Mobile Pascal TDP Tweaker Update and Feedback Thread

Mobile Pascal TDP Tweaker v1.2.1.0 (07/07/2019):

1. Added support for mobile GTX 1050 Ti (credit to MultiRW/DymOK);
2. Fixed issue that RTX 2060 was recognized as RTX 2070;
3. Added option to fix black screen issue for some vBIOS;
4. Added display for device id;

Mobile Pascal TDP Tweaker v1.2.0.0 (06/30/2019):

1. Better support mobile RTX cards (2060, 2070, and 2080);
2. Added support for mobile Quadro P3000, P4000, and P5000;
3. Added support for mobile GTX 1050;

Mobile Pascal TDP Tweaker v1.1.3.0 (03/24/2019):

1. Added experimental support for mobile RTX cards (not tested);
2. Bumped TDP preset values for GTX 1060, 1070, and 1080;
3. Temporarily disabled support for desktop cards;
4. Updated target .NET Framework to 4.6.1;
5. Source code is available now;

Mobile Pascal TDP Tweaker v1.1.2.0 (12/22/2018):

1. Extended support for newer OC vBIOS (88W 1060, 125W 1070, and 180W 1080);
2. Added more options for temperature controls (experimental);

Mobile Pascal TDP Tweaker v1.1.1.0 (06/26/2017):

1. Added support for vBIOS extracted directly from the chip-set;
2. Added support for Titan X Pascal and Titan Xp(experimental);
3. Better support for some vBIOS that have no name;

Mobile Pascal TDP Tweaker v1.1.0.0 (06/20/2017):

1. Small bugs fixed;
2. Black screen fix for some non-gsync vBIOS and old vBIOS;
3. Interface redesigned and polished;
4. Added "Preset" button (only for notebook cards);
5. Added "Extreme" section for hard-core users to completely remove power throttling (not recommended, inappropriately modifying these values can easily burn your card!);
6. Added convenient "Tools" section for hex-editor users;
7. Support desktop Pascal vBIOS (experimental; 1060, 1070, 1080, and 1080Ti only);

Download, details, and disclaimer:
https://github.com/LaneLyng/MobilePascalTDPTweaker/releases

How to check chip-set model:
http://forum.notebookreview.com/thr...-feedback-thread.806161/page-42#post-10584406

Clarification of " Pin 1" on the chip-set:
The Pin 1 on the vBIOS chip-set is the one next to a tiny concave dot on the chip, it has nothing to do with the White Dot.
The white dot is just/only an indication of that chip is a vBIOS chip. Incorrectly connecting your chip with the clip can result in damage of the chip-set.
Please see below links:
http://forum.notebookreview.com/thr...d-feedback-thread.806161/page-4#post-10551261
http://forum.notebookreview.com/thr...-feedback-thread.806161/page-17#post-10561646




Experiment of how TDP values affect performance under heavy load on the GTX 1070 (this can be a reference if you would like to modify your TDP value):
Core clock added 189MHz, memory added 450MHz during the 10 runs.


Another TDP test with default clock and under-volted:
http://forum.notebookreview.com/threads/clevo-overclockers-lounge.788975/page-1317#post-10561586

Successful examples:
GTX 1060:
MXM: http://forum.notebookreview.com/threads/mobile-pascal-tdp-tweaker-update-and-feedback-thread.806161/page-61#post-10615246
Onboard: http://forum.notebookreview.com/thr...bios-modification.810393/page-2#post-10636802
GTX 1070:
MXM: http://forum.notebookreview.com/threads/mobile-pascal-tdp-tweaker-update-and-feedback-thread.806161/page-29#post-10565209
MXM: http://forum.notebookreview.com/thr...0b-alive-thread.795048/page-152#post-10550587
MXM: http://forum.notebookreview.com/threads/mobile-pascal-tdp-tweaker-update-and-feedback-thread.806161/page-8#post-10554355
MXM: http://forum.notebookreview.com/threads/mobile-pascal-tdp-tweaker-update-and-feedback-thread.806161/page-18#post-10561701
Onboard: http://forum.notebookreview.com/thr...-feedback-thread.806161/page-36#post-10568711
MXM: http://forum.notebookreview.com/threads/mobile-pascal-tdp-tweaker-update-and-feedback-thread.806161/page-64#post-10623678
MXM(top score): http://forum.notebookreview.com/thr...-feedback-thread.806161/page-76#post-10641592
, etc.

For Clevo laptops, if there is no effect observed after successful flashing, please try:
http://forum.notebookreview.com/thr...-feedback-thread.806161/page-71#post-10634716

Tutorials of modding your own Pascal vBIOS through hex-editor (credit to Dreamonic and alexio):
http://www.voltground.com/haven/threads/94/
http://www.xtremesystems.org/forums/showthread.php?180607-Tutorial-How-to-fix-a-bios-checksum

 

Requires a BIOS hard programmer, hmmm.......

 

oooh liking likable motherliking like.

 

Temp slider doesn't work.

 

Sweet! I modded the 1070N vbios to match a the 1070 FE vbios.

