Dell Precision 7670 & Dell Precision 7770 owner's thread

[Aaron44126]

Realtek audio driver

https://www.dell.com/support/home/en-us/drivers/driversdetails?driverid=PPC6R

 

Intel graphics driver

https://www.dell.com/support/home/en-us/drivers/driversdetails?driverid=72D84

 

(Both just made public today, despite what the "release date" says)

 

——————————

 

Over one week in and no hint of movement on my replacement cooler.  Their estimated ship date is March 24, so another week to go.  I've been thinking about it and I think it's possibly a paste pump-out issue.  When I took the cooler off to repaste it seemed like an extremely barely there layer of paste was left between the CPU and the cooler.  @win32asmguy mentioned something similar with the stock (Delta) cooler after repasting in the early days of this thread.  I've been playing a more CPU-intensive game lately, so.....

 

Anyway, if that is the case, hopefully the factory paste on the new cooler holds up better (I think it should).

[MyPC8MyBrain]

1 hour ago, Aaron44126 said:

When I took the cooler off to repaste it

The way I do repasting is similar to what you described up until the cleaning portion.

From there, I use electrical tape all around the CPU and dGPU silicon die as close as I can. I then create a flat layer of paste that covers the entire die evenly. After that, I remove the tape, which leaves a perfect flat two-millimeter layer of paste perfectly on the die.

The next step is critical, which is mounting back. Do not just tighten it down based on number order. Place the entire mount gently in its place, using its screws as guides. I like to go over each screw and turn it counterclockwise until I hear it pop on top of its screw slot. I do that for all screws to ensure that they all start in the same position. Then, go by the numbers, tightening them back, but only do a 1/4 turn at a time, going around until they are all tight. This ensures that the plates are not deflecting the paste to one side or the other as it's being seated back.

[Aaron44126]

10 hours ago, MyPC8MyBrain said:

The next step is critical, which is mounting back.

 

Yeah, I follow a similar procedure turning each screw just a little bit until they all bottom out.

 

—————————————————

 

Had something else weird happen.  I've been having some more unexplained crashes in recent days.  Just this morning, the system went and just locked up.  After forcing a reboot, the Dell BIOS really wanted to run a hardware diagnostics, even though I have the SupportAssist recovery options turned off in BIOS setup.

 

So I dug in today and saw that the last thing showing up in Event Viewer before the crash is a PCIe device error:

 

I checked out that device name PCI\VEN_8086&DEV_7ABA, and it turns out that it corresponds to the PCIe slot hosting the SD card reader.

 

I've been running with an SD card inserted in order to facilitate moving files between different installed OS's (Windows 10, Windows 11, Linux) — and actually I rarely boot into one of the others so I don't use this often, but I've just left the SD card inserted so that I don't have to go looking for it in the case that I need it.  I need something like this since I use BitLocker and I have it purposefully set up such that the OS's can't see each other's partitions.  (I guess I could just make a tiny unencrypted shared partition.)  The SD card in question is just 8GB and it's probably like 15 years old, so maybe it is flaking out?  Can a misbehaving SD card take down the whole system .....?

 

(Quick edit — Looking at some online stories, looks like the answer regarding whether a bad SD card can cause system crashes is "definitely yes".  It would probably be less likely if the SD card reader was attached via USB and not direct PCIe.)

 

Anyway, I just pulled the card out and I will run without and see if that takes care of the problem.

[MyPC8MyBrain]

i think you're chasing chasing the wrong lead.
if the SD card itself was bad you'd get read or write errors not PCIE errors.

it seems to me that everything started after you created a stub for NVidia device.
something happened that made you do that first, then two weeks later you notice the system is over heating.
and now this PCIE issue.

 

[Aaron44126]

Man, you really don’t like that stub. It’s literally doing nothing, by design, except from prevent the NVIDIA HD audio driver from being loaded — as if the driver wasn’t even installed (except I don’t have a Device Manager error).

