XPS 9550 Bios 1.14 - any issues? undervolting?

Hi,

I have an XPS 15 9550, 64GB of RAM (yes it works).

I'm running old BIOS 1.9.0 Dated 10/11/2018. I'm undervolted and running under Debian and all is working really nicely. It's rock-solid stable. So, normally, I would not mess with updating BIOS on something that's working.

However, I just got a couple of Dell docks WD-19TB and TB16 and I can't get the USB-A ports to work on either of them. I updated the Dock BIOS and the only thing that seems to be left is the PC BIOS.

The latest BIOS is 1.14 October 2020. Before trying this, I wanted to see if there are any issues? I want to keep undervolting and keep being rock-solid stable.

Any help much appreciated.

Thanks!
Mark

 

I run 1.13.1. Undervolt works perfectly with Throttlestop and I don't have any issues. I run an i5 on Win 10 Pro.

You will see that 1.14 addresses INTEL-SA-00289 so may restrict undervolting, which is pretty well documented at NBR. There are some workarounds (which I thought was rolling back to an older BIOS version but not fully up to date there.).

BTW are you running the i7? Can you run 64GB of ram in Windows? Exactly what RAM part number are you running? Any tricks? I think I saw a couple of people running 64GB on Skylake i7 laptops in the past few years. I run 32GB but could use 64GB.

 

Thanks much! I will try 1.13.1

I am running an i7 with 64GB of RAM. I have not tried Windows with it but there was another thread here where somebody had. I bought Crucial CT2K32G4SFD8266. The trick: I did have to adjust the undervolt as I was getting hangs. Under Linux, the Analog I/O worked at -120mv with 16GB, but I had to set it to -40mv with 64GB. The rest of my undervolt setup did not have to change. CPU -175, GPU -120 (although I have the GPU disabled anyway), CPU Cache -175

 

Thanks Mark. Check out the ThrottleStop guide for recent undervolting experience with Skylake. It appears we do not need the same undervolt for CPU core & cache. I run -180 and -160, respectively rock solid on the i5 which I think is more ammenable to aggressive undervolting than the i7 but that varies by individual chip so you just need to stress test.

Since you are running such aggressive undervolts, maybe that is causing the USB port instability. Try without the undervolts.

Also, I found undervoting the iGPU can cause some latency issues, so I get better performance on my applications with on iGPU undervolts.

 

Thanks much. I'll try the BIOS and if that doesn't work, I'll try without the undervolts. I did the repaste and run i8kmon to control the fan temps. I run a couple of VMs and a couple of browsers and without pegging the CPU, the fan is silent with CPU < 60*C. That's probably the other thing I should have asked about -- any issues with overriding factory fan control with the newer BIOS?

 

You can still have undervolt wonkiness even at low temps and low CPU use.

Dell Command | Power Manager give you access to fan registers that are not accessible elsewhere. v2.2.1 is the latest that works with the 9550 as far as I could tell as the newer versions were for newer laptops; I haven't checked in a year. I don't have any of the other Dell spamware installed.

 

So, I got this resolved. It wasn't the undervolt or the bios. There is a separate chipset firmware for the XPS 15 Thunderbolt. I had version 12 and updated to version 16. This took some learning under Linux.

So there's firmware on the Thunderbolt dock, Thunderbolt firmware in the XPS and BIOS in the XPS. Wow.