JadeRover

Hello eveyrone, Ok so this problem applies only to MXM gpus that have a "pwr" limit (as well as "thrm" limit sometimes) in GPU-Z when doing a benchmark. This is also known as the "40w power bug" Such example : Notice this is not due to hitting the power limit of the card (33% of TDP). Symptoms of the low power bug is a TDP usage around 50% = 40w if you have a TDP of 80w. I think I found the issue. It has to due with a digital signal that is sent to the GPU, the PWR_LEVEL signal : A previous fix was to block the SMB_DAT or SMB_CLK signal so that the EC chip of the laptop wouldn't be able to communicate with the MXM dGPU. This has the disadvantage of stopping all data communication between EC and MXM dGPU = no more thermal data coming in = EC usually ramps up the GPU fan to maximum. According to MXM 3.1 documentation : This signal is not mendatory and it does apply some kind of power reduction when enabled. Therefore I took a gamble and decided to cover it up and see what happens. My fix is to block off the PWR_LEVEL pin with a bit of tape. - I first put some normal paper scotch tape on my plastic desk (to get it nice and flat during cutting process) - I cut it into a thin strip with a very sharp cutter(width of MXM pin) - I removed it from my plastic desk, then applied it with tweezers. The pin is the 9th counting from the first "thin" pin, on the TOP of the MXM module aka side of the MXM gpu where there is the actual GPU die : It's okay if you cover up the 8-5th pin as well as these are reserved pins (no connection inside a laptop) I did this on my dell precision 7720 with GTX 1060M and this has so far resolved the issue, my card can now boost to 100w thanks to the modded bios. Hope this helps someone !

You are right, I think it should work afterall, using nvflashk. But I would really highly recommend a flasher anyway as the first vbios I can make will prob result in bricking the gpu unill fine tuning is done. Also, my program is coming along nicely, I can reliably find the Core clocks and mem clocks of pascal. I can also reliably get the core + mem clocks of Turing and above gpus (tested up to 4060m vbios) . But this is only using the virtual p state table. Meaning I can only edit the values of "boost" you see in gpu z. This is not the actual boost value : rtx3000 can boost to 1820mhz and that clock is not in the v p table. It must be somewhere else. This is not the case for pascal as this second "real boost" clock is in the v p table. They must have done slight changes between pascal and further generations. Because of boost 2 -> boost 3 modes. Mem overclocking is possible though, as well as power limit modification.

I can probably write the code this week, thing is I would need people to test the vbios with fixed checksum that are made, I only have a GTX1060 (in a dead precision 7720 I gotta fix) + Flashing would only work with an spi programmer, CH341a and such. modifying GPU clock + MEM clock is usefull only for quadro cards that can't overclock in software, all GTX + RTX cards have OC sliders unlocked in MSI afterburner. Power limits modification is usefull for Quadros + GTX cards I guess.

Here, look in hex for "28 4F 4C" it's the same from pascal -> blackwell The actual TDP values is the "text" are after the (F...(F -> Here it's oS...oS -> in hex it's F8 24 01 -> after little indian 01 24 F8 = 75 000(d) so 75w. (example is p3000 vbios) Def + Max value (not sure of the order though) I could probably write a small tool that can sniff out and find the power + GPU clock + MEM clock for pascal -> blackwell gpus. Shouldn't be that hard, problem is fiwing the checksum after modifications.

Actually out of curiosity, I went to inspect the "low power" ram frequency string out of a P5200 vbios that has ram speed of 1802Mhz. As you can see here : Despite the different layout : The first ram speed string with DDR value and 2xDDR value is highlighted in dark blue and the second string (that I called "low power" in the post before) is actually set to the same value of : you guessed it : 03 85 CE 17h = 59 100 695d -> devided by 2^15 = 1803 Mhz !! So I guess to overclock ram : users should change the first ram string containing the two DDR speeds as well as this second string that is located between the first P state entry and the second P state entry. This should, in theory, levitate the 1752Mhz memory lock. Testing needs to be done ! I won't have my p3000 for a while unfortunately.