 

In this case, the first crash was February 27, a few days before I implemented the stub. Of my hypothesis about the bad card is wrong, we’ll find out.  In any case, I have been looking and see numerous cases reported of a bad card causing BSOD or other system issues. It doesn’t always result in “clean” read/write errors.

[MyPC8MyBrain]

43 minutes ago, Aaron44126 said:

you really don’t like that stub

I have nothing against your stub solution, as you said its just HDMI sound driver.
I cant recall what brought up the stub solution, I just remember something was acting up for you prior.
 

the point i was trying to make is its clear that something is slowly creeping up on you there.
i was just trying to draw your attention to possibly a pattern you maybe overlooked that could have started way back.

 

[Aaron44126]

8 minutes ago, MyPC8MyBrain said:

I cant recall what brought up the stub solution, I just remember something was acting up for you prior.

 

That one was pretty clear-cut (IMO).  I "caught" the NVIDIA audio driver gobbling CPU time on the "System" process and causing stutters in videos/games.  The problem went away immediately when I disabled it.  That's why I went to stub it out.  (It is useless to me, I don't have any situation where the NVIDIA audio driver would actually be needed.)

 

I will of course keep in mind that there may be another (deeper?) issue.  Overall, I wish this system was a bit more solid, feels like I am always fighting with something.  😕

[MyPC8MyBrain]

did you ever do a clean install or you running base OEM image?

if you are with OEM image your problem could be rooted there.

[Aaron44126]

38 minutes ago, MyPC8MyBrain said:

did you ever do a clean install or you running base OEM image?

if you are with OEM image your problem could be rooted there.

 

I did the Windows 10 install myself.  Very little Dell software installed.  I do still have the stock OEM Windows 11 install available to boot from.  Thermal/fan issues seem to be present in both, making me think a hardware issue rather than a software one.

[heikkuri]

On 3/17/2023 at 4:17 AM, Aaron44126 said:

Realtek audio driver

https://www.dell.com/support/home/en-us/drivers/driversdetails?driverid=PPC6R

 

Intel graphics driver

https://www.dell.com/support/home/en-us/drivers/driversdetails?driverid=72D84

 

(Both just made public today, despite what the "release date" says)

 

 

 

 

New Realtek audio driver 6.0.9472.1, A46 really fix the problem, which was with the previous driver 6.0.9418.1, A19.

Great jump from A19 to A46.

 

Now the folder C:\Program Files\Waves has just one subfolder IntelOpenVINO. No more constant pending reboot state

of bad audio driver.

 

Fixes & Enhancements does not tell you anything about the fixed Vino problem or pending reboot case.

 - Fixed the issue where there is no audio from the external USB speaker.

 

Dell Precision 7560 and 7760 are still missing the same audio driver fix.

 

Intel Management Engine Components Installer has also a new driver 2302.4.5.0, A05.

Do you install the driver in home environment?

 

 

 

[Aaron44126]

1 hour ago, heikkuri said:

New Realtek audio driver 6.0.9472.1, A46 really fix the problem, which was with the previous driver 6.0.9418.1, A19.

Great jump from A19 to A46.

...

Intel Management Engine Components Installer has also a new driver 2302.4.5.0, A05.

Do you install the driver in home environment?

 

I also had a problem with the 9418 version and I am still running the previous one, 9363.

I have not installed any of these new driver versions just yet, I am in "change freeze" while waiting for my replacement heatsink assembly.

 

I do install the Intel Management Engine drivers.  I have seen recommendations to disable the Intel Dynamic Tuning service but I have not done this myself.

[Aaron44126]

BIOS update 1.10.0.

https://www.dell.com/support/home/en-us/drivers/driversdetails?driverid=C9NWC

 

Note, if Dell included an Intel CPU microcode update, undervolting may be non-functional.

