Aaron44126

You can use them for internal, external, or network drives if you like. I had a B: SSD drive for a while, no trouble with it at all.

@brunooo84 I played with the C-states options and found that disabling dGPU C-states causes a definite hit to dGPU performance. The top C-states option seems fine to disable. (This one has existed in the Precision BIOS for a decade or more.) The bottom one regarding dGPU C-states is new and seems to actually cap the dGPU speed. Don't turn that one off.

Oh nevermind then, I was thinking 10 minutes. I've also gotten some higher single runs.

No other tuning? Man, why is 7670 faster than 7770... 😕 (15K out of box, 18K after repaste + AC/DC fix -- I know there is other stuff I could still do)

I was poking around in the BIOS to mess with C-states and I noticed another option in the performance section labeled "Intel maximum turbo boost 3.0" or something like that. It says that it will enable the CPU to boost beyond the max turbo speed, but also disable NVIDIA Dynamic Boost 2.0. It is disabled by default. Has anyone messed with that ...?

See @win32asmguy's posts on page 5 to see what steps he went through to improve performance. Then this one on page 9 where he discusses the new heatsink. Basically, it can produce similar results without having to go through the hassle of repasting / repadding / etc.

I rather doubt it. The power draw cap is so specifically right at 130W, it has to be an intentional limit. Unlike the CPU, the dGPU has no problem staying cool. (It’s designed to be used at over 2x the power level in a desktop, after all…)

HWiNFO reports a consistent 130W power draw despite NVIDIA control panel showing 150W TDP. I believe 150W would include the boost TDP, but the boost TDP doesn't seem to kick in. (This was also reported with last year's Precision 7760, 115W "actual" vs. 140W "max" TDP.)

Your image has a mismatched subsystem ID. This is pretty normal in laptops. The subsystem ID is essentially assigned by the motherboard, not by the GPU. This is why you need to mod the INF to get the NVIDIA driver to load if you, for example, take a MXM GPU card from a newer system and drop it into an older system. Its subsystem ID actually changes depending on the system that it is in. The value "burned in" to the BIOS might not match the value that is observed while the card is up and running. We can even pick out the values from the nvflash output. 10DE = NVIDIA's vendor ID 2420 = NVIDIA's identifier for GeForce RTX 3080 Ti (laptop version) 0000 = A filler value... 1028 = Dell's vendor ID 0B2A = Dell's identifier for Precision 7670. (You see these same values if you poke around NVIDIA's driver INF files.) Anyway. Comparing your image and my image, they mismatch on the subsystem ID and also the board ID. You can override the subsystem ID mismatch (-6) but not the board ID mismatch. nvflash has an option to override the board ID mismatch (-5), but, NVIDIA stopped allowing it with Turing cards. So, you need a patched version of nvflash that ignores this mismatch in order to complete the cross-flash. The problem is that the 3080Ti is "too new" and a patched version of nvflash that supports it has yet to surface. I think that it will inevitably appear. Maybe in just a few weeks/months when people want to start cross-flashing GeForce 4000 (desktop) cards and someone is motivated to mod a newer version of nvflash. In the "old days" you could modify the vBIOS file itself and make all sorts of things work. You could just change the IDs in the vBIOS file to match what you needed (as well as arbitrarily set the clock speeds, power limits, etc.). My own flashing experience comes from messing with overclocking the old Quadro K5000M. Anyone could produce vBIOS images to make it run faster than stock. Since Pascal, NVIDIA GPUs are enforcing digital signatures on the vBIOS, so only NVIDIA can produce valid vBIOS images.

NVIDIA Inspector shows it on the main window. P0 is "best performance", P8 is the low-performance "idle" state, and there are some in between states. NVIDIA Inspector also has command-line flags and you can force the P-state to a certain value using those. We've taken notice in the 7X70 thread that Dell blocks P0 if you have the system set to the "quiet" or "cool" thermal mode so you take a big GPU performance hit in those modes. Got me thinking about if something is causing the GPU here to drop to a lower mode or if it actually staying in P0 but at a reduced clock speed.

You just want to have them send the "same part" you already have... So far, I think that everyone who has tried has ended up receiving the Sunon part. They should be able to figure out the part number (they shouldn't even be allowed to send you a part that isn't already in your system), but if they can't, you can find it yourself. Go to dell.com/support, put in your service tag number, and then click "View product specs" on the next page and it will take you to the full parts list. Look under the GPU section for one labeled "ASSY,HTSNK".

I am curious about this, it seems pretty odd. I am wondering if maybe just an early/testing vBIOS was included on their card by mistake. Did this user give the version number for the vBIOS in their system that was reporting 157W TDP? It should be visible on the very same screen in NVIDIA control panel where you can see the card's max TDP. NVIDIA is good about never reusing the same version number and kicking out a new one for each vBIOS variation that they produce. (Just another thing that could maybe be used to quickly tell which vBIOS you have, or maybe even tease out which one is "newer" than the other.)