Hello @Ralph. I know it's been about a year since you posted. I have, like you, found where the GPU clocks and memory clocks are stored in the pascal quadro vbios -> the Virtual P state table that can be found thanks to ImHex... I will make a tool to allow users to change clocks in vbios once I have gathered all info needed + done tests on my quadro p3000. Did you find a way to reference the voltage for each clock ? Or is it just when applying a clock value that the card automatically decides what voltage to go to ? This is based on the vbios you posted here. Clock values are pretty close. Highlighted blocks must be read in little indian (17 CE 85 03 = 03 85 CE 17). And at the exception of the memory frequency 2x DDR. All values must be read by dividing the decimal value of the hex block (in little indian) by 2^15 Apparently memory frequency can't go above 1752Mhz on pascal quadro cards.

That's unfortunate, I can take a look at your vbios if you want

OK, no choice unfortunately, you must get a 3k motherboard + 3k cable for a 4k upgrade

Yes, the m1000m MXM card was shipped in both dell precision 7710/7510 and zbook 15 g3/17 g3. The display outputs through the MXM slot to all the connectors of the laptop are implemented in the vbios. Flashing the bios that you found working would have made this first card work.

@loopster https://notebooktalk.net/topic/225-precision-m6700-owners-thread/page/6/?&_rid=2023#findComment-61191 Here you go, this guide is for the m6700 DC screen. I applied it on my m6700 and the 10 bit mod worked. However I couldn't tell the difference with it on or off, even trying specific 8bit vs 10 bit downloaded videos. Also the DC screen heats up a lot at 100% brightness. Probably the interposer board that is working hard.

Hello, This guide will explain how you have to actually modify a pascal/turing vbios with pascal mobile tdp tweaker in order to not get error 43 and black screen. latest vesion 1.21 : https://github.com/LaneLyng/MobilePascalTDPTweaker/releases/download/1.2.1.0/Mobile.Pascal.TDP.Tweaker.1.21.zip tools needed : any hex editor -> HxD for example I have embarked on a long vbios modding adventure. End goal : custom display tables in vbioses utility -> Extremely useful for MXM gpus -> looks quite possible thanks to released nvidia documentation but I must invest lots of time in research into this (my time allocation will vary greatly week to week...). Side goal : create a core + memory overclocking utility -> Possible as well, simpler than the display table Anyways, early on I noticed a flaw in the pascal bios tdp tweaker in the way it calculates checksums : it only corrects the checksum to the correct value if your vbios has no header ! If you use a vbios with header the app will calculate a wrong checksum and you will get error 43 or black screen as the security chip on the GPU will refuse to post. Step 1 : Check if your vbios has a header -> Open it in a Hex Editor and see if the starting string is "NVGI" -> if no header, you can mod the base directly in TDP tweaker with no extra steps. example of a vbios with a header : Step 2 : Remove the header and keep it in another file, select all the first lines of the vbios untill you see the string "UxxxK7400". In my case the actual vbios data starts at offset A00 -> Save the modified vbios as "vbios_mod_applicable" Step 3: Open this "vbios_mod_applicable" in mobile TDP tweaker, apply your edits and save the bios, now the checksum is correctly calculated automatically when you save this modified version. Step 4: Open the TDP moded vbios in a Hex Editor and reinsert the header at the very beginning of the vbios. Make sure the file size matches between the OG vbios and the TDP modded one. Step 5: Flash the modded vbios with an external programmer -> nvlashK and OMGflash will not work. You need to have a flasher handy anyway if you are attempting a vbios flash that risks bricking your card. Step 6: Enjoy higher TDP/other changes you applied

You can get a dell 240w charger to plug into the barrel jack of the zbook 17 g2 + add 100k resistor between the ID pin and the V+ pin so that laptops sees it as 200w charger. Then shunt mod the input power resistor so the system won't shutdown from over current at >210w. You should be able to run 230w load with 150w + 80w cpu. However you will probably need a heatsink mod as single fan cooler will quickly be overloaded. Luckily people using the zbook 15 g1/g2 frequently do this by adding a fan in the ODD space + extra heatpipes.

