Precision 7530 & Precision 7730 owner's thread

I doubt GPU-Z will be able to extract the BIOS on these cards. It failed on my Thinkpad's P600.
Reason being that BIOS on mobile Quadro is not stored the same than their desktop counterparts (don't remember the specifics).

Under linux, you could try this tool using a downloadable BIOS update for the 7730:
https://github.com/coderobe/VBiosFinder

Note: if you want to extract the VBIOS for doing GPU passthrough on Linux for running Windows, this does not work:
nobody has been able to do it with Pascal and Turing mobile GPU.

 

GPU-Z worked on my P3200; nvflash works, too. The VBIOS in the Dell Quadros are stored in a separate EEPROM chip, same as desktop cards.

 

Ah, good to know. I thought all mobile Quadro were equal in how the vBIOS is stored.

 

I did a dump on my 5200. This function works in GPU-Z. I need somebody owns the 5200 made a dump of your video card. I really need it for that would revive the video card.

 

And let Dell off the hook for a half-baked product? So many annoyed users doing rain-dances at https://www.dell.com/community/Prec...Dock-does-not-work-as-advertised/td-p/6245138.

I think at this point I'm going to fight for my money back.

 

I observe precisely the same behaviour that you did: after closing Lightroom, the GPU appears to go into idle. However, my power draw remains high (~ 20 W), which is is a telltale sign that the GPU is not sleeping.

I am really quite lost. I think I might demand a motherboard + GPU replacement.

 

How do you measure the power draw of the graphics card?
I also have the combination of i9-8950HK and P3200.

In Idle the P3200 is not active and so I can't see the power draw of the card in HWiNFO. Or have you turned off the graphic switching?

 

I have left graphics switching ON—this entire discussion is about high power draw even with Optimus (or graphics switching) left on.

If your P3200 isn't displayed in HWiNFO, you'll need to enable the option below:


And then restart HWiNFO. Don't be misled by the 0.000 W: if your unplugged idle power draw is on the order of 15-25 W, then your NVIDIA GPU is almost certainly still running.

 

Hm, I'm seing some Schrödinger's cat situation. You try to verify Optimus not working because your P3200 is always active. But to do so, you need to wake (activate) the P3200 to measure it's power draw. This way you most likely bring in some additional unkown factor into your observations.

Already checked what happens if you disable that point and how long your battery lasts without even using HWInfo ? Maybe HWInfo is one culprit...

 

Ah, that's why HWiNFO has three options pertaining to the GPU: one to wake it, and the other to continuously poll it (this is the one that keeps it awake).

As mentioned earlier, the power draw charts I put up were with data extracted with HWiNFO, with that option un-checked.

Let me summarise everything until now:

1. I observed roughly equal and high power draw and hence battery life regardless of whether or not the Graphics Switching option in the BIOS was checked or not, and regardless of whether or not the Discrete Graphics Controller Direct Output Mode option was checked or not. Two options each with two toggles = 4 total combinations, all of which I tested.
2. I recently chanced upon the observation that:
3. Opening Adobe Photoshop Lightroom Classic CC v8.2
4. Importing a photo
5. Setting Lightroom to 'Develop' mode
6. and minimising it, led to massive improvements in battery life after it was minimised (a doubling, as shown earlier)
7. I proceeded to test this hypothesis with HWiNFO logging data, with that option both checked and unchecked. There was marginal difference in the power draw, and the biggest contributor was Lightroom's odd behaviour.

This led me to the conclusion that Optimus on the Precisions is poorly implemented; this is further cemented by the fact that @SvenC owns a Precision 7730 with the 8950HK, that is supposedly a power-hungry machine, and his notebook reaches ~9.5 hours of battery life with light use.