped with the laptop, but I know Cicichen, the guy who sells the 4 pipe CPU heatsinks, also made a 3 pipe heatsink.

As for the 3 pipe GPU heatsink, the part number is N3P1N.

 

also another thing i have a quadro m4000m in my alienware 17 r1 i need a modded driver to get it to work its a dell card and the bios recognizes it if i post the hardware ids could someone make the proper inf or mod a driver for me ?

 

If you post the hardware ID then I'll help you with the mod. It's not too hard to learn and then when you do, you can keep it updated all the time

 

Problem is my eyes are bad do staring at a bright white screen full of texts pretty much kills them I'll post the hardware ids I appreciate it

 

I'm at work and I don't have admin permissions to install software on the work PC so I can't extract the .inf from the newest driver. Best I can do is write the line and the string that you'll need to insert into the .inf.
I'll walk you through it

 

this is my hardware ids also it says the driver trying to start is not the same as the posted device

 

this is also popping up i have removed all nvidia drivers already though lol

 

Ok here are the lines you need to put into the nvdmi .inf

%NVIDIA_DEV.13F9.05AA.1028% = Section049, PCI\VEN_10DE&DEV_13F9&SUBSYS_05AA1028

NVIDIA_DEV.13F9.05AA.1028 = "NVIDIA Quadro M4000M"

- Use DDU in Safe Mode to remove any traces of an old driver
- Download and attempt to install the latest driver
- When it fails, keep the installer open and go to your C: Drive. Find the NVIDIA folder and copy and paste this in the same C: drive. It will auto-name itself NVIDIA - Copy
- Close the installer and the original NVIDIA folder should disappear
- Go into the NVIDIA - Copy folder and get into the International folder and then the Display.Driver folder
- Scroll down to the nvdmi.inf and open it
- Scroll down to the list of device ID's and make a space for the first line that I put above. I like to put it in order (below the last 13F9 in the list) but it doesn't really matter
- Fast scroll to the bottom of the .inf and find the list of Strings (that look like the second line above) insert the M4000M string (again I like to put it in order beneath the last 13F9 String)
- Save the inf, close it and close the NVIDIA - Copy folder
Driver is now modded and ready to install.
Restart the laptop with Disable Driver Signature Enforcement
Hold SHIFT while going to the Start button and pressing Restart
Troubleshoot - Advanced Options - Startup Settings - Click the restart tab on screen
When it restarts press Option 7
The laptop will restart and you can go straight back to the International folder and run the Setup.exe at the bottom of the folder.
During the installation you will be greeted with a warning message. Just accept it and the driver will install.

 

okay so i did all that but any time i do this then go to the setup icon and double click it now says i lack permission . driver signature is off used ddu

 

justr isnt working i even dfirectly modifed the installer ie extracted it im about to tear this thing apart

 

further more even after removing all nvidia driver info it stil shows that same error

 

If you followed the guide then it should work so my guess is that you’ve done something slightly wrong. Did you disable driver signature enforcement?


Sent from my iPhone using Tapatalk

 

Yes I did. I rebooted shd hit 7 when prompted with the advanced start-up

 

You you could modify an inf for me I could then paste it into the display driver folder ruling out user error

 

There is an application that handles that for you now from what I have seen. Cant remember the name of it but its here somewhere in the forum

 

I've looked all up and down can't find it plus with my bad eyes I could be missing it

 

nothing on the internet about an auto application either LOL

 

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

 

not sure what that accomplished lol it just launches the nvidia installer which then fails to detect compatible hardware . im going to try redownloading another driver modifying the inf again and use this thing

 

Its probably best for you to make your own thread

 

yeah that app doesnt work . idk what is up here this doesnt make sense at all as soon as i edit the nvdmi.inf filed the setup is broken ill upload the inf and one of yall can see what i borked lol

 

yeah but you know how many dead questions are on here ? please help im literally begging im visually disabled so this hurts my eyes

 

and now for some reason it wont let me delete the files . sighs

 

after modding the inf it now says i do not have permission to delete it or anything ..... sighs . well looks like i wasted 170 bucks

 

i might jst pull the 880m out of my clevo and stick it in the Alienware instead should just work do to the fact the card actually ships with this model lol ill sell the clevo for an m6800

 

5 posts in a row? This is not a messenger, its a forum with its own set of rules. Use the edit function, you are actively robbing exposure from other people just like you who are requesting aid. If you cant respect the forums rules it makes people less inclined to help you.

https://null-src.com/posts/nvidia-notebook-driver-inf-mod/post.php

Or you can look at LaptopVideo2Go and/or https://premamod.wordpress.com/2017/08/15/j95-nvidia-inf/

 

Sorry about that it's because the visual assistance app I have is kinda buggy abd instead of putting the posts into one like it sends them as sperate replies

 

Can you list the process you are going through to mod and install the driver? Something doesn't seem right.