 

Right, the temp slider doesn't really work, as well as the throttling temp for dynamic clocks.

 

Did it work before?

 

Not really. Checking the "Adjustable" for temp slider only makes it showed up in Afterburner or Inspector, it couldn't really set the value at this moment.
So you can just leave it "Fixed". The same with throttling temp, even thought it set to 91C, the dynamic clocks still start dropping around 54C.

 

hopefully you can find a solution.

EDIT: I had a look at the MSI 1070 FE (as well as other 1070 vbios from MSI and other vendors) and the throttling temp was only 60 degrees compared to the MSI 1070N's 54. A 6 degree difference isn't a big deal but being able to alter it to 60 and work would be nice.

Also, does altering the rated and max temps have any effect?

 

Nice. Thank you for creating the tool, it should come in rather handy, especially once we figure out how to flash the modded vBIOS.

Curious. Is there any chance you're going add support for the Titan X (Pascal) and Titan Xp Desktop vBIOS in the future as well?

 

I'd prefer the throttling issue to be fixed first before support for more cards. I'm sure there's people who can work the problem.

Sent from my SM-N910G using Tapatalk

 

No problem. I can add the X (Pascal) and Xp for the next update, .

 

If you keep the GPUs cold it will fix that. It's really stupid how NVIDIA designed their firmware. WIth proper firmware, the GPUs should run at full boost until they reach the therrmal shutdown threshold. That's the caveat. NVIDIA did not release "proper firmware" for any Pascal GPU. The retarded dynamic clocking nonsense is for the birds.

 

Right. It's like every 6C higher or so it will drop 13MHz, started from 54C.
By keeping the card as cool as possible and playing with the voltage curve in Afterburner, it helps a little bit in mitigating the unwanted drop.

 

I meant in terms of the slider not working.

And it actually starts at 42C. At least it is for the 1080N.

 

Do you mean the temp slider is locked in Afterburner or the temp slider value can not really be set?

Sounds about right:

 

With stock bios back in Dec 2016
http://forum.notebookreview.com/threads/clevo-overclockers-lounge.788975/page-721#post-10403967

http://forum.notebookreview.com/threads/clevo-overclockers-lounge.788975/page-737#post-10408725

 

I keep my fans at 100% even when idle. In games that equates to about 51-52 C in an Australian winter or mid 60s in the summer. Both cases the clocks don't hold steady.

 

Power/Voltage limits. An obviously temperature to a certain extent, if the temperatures are low enough.

 

I don't know if bumping the TDP would fix that

Sent from my SM-N910G using Tapatalk

 

It will , again, only to a certain extent and for certain loads.

 

It's a little bit weird that my card scores higher(in Heaven) when feed with higher voltage for the same clock, after the core temp has settled down and made no changed. For example, it scored almost 100 points higher in Heaven with [email protected] than [email protected] (sharp rising curve).

 

I think we'd all be better off without the stupid GPU Boost gimmick like we always have been in the past. It would be better to be able to just set temp target, power target, voltage and clock speeds at their maximum stable values and have them stay there forever, or until the GPU overheats, whichever comes first. Dynamic "optimized" crap has always been retarded. I want it to follow orders and stop trying to figure out what the best approach is. That's my job and I'd much rather have the same kind of control over the GPU like I do the CPU. Pretty much everything NVIDIA does with firmware is just bull-feces and it has been that way for years. It's just more spastic now.

 

I can't really say at that point. We do things differently so What works for you I say go for it!

 

Pure old school And it worked best... no performance drops, no issues with stability, full knowledge about current voltage curve... and now it's a damn temp-lottery. Personally I hate this boost since 980M - it was problematic there, but with Pascal... it's just annoying af... "Funny" games finding sweetspot for OC (that minimalizes max stable OC by much) etc...

 

Somewhere around there; I think it might be closer to 45-48c for me. Was able to see 1911MHz boost for a bit last night I think around 43c, will check my screenshot

Sent from my OnePlus 1 using a coconut

 

42C is the reset back to it's original clock before it dropped.

 

Thank you for this tool!

Stupid question: is it possible to flash the vbios made with this tool or flashing method is still a thing of the future?

 

No. you can not flash with this tool. You still need a spi programmer.

 

That is not a stupid question, but the answer takes stupidity to an unprecedented level.

Thanks to the Nazis at NGREEDIA, flash protection implemented on their latest filth pretty much ruins the experience of owning a Pascal GPU for 99% of the enthusiasts out there who would normally expect to have more than a belly-button experience.

This is an example the retarded mierda perro you will see now if you try to flash a vBIOS that is unsigned or had its signature broken by modding.


There is only one way to effectively deal with this immeasurable wickedness from the Green Goblin, and it's not for everyone... unfortunately. If you want a better vBIOS, this is the only way you'll get one on the GPU. You'll need the special tools, have to disassemble things for every flash, and a second computer to do what used to be done in just a few second with free software. Now, isn't that special?

 

i'm still waiting for my SOCI8 clip adapter to arrive as it should have arrived already.

i hate the postal service sometimes.