@Vladka76 Yes, this should work, nice that 8770w has DP_D -> eDP, this will make GPU compatibility greater. I noticed a trend that up to zbook 15 g3. All 15 inch HP worstations with MXM3 connector have DP_C -> eDP which is a shame = special HP vbios needed or mac rumors vbios needs to be used (best card is the P5200). If not, there will be no picture on LCD. Unless you have a vbios that has standard DP_C -> DP (standard display port), then a backligh mod can be performed but screen is stuck at 100% brigthness. Couldn't find boardview, schematic is preffered tbh so good that you have it. I think you will need to mod the specific "Dream Color" cable that came with these laptops, the standard LVDS cable won't have cables connected to the pins associated with eDP, they will be unpopulated as why would your LVDS cable need to carry eDP wires ? Alternatevely you could repin a LVDS cable to connect to the eDP lines. What I think is easier is : getting DC cable, cut off the end that connects to the interposer board and solder a standard 40 pin ipex connector used by eDP displays (if cable length is enough to connect to display, it should be). Only problem is that eDP cables used or high bandwidth, which is the case here, need special coaxial cables that are HELL to solder (experienced this the hard way when trying to mod a zbook cable). These are usually soldered with lasers in factories D= Might be worth it to find a manufacturer who can build you the cable, on tabao or even aliexpress, sellers you make custom cables should offer it as it isn't much work for them. PS : Your pinout is correct, but you do need pin 44 that is eDP HPD (hot plug detect), required for eDP, not sure what the MARS signal pin is, you can trace it back on the schematic, probably a GPIO pin to tell the PCH or EC chip that a DC display is present in the system.

I double checked again on the boardview of the 8570w motherboard and this is not the case The motherboard never converts LVDS to eDP, not possible, same for the cable or interposer board, the technology to convert LVDS to eDP came after these laptops were built. Famously the thinkpad T430 'IPS' mod uses a third party LVDS -> eDP converter. On the screenshot I showed just above, the motherboard LCD connector has both eDP and LVDS on the same connector, both coming from the GPU in the MXM slot, in a direct line, no conversion going on in the motherboard. The eDP cable you are talking about doesn't convert any signal, it just connects the eDP found on the motherboard screen connector to the eDP display. A LVDS cable that frequently came with these models connects the LVDS found in the motherboard screen connector to the LVDS of the screen. I talked about how zbooks do this. All zbooks, (starting g1, 15/17 models) have only eDP on the motherboard screen connector. After that, the cable carries the eDP signal to an interposer board -> see my post at the top of the page showing pictures of such interposer that converts eDP to LVDS. The problem is that the 8570w never shipped with a standard eDP cable that could connect into a modern eDP display. It was made to connect to the 40 pin eDP interposer board used to communicate to the Dream Color LVDS LCD. On the pictures of such interposer board, the connector that get's data from the motherboard is non standard to eDP display. Good news is that, unlike thinkpad T430 that needs an added converter board to get modern eDP displays working, 8570w just needs a modded cable that has it's end modified so that it can connect to a standard eDP display connector.

Some 15.6 and 17.3 inch laptops of that era (ivy bridge + Haswell) used eDP to LVDS converters so that they could connect to LVDS displays that were more common at the time. You are in luck as your model came with a 3K option, what you have to do is hunt down that 3K eDP cable as it can only connect to a 3K eDP display (LVDS tops out at 1080p). Then, you'll have to check if the connector is 40 pins 0.5mm pitch or 40 pins 0.4mm pitch, after that you can buy a 40 pin display of your liking (making sure the pin pitch + connector position will work with the 3K cable). You can chose 1440p 144Hz for example I would not go up to 4K resolution as older systems can't deal with the bandwidth if the internal eDP connection is limited to DP1.1 (not always the case : happens if there is mux switch in the way, example precision M6700 has no mux switch = DP version limited by MXM gpu installed). If you really want 144Hz, the safest way is getting 1080p 144Hz as even if it is DP1.1 limited because of mux switch, you can still get 154Hz max