Precision 7530 & Precision 7730 owner's thread

Precision 7560 and 7760 were announced on Tuesday. You can see some details in the 7560/7760 pre-release thread. Availability date is still unknown but it should be "soon". (I would guess that you'll be able to order one within eight weeks... hopefully a bit sooner than that.)

 

Oh, excellent. I missed that and didn't see anything when I checked the Dell website this morning.

 

I've been getting a lot of BSODs when my monitor is plugged; pastebin. Can anyone make head or tail of this? The key stack trace is reproduced below, also:

Code:

nt!KeBugCheckEx
nt!MiRaisedIrqlFault+0x184142
nt!MmAccessFault+0x4ef
nt!KiPageFault+0x35e
nt!MiWsleFree+0x17e
nt!MiFreeWsleList+0x1bb
nt!MiTrimWorkingSetBuildup+0x87
nt!MiTrimWorkingSetTail+0x28
nt!MiWalkPageTablesRecursively+0xc6d
nt!MiWalkPageTablesRecursively+0x88b
nt!MiWalkPageTablesRecursively+0x88b
nt!MiWalkPageTablesRecursively+0x88b
nt!MiWalkPageTables+0x36b
nt!MiTrimWorkingSet+0x1db
nt!MiReduceWs+0x71
nt!MiPreUnlockWorkingSetShared+0xb5
nt!MiUnlockWorkingSetShared+0xdd
nt!MiUnlockWorkingSetExclusive+0x117
nt!MmAdjustWorkingSetSizeEx+0x228
nt!PspSetQuotaLimits+0x2e8
nt!NtSetInformationProcess+0x7e4
nt!KiSystemServiceCopyEnd+0x25
0x00007ffe`4584d1a4

The error was ATTEMPTED_WRITE_TO_READONLY_MEMORY.

 

No, that all looks like internal Microsoft stuff on the stack, there's no hint which driver could be causing it ... (Also assuming that a .SYS file wasn't blamed right on the actual BSOD screen?)

Which GPU is the external display attached to? I would try to attach it to the other GPU and see if that makes a difference, as a first troubleshooting step. (I.e. If the monitor is connected to the rear HDMI or mDP port, connect it via USB-C instead so that it can be connected to the Intel GPU ... assuming that Optimus is enabled.) If it does make a difference then the blame possibly lies with one or the other GPU driver...

 

The Quadro. I've been getting approximately two BSODs every month, and they're only when I'm connected to the monitor... I get a feeling it has to do with the NVIDIA driver, as it's quite famous for causing BSODs. I don't get kernel panics or crashes in Linux, so it must be a Windows problem, because I initially thought the memory was problematic.

All in all it feels like the 7530 was a bit of a beta release, because non-JEDEC memory doesn't work well, the docks don't work well, and there are weird boot-up issues such as no iGPU output through the HDMI/DP ports at all...

 

Is XTU undervolt possible again within the latest bios version 1.15.3 ?

 

I installed it; undervolts still appear to stick with ThrottleStop. A bigger problem is that I now have an extra error message every boot:

Code:

(A7) Me FW Downgrade - Request MeSpiLock Failed

I might just source out a 7540 motherboard with an i9-9880H and change my RAM out to non-ECC modules to set everything to rest, because this is annoying.

 

So you mean undervoltage is possible or not ?

 

To be fair, you are using an "unsupported" GPU with an "unsupported" power limit so who knows if that could be part of the issue...
(I ended up having to get rid of the Quadro P5000 in my Precision M6700 because it caused BSOD, only once a week or so ... and they only occurred when an external display was connected. Problem went away when I put my M5000M back in, though that is also an "unsupported" GPU for this system.)

Pretty sure that this is the case with all future models as well and it's because these ports are directly on the dGPU. They never implemented pass-through from the iGPU (unless you have the dedicated pass-through card).

 

Yes, I can still undervolt.

It very possibly is part of the issue, not to mention the self-signed drivers (which also causes problems with multiplayer game DRM/anti-cheat). Like I said above, I am in half a mind to buy myself a new motherboard, which might end up being cheaper than buying additional warranty...

A bit of a bummer, if even new models don't have all three ports from the iGPU/dGPU. It makes no real sense, as it should be straightforward to implement: simply mux the outputs, and send a signal to choose the input device. Maybe the DGFF is truly closer to the desktop PCIe form factor than one might expect it to be...