Mr. Fox's GeForce 344.75 Desktop Driver Mod for 780M, 880M and 980M Mobile GPUs

I did this driver mod a few days ago and it seems to be working well for me in benchmarks and games, so I am sharing it with you folks to see if it works well for you also.

It should work on any Alienware or Clevo laptop because it only identifies the GPU hardware ID, so Clevo owners should feel free to try it out. On-the-fly-SLI mod is not necessary, as it works by default using this desktop driver.

Note: The nvcvi.inf and nvdmn.inf files have been intentionally messed up, so don't try to fix it. Those files have to be present for the driver to install. I tried removing them entirely, but the driver installation apparently needs to see them and fails without those files sitting in the folder, so I screwed up the hardware ID code for 780M, 880M and 980M to force use of the nv_disp.inf file exclusively. (Clevo owners will need to mess up their hardware ID in the nvcvi.inf like I did the nvdmn.inf for Alienware.)

Windows 8 users will need to reboot with driver signature enforcement disabled or the installation is guaranteed to fail. Temporarily disable HWiNFO64 (rename the installation folder works easy enough) to avoid hard freezes. Use DDU in Safe Mode before installing to remove all traces of your old driver and all registry entries from the old driver before installing it.

Download Mr. Fox's 344.75 Desktop Mod v2.0.7z

Here is a moderate GPU overclock on Fire Strike showing the driver works fantastic for 780M. This should be achievable on an SLI machine with a single AC adapter. With 980M stock core clocks are higher than this 780M overclock.



I hope this works as well for other 880M users as it has for woodzstack...

---

Try this option to disable DSE. It works and I just recently discovered the command.

Open an elevated command prompt and run the syntax below. Reboot, and DSE should be disabled for good. No need to temporarily disable DSE the old way.

Code:

bcdedit.exe /set nointegritychecks on

I am now able to install modded drivers without having to boot with DSE temporarily disabled. Works on the Sky X9 and Panther with W10 TH2. I've done it probably a dozen time in the past two weeks.

 

I won't have them until tomorrow. It will be next week before I share anything on 980M. I am fully expecting Windows 7 will not work on Alienware systems, but there is always hope. I even ordered a non-Alienware retail Windows 7 SP1 x64 Pro DVD to see if I can get it to work by some bizarre miracle, but I am not optimistic. There is something about the Alienware system BIOS not playing nice with the vBIOS and only GOP booting is supported. It seems that Windows 7 doesn't function without CSM/Legacy Option ROM except for a handful of lucky people with unusual circumstances that mysteriously worked out for them. I have torn up the internet searching for a way to run Windows 7 in pure UEFI mode (without CSM/Legacy mode) and the inability to do so seems almost universal, not unique to Alienware.

Those curious about the Alienware 18 working with 980M SLI already have proof that it works courtesy of Brother louis1978uk, so go ahead and grab them if you want to upgrade. M18xR1 owners will need to wait until I have time to test it. Testing on the M18xR1 will be super easy since I don't have to swap around the heat sinks. Alienware 18 uses different heat sinks and they will not cross fit with the M18x chassis. They are almost the same, but just enough difference to not fit and I don't want to damage the fins on the radiators by wedging them in.

 

Mr fox., what did you change?

 

That's a secret, but if you look hard enough you can find it.

 

Thanks, Mr. Fox. You are the best. Enjoy your 980M cards. If you have those, you will make miracles.

 

I'm a little lazy right now, but sometime in the next couple of days I will give these a shot (mainly because I hate redoing my per-game profiles in NCP). Maybe I'll finally be able to keep a daily 1006/1500 clock speed on these things. Just because. JUST BECAUSE. *strokes 780Ms*

 

Okay, got pissed because I can't keep 1100/6000 in Valley. Gonna try these out right now XD. (There's this one guy I'm determined to beat. He has a single 780 and is beating my setup, this cannot be allowed.)

 

I notice your 780M is running at 3.0 x8 only and not full 3.0 x16... though its an improvement form somepeople who get it stuck on 1.2 or 2.0 @ x8

you will get better performance when you can keep it tripped at 3.0 x16 of course. Which is what you get with the stock drivers using 680M I believe.