EDIT: Any update on the temp throttling issue?

 

Location of the chip-set (MXIC 25U8033E) in my MSI 1070, usually the chip-set is marked with a painted dot by factory:

Inline flashing through connecting a compatible programmer with another computer:
Edited: The Pin 1 connection showed on the below picture is not correct, please refer here for where Pin 1 is located: http://forum.notebookreview.com/thr...d-feedback-thread.806161/page-4#post-10551261

 

I assume you're matching the pink cable to the corner with the dot in one of the corners (pin 1 I think)?

Sent from my SM-N910G using Tapatalk

 

Yes, that is pin #1. If you don't do that the programmer will error out and you won't be able to ID the chip correctly or complete the flash.

 

Like this in the below picture. My second picture above showed a mismatching connection on the chip-set, please neglect that.

 

Awesome. My programmer is on its way. Still waiting for the clip POS to come...starting to annoy me.

Sent from my SM-N910G using Tapatalk

 

Be sure the programmer supports the lower 1.8V chip or get a 1.8V adapter or it will fry the vBIOS chip using full voltage. (You may already know, but I mention it for any readers that do not.)

https://www.amazon.com/gp/product/B01GFMYICM

 

THIS ->>>

Do not even attempt with out the 1.8V adapter. Also which programmer did you order?

 

I bought the same one as @Coolane, the sofitech sp16-b. I even looked up the supported chip list and it was listed. Even played around with the flashing app and it lists the chip.

Sent from my SM-N910G using Tapatalk

 

I think @Coolane (and the one you ordered like his) are one of the few that does, but you have to use the correct section of the connector to get the lower voltage. If you connect the alligator clip to the area one would normally connect it you will get full voltage and may torch the chip.

Most programmers require using the adapter. My TL866A requires that I use the 1.8V adapter, but it directly supports the chip and the chip is in the list even though I have to use the 1.8V adapter. So, the moral to the story is that it is not safe to assume anything based on the supported chip list.

 

Duly noted. So long as I plug it into the right pins and select the right voltage it should be ok...hopefully.

Sent from my SM-N910G using Tapatalk

 

Yes, exactly.

For the benefit of anyone else that does not know, this photo that @Coolane posted shows it. Notice the connection is in the area farest away from where the locking arm is placed. That is the 1.8V connection area. In all other circumstances (on all SPI flash programmers) the area where you insert the connection for BIOS/vBIOS flashing (SOIC8 chips) is in the slots closest to the locking arm.

 

Yup, even the app demonstrates that but you gotta find it LOL.

Sent from my SM-N910G using Tapatalk

 

Yes, I'm sure that you're ready to go because you have been reading up and preparing yourself. I just don't want to see anyone else make a huge mistake that costs them a lot of money or hassle due to lack of knowledge (or not reading instructions). This is really a safe way to flash things as long as it is done correctly. I think it's safer that software flashing in DOS or Windows when done correctly.

 

Fair enough. Looking forward to giving it a try.

I feel so 1337 man!

EDIT: The SOIC8 clip adapter came

 

My programmer came this week (a TL866CS converted to TL866A) with a 1.8V adapter and the SOIC8 test clip.
Before using it on my Alienware's 15R3 mobo, i tried to read an old Gigabyte's Mobo Bios Chip and got the OverCurrent Protection actions 'Message.

I read this doesn't happen with all mobos, and if it does, you have to cut VCC on the chip or hard mod the programmer to bypass this warning. I tried the linux minipro's cli too without success .

I don't know wich ship betwen UT2 and UG13 on the 15R3/17R4 mobo is the VBios, can i try to read those chips safely ? I read keeping AC pluged may help to read, but doesn't allow to write.

Any advices ?

 

Did you connect the clip with the chip correctly with the Pin 1 matched the Pin 1 on the chip? The vBIOS chip looks like this, model name may vary.
I got something like "Current Protection" before because I clamped the chip in the opposite side.
I feel like taking out the batteries and disconnecting the power supply is safer to do the inline flashing.

 

Which chip is it?
Either ways, try disconnecting pin 4 on the clip, or put a thin strip of tape or something to prevent it from making contact with PIN4 of the SOP8 chip.

In some cases you will only be able to read the chip using this method, but not write.

The easiest method is to get a SkyProII , that can easily read and write inline. And its cheaper than the TL866CS/A.

@Mr. Fox This is the third time other than my own experience, that a recent TL866 is behaving this way.
They have definitely changed something in the recent/last years batches.
End of the day i just got rid of the OvP protection with a hardware mod to the programmer.

 

Very nice! It has a built-in screen and supports low voltage 1.8V, also cheaper than the SP16B.
http://www.coright.com/product/showproduct.php?lang=en&id=55

 

Thank you!
@Coolane @bloodhawk do any of you have a direct link for buying the SkyproII? I can't find it on Amazon D:

Do I really have to buy it from China or something? If there are any other options (even aliexpress or banggood) let me know. I wish I could use my amazon prime shipping

And I can't read chinese and have no idea what to do on their taobao link....