1. Run DDU in Safe Mode to Uninstall the old driver files
2. Download and Install the latest Nvidia Driver from the Nvidia website for your GPU
3. Keep the failed Installer open and Copy/Paste the Nvidia Folder in your C: Drive
4. Close the Installer
5. Open the Copied Nvidia folder and insert the Device ID line and String in the correct places in the nvdmi.inf
6. Save the .inf and exit.
7. Restart the laptop with Disable Driver Signature Enforcement
8. Run the Setup.exe in the International Folder of the copied Nvidia folder
9. Done

 

yeah did all that but as soon as i save the INF the driver breaks and makes it near impossible to delete the folders lol i found a way around that by booting linux . i really appreciate your help . ive done this before on an old m6400 with quadro 5000m .that was a few year back using the same methods just a bit different . as the drivers were not as picky im starting to think it was windows preventing this so what ill do is try an older build and do this again again thank you

 

also i see your parting out that 17 r1 i just got another with no gpu LOL what would you ask for the 880m

 

Saving the inf shouldn’t do anything and I don’t know what you mean by deleting the folders.
I do this all the time with my 1070 so I’m not sure why your card would be different. Maybe over the weekend I can do the inf mod for you.


Sent from my iPhone using Tapatalk

 

@Tenoroon knows how to contact me if needed.

Im not going to participate in thread hijacking nor do I want to see it either anymore. Unsubbed.

 

Got it working. With the help of another in a different thread . Turned out my ranger ignored driver signature disabled even though it was set to disable so I had to out it in test mode thank you for all your help

 

Long story, need to sleep for school tommorow, but IGPU is ****ed on the 4930, i put it in an optimus based laptop and got this, same issue Zero1 had:

 

Alright, I hate doing this, but this is now stuck in my brain, and I am frustrated that I mostly likely can't fix it.

I'll re-tell the story, but I'm going to need some input from the brains of the forums:
@unclewebb @Mr. Fox @Papusan @Reciever and any other people who have a lot more experience than me.

So over the summer, I got some LM for 4930mx so I could overclock it a bit. I applied it very carefully, and everything turned out fine, temps dropped and I was running at a stable 4.0ghz across all four cores. About a day later, when I was downloading a game, I noticed that my Package temperatures were in the high 90's and almost 100, I though that maybe the LM got absorbed into the copper heatsink, so I re-applied the LM, and still same issue. I noticed that while the Package temperatures were hot, that the core temperatures were perfectly normal. I immediately thought the LM messed the CPU up, but I did a really good job making sure the pins on the substrate next to the die were covered (Haswell CPU's have that row of pins or whatever they are above the die.)

I asked what was going on, and @Clamibot was just as confused as I was because I'm using the 120hz eDP panel on the laptop, the iGPU wasn't being used at all, just my dedicated 1070 (Keep this in mind for later.) I finally decided to put in my 4710mq to see if it was maybe a motherboard issue, and everything was normal. I then put the 4930mx back in with regular thermal paste, and the same issue happened, normal core temps but really high iGPU temps. Since then, the 4930mx has been sitting on my desk, and I've been quite angry as I couldn't figure out the issue.

Fast forward to yesterday, and I forgot that I had an older MSI laptop with a 4210m laying around. So, last night, I put in the 4930mx with normal thermal paste, and lots of artifacts happened as seen in the post above this. It was quite late, and I wasn't able to think very well, but I realized that the 4930mx was utilizing the iGPU at the time as the MSI laptop uses Optimus. So when I got home from school today, I immediately booted into safe mode, and the artifacting stopped, but the iGPU temperatures were insanely high still. I then disabled the iGPU drivers and booted normally, which also gave me no artifacts, but as soon as I re-enable the iGPU drivers, mass artifacting starts happening, and the iGPU temperatures get even higher. I will include some pictures at the bottom of this post with explanations of what I was doing at the time.

Now yes, this issue could be the LM, but its strange that @Zero 1 had the same issue (lots of artifacting with his 4930mx,) without using LM. The issue could very well be the iGPU just dying, but I find it odd that it took temperature sensor with it. You could also argue that the temperature sensor for the iGPU isn't broken, and it really is that hot, but trust me, I used my finger and a thermal gun, the thermal gun never said anything close to the reported iGPU temperatures and the heatsink was never hot to the point of pain.

What makes me so frustrated is that this issue could easily be fixed if I could completely disable/kill the iGPU. I literally have 0 use for it as my Ranger uses the dedicated GPU only and the broken iGPU screws with the Package temperature, which is what all motherboards refer to when deciding to thermal throttle/thermal shutdown. My main question is, would it be possible to disable/kill the iGPU, or at least screw with the sensor to give it false readings so that the Package temperature doesn't think the iGPU is really hot?

Pictures: (I know they kind of suck, but I'll try to explain what was going on.)

In this picture, I was running Unigine Heaven with the dedicated 920M this laptop has, the iGPU drivers had been disabled and the CPU core frequency was being throttled to around 2.3-2.4 ghz.
The main core temps are a bit higher as Unigine Heaven lightly taxes the CPU.



In this image, nothing was running, but the iGPU drivers had been re-enabled within Windows. You can see the artifacting and the insanely high idle temps of the iGPU.


I appreciate anyone's input/help with this, it means a lot, and I'm hoping I can revive this 4930 as finding them in the U.S for a decent price is difficult.