 

- Fixed the issue where the system cannot detect the Dell DA310 USB-C Multiport adapter after updating the BIOS. This issue occurs after connecting the adapter to the system.
- Fixed the issue where an error message is displayed when you connect the 45 W and 65 W adapters to the system.
- Fixed the issue where the system charges beyond the Custom Charge Stop set in the Battery Configuration in the BIOS setup.
- Fixed the issue where an error message is displayed when you connect the 45 W and 65 W adapters to the system.
- Fixed the issue where an application does not close when you press the Alt+F4 hotkey.
- Fixed the issue where the system cannot turn on after entering storage mode and plugging in the 130 W adapter.
- Fixed the issue where the system stops responding when the system boots to the operating system.
- Fixed the issue where the system reboots repeatedly when USB bootable devices with longer product name and serial number are connected to the system.
- Fixed the issue where system cannot connect to the BIOSConnect server with all the supported WLAN cards.
- Fixed the issue where the system cannot detect the SmartCard in the Preboot Execution Environment (PXE).

[Aaron44126]

Not sure when they started offering this but Dell has (expensive) CAMM modules available to order standalone from their web site.

16GB - https://www.dell.com/en-us/shop/dell-camm-memory-upgrade-16gb-4800mhz/apd/370-ahfq/memory

32GB - https://www.dell.com/en-us/shop/dell-camm-memory-upgrade-32gb-4800mhz/apd/370-ahft/memory

64GB - (not there for some reason)

128GB - https://www.dell.com/en-us/shop/dell-camm-memory-upgrade-128gb-3600mhz/apd/370-ahfr/memory

[Aaron44126]

...I'm sorting of doubting that they're going to hit their ETA?  😕  Next-day service is the best.

[brunooo84]

has anyone updated to latest bios? any issues? thanks!

[RYANKO]

On 3/25/2023 at 11:18 AM, brunooo84 said:

has anyone updated to latest bios? any issues? thanks!

I have updated the latest bios three days ago.
Use the laptop 14hrs per day.
Nothing special happened so far.
I didn't do undervolt by the way.

[Aaron44126]

On 3/24/2023 at 8:23 PM, Aaron44126 said:

...I'm sorting of doubting that they're going to hit their ETA?  😕  Next-day service is the best.

 

Called Dell this morning to get an update on the heatsink.  (Wanted to wait until regular business hours, figuring that even though they have 24/7 support people on hand, the people who can dig into parts inventory are not generally working on the weekend.)  I was told that on March 21, the ETA on their end was bumped back to April 7.  This came after an earlier bump back to April 1.  I can't see those update on the ticket; for me it is still showing March 24.

 

He suggested that I give it another week to make sure that they haven't pushed it back anymore, and that if it keeps getting bumped back then I should put in for a system replacement.

 

So, more waiting, and in the meantime...  I've ordered a heatsink from a third-party supplier, and it should arrive next week; we'll see which one gets here first.  (I'm hoping to avoid going in for a system replacement yet.)

[Aaron44126]

New heatsink from Dell arrived.  (Still have another one on the way; I'll keep it as a spare.)

 

Here it is, as it came:

 

 

Observations:

• My original heatsink had Delta fans and this one has Sunon fans.  This is the same swap that @win32asmguy got way back in August or September; I saw that @MyPC8MyBrain got the opposite swap back in December with one of his systems (system came with Sunon fans and Dell sent a Delta fan replacement).
• Compared to the one that @win32asmguy got (pictured below in the spoiler block), mine has a much yuckier job with the GPU thermal paste.  It looks like they might have added it afterwards rather than used a factory "stamp".  It was "goopy" and some of it got stuck to the plastic packaging material (as you can see here), but it still seemed to be covering the copper surface well enough, so I put it in without attempting to clean and repaste.  (I'm more concerned with the CPU paste application than the GPU; it's really hard to overheat the GPU in this thing.)

Spoiler

Replacement heatsink sent to @win32asmguy did have a proper factory stamp on the GPU side.

• They didn't do a great job with some of the thermal pad positions and I fixed those up before putting it in.  (Also, different colored thermal pads compared to the one that @win32asmguy received.)