If anything it will improve performance at the cost of higher power use / lower battery life. It prevents the CPU from dropping to lower power states in certain conditions. ...Though I did some testing on my Precision 7530 which had this issue and I found the impact to both performance and battery life to be pretty negligible.

Classic Precision issue. Try going to BIOS setup and disabling C-states. It should be under the "performance" section. I just noticed that there is a setting there about C-states and discrete graphics. I wonder if that "feature" is involved with the dGPU "stuttering" issue that I described above... I will experiment with it. Unfortunately, Dell doesn't have separate part numbers for variations of the same heatsink assembly. (You'll get different part numbers for i.e. dGPU present or no dGPU present, but not for which type of fans are in there.) Did you get a 7670 with Delta fans?

I got to thinking. Did you notice what the NVIDIA GPU's power state was when it was doing this thermal throttle? Did it bump to P0 down to something lower like P3? I wonder if forcing it to stay in P0 with NVIDIA Inspector would change the behavior...

I noticed some regular stuttering behavior in multiple games using the dGPU. Like, there would be a noticeable blip of perhaps a quarter second with the graphics “frozen”, and it would occur every few seconds. It was not consistent and rebooting would fix it for a while. I couldn’t blame it on DPC latency and there was not a corresponding audio blip. I think that I have “solved” it by disabling the “hardware GPU scheduling” option in Windows advanced graphics settings. (A search through Reddit shows many people complaining about this option causing odd behavior.) But, that was just a few days ago that I toggled that and it hasn’t been long enough for me to be sure that it is fixed. I was planning to mention it here after being more certain about it. I am on Windows 10 21H1. (22H2 seems to have no real changes from 22H1.) I’m using the Intel drivers from Dell but recently moved to the Studio drivers from NVIDIA. (I experienced the stutter with both Dell and Studio drivers.) I have not experienced the black Chrome tabs issue that you mention… but I primarily use Firefox ESR. I do have NVIDIA control panel set to use the integrated GPU for all apps by default and I manually whitelist apps for the dGPU, so there is no reason why the browser would be invoking the dGPU.

7770 + 3080Ti does have switchable graphics. I’ve been using it. I have found the system to be unreliable with switchable graphics off. The system works but sometimes will not “wake up” after having the display power to off. Once, I had a garbled display while trying to get to BIOS to turn graphics switching back on and I had to use an external display to do that. (So, it does not seem to be a Windows/driver problem.) No such issues with graphics switching on. I’m generally fine with this configuration so I haven’t decided what I’m going to do just yet, but it will probably be an eventual warranty claim. I sort of want to see if other systems show this issue (I hate doing warranty parts replacements only to end up with the same problem). I have been gaming on the system, so I can say that the 3080Ti otherwise works fine. I did record a couple of videos of the garbled BIOS boot graphics.

Cinebench runs at low priority by default so anything else running on your system (including background services) can chip CPU time away from it. If you bump it up to "Above normal" priority then background services should become less of a concern.

This thread is starting to get pretty long. I've updated the OP with some suggestions for performance improvements and links to key posts to try to make this information easier to find. Any suggestions, let me know...

Check:https://www.reddit.com/r/GamingLaptops/comments/np5vrp/comment/h05i740/

Ok, so it got past subsystem ID, but not board ID. Looks like nvflash is pickier about this than it used to be. I don’t know if there is a patched version that works with Ampere cards… https://www.techpowerup.com/download/nvidia-nvflash-with-board-id-mismatch-disabled/

Did some looking. Just put “-6” in the flash command.

This is normal since we have different systems. There is a command-line flag to override it and flash anyway, "overridesub" or something like that.

Yeah, as stated above, NVIDIA control panel shows 150W TDP. Nope, if graphics switching is disabled then the iGPU is disabled, even in BIOS. If there was trouble with the dGPU starting up, you'd have to reset the BIOS to enable the iGPU again. (That was easier back when the coin-cell battery was readily accessible...)

NVIDIA control panel shows TDP of 150W, but in practice it doesn't seem to go above 130W. I dumped the vBIOS with nvflash. I recommend that if you are going to flash a new one: Back up your original vBIOS in case you need to roll back to it. (@Ionising_Radiation tried this trick on Precision 7560 to improve GPU performance but ended up having to roll back because it was causing issues, I do not recall exactly what.) Have graphics switching enabled (easier to recover in case of trouble). Use Linux to do the flash if you are comfortable with it. You can do it from a "live USB" environment. (Less chance of trouble. I've had multiple occurrences of BSOD in the middle of a flash on Windows. Can be tedious to recover from if it is interrupted.) [Edit] Here is @Ionising_Radiation's post after the flash. https://notebooktalk.net/topic/24-precision-7560-precision-7760-owners-thread/?do=findComment&comment=12168 7770-3080Ti.zip