Also, do the 980M's work in the M18X-R2 and the M17X-R4 ? (People have stated it works well in M17X-R4, but have not shown me any proof, and Ive yet to really research it thoroughly)

Also, whats so special about this driver mod ? The normal INF mods worked fine on the 344.75 driver and I have them installed, and have them working for 780M on my m17x-r4 and the 880m's on the m18x-r2

 

This is impossible on a SLI machine. 3.0 @ 8x is the highest you go because the CPU only has 16 PCI/e lanes. It also won't affect bandwidth much; these cards don't saturate the bus at all even at 8x.

----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Mr. Fox, the drivers don't work for me. I totally removed them and used command prompt to fully disable driver signature enforcement and rebooted, and I tried multiple times but I kept getting the same error below:

 

That's correct.

Hmm, strange. I have clean installed it a couple of times and mikecacho had no trouble installing it. I did not do anything unusual for Alienware versus Clevo. I think you may be having a Windows 8 cancer thing going on because you got as far as you did. Otherwise, the installer would have stopped at the point of not detecting GPUs. Your screen shot is EXACTLY what I see when I try to install modded drivers without disabling DSE first.

Open a command Window with Admin rights and run the commands bcdedit -set loadoptions DISABLE_INTEGRITY_CHECKS and bcdedit -set TESTSIGNING ON, then reboot with the Windows 8 Troubleshooting options menu and choose F7 one more time. If that works and you have a watermark test mode on your desktop, kill it with this application. (I run Windows 7 and 8 in test mode with integrity checks disabled 24/7. You can turn them back on again if you want to.)

Or, if you prefer the Windows 7 legacy menu that works with F8 before Windows logon instead of Windows 8's retarded way of doing things, you can turn that on with the command bcdedit /set {default} bootmenupolicy legacy (also requires command window with Admin rights).

 

The "disable integrity checks" is what I used. I didn't try "test signing on" yet. I think I'll try that bootmenupolicy legacy factor later too though; I can't stand the way I have to ask win 8 to start it.

Also, I tried doing a clean install in safe mode too, and it failed there. Usually with normal drivers, it'll finish and ask me to restart the PC. I also did check; there were no leftover drivers when I was trying to use it. So that's weird.

I'll do all that stuff a little later though. I hope it works.

 

It should work fine, but even with test signing on and disabling integrity checks you still have to use the boot menu and select F7 to disable Driver Signature Enforcement. Unfortunately, there is no way to absolutely eliminate it from Windows 8 permanently. All you can do is make it less aggressive with those commands and temporarily disable it for unsigned driver installation. I wish there was a way to just turn it off and be done with it for good.

 

Yeah. Maybe Windows 10 might ACTUALLY be the perfect OS. Win 8 has a lot of things in it that I use a lot, like the better monitor capture in windows and the way it supports my multiple monitors (as their resolutions are mismatched) etc, but it ain't near perfect at all. *sigh*. Welp, here's to hoping Win 10 might be good all around. And STOP HIDING THINGS. At the least, if you wanna make it hard for driver signature etc, do it for the BASIC version. The "professional" edition like I have should have ZERO problems at all with me doing something a "professional" or "expert" would do. Like this.

 

Even desktop cards don't saturate the bandwidth at PCIe 2.0 8x, and 3.0 8x has double the bandwidth of 2.0 so no worries there.

 

^ that too

 

tOp DrIvEr Mr Fox still wiggling these 880ms into submission i know there hot cards but really at 1200mhz there pretty nippy tbh even if it takes 1.150 to maintain stability

These 980m's though are just LOL..
I sit there benching away all happy with myself getting a few more points on the hairy limit of metling my 880s then switch to the 980ms and watch them absolutely STOMP the 880ms in stock,its hilarious yet somehow sad all that time wasted...
god i hope amd produce something special i need moar toys...

 

Well after tesing you driver vs the same driver stock, the 880M seems to output 50% more. I literally am getting 50% more FPS then before... for once, it