Before:

 

 

My system has a fair amount of apps open but not much going on.  CPU load ≤4% (mostly on P cores), turbo boost disabled.  Average CPU core temperatures in the upper 50's °C, and fan speed RPM in the 1400-1500 range.

 

After:

 

 

Fan speed consistently lower by a couple hundred RPM, temperatures lower by a few degrees, more than I got from repasting but not a major difference by any stretch.

 

I'll say that I like the fan better though.  I can't hear it powering on and it is generally quieter, and I might be making this up but it also seems to be kicking more heat out the back at low RPM.

 

I'm not going to mess with it anymore for now since the system seems to be stable and performs within my needs.  But if it gets worse I'll probably ask about a system replacement.  (I still have an outstanding issue that pops up when running Optimus disabled, too...)

[zexspart01]

Does anyone else have an issue with throtlestop ? It seems like after some recent windows update, throtlestop doesnt work anymore. 

[Aaron44126]

49 minutes ago, zexspart01 said:

Does anyone else have an issue with throtlestop ? It seems like after some recent windows update, throtlestop doesnt work anymore. 

 

Are you on BIOS version 1.10?  I don't think anyone has confirmed whether Dell has updated the Intel microcode yet.  It is expected that this will happen with an upcoming BIOS update, and when they do, undervolting will not work anymore.

[win32asmguy]

@Aaron44126 Did the new heatsink reduce or eliminate the noise the fans made when cycling on? Just curious as the 7670 sunon variant definitely does not have that issue but I did experience it with both variants on the 7770.

[Aaron44126]

12 minutes ago, win32asmguy said:

@Aaron44126 Did the new heatsink reduce or eliminate the noise the fans made when cycling on? Just curious as the 7670 sunon variant definitely does not have that issue but I did experience it with both variants on the 7770.

It absolutely did. I did not hear them cycle on at all when I first powered it up which is a big change from what I normally got from the Delta fans. I have done close testing to see if I can hear them turn on at all, since I rarely reboot and the fans are just always on rather than cycling on and off now …..

[win32asmguy]

1 hour ago, Aaron44126 said:

It absolutely did. I did not hear them cycle on at all when I first powered it up which is a big change from what I normally got from the Delta fans. I have done close testing to see if I can hear them turn on at all, since I rarely reboot and the fans are just always on rather than cycling on and off now …..

 

That is really good to know. After using the 7670 quite a bit under combined loads it is definitely not as good as the 7770 was. It can really only sustain about 85W GPU and 45W CPU, and needs MMIO PL1/2 locked via TS for it to not drop the CPU down to 30W.

 

Oh, and another interesting note about the 7780, it looks like it now has modular wifi again which is nice.

[yslalan]

On 4/2/2023 at 8:30 PM, win32asmguy said:

Oh, and another interesting note about the 7780, it looks like it now has modular wifi again which is nice.

Yea, that's also changed on 7680. They finally made it.

[Aaron44126]

Playing with Linux again.

 

Did a Linux Mint install.  Worked fine, but unworkable; high-DPI support in Cinnamon is just not there.  Everything was tiny despite being set to 200% scaling.

 

So, I did a stock Ubuntu install with GNOME 3.  I don't really like GNOME 3, but it is pretty solid when it comes to just "working well" with plenty of settings and good high-DPI support.  Observations:

• I tried Ubuntu 22.04.1 some time back and it would boot the live media but fail to boot after an install was done; Ubuntu 22.04.2 installed and booted fine.
• I was not able to select 120 Hz for the display, it was locked at 60 Hz.
• Before I could mess with that more, I did package updates (which included a kernel update) and also installed the proprietary NVIDIA driver (525).  After rebooting, the system was dead with just a black screen and blinking cursor.  (Note, graphics switching enabled so it should have been running off of the Intel GPU.)

Alas...  Don't have more time to mess with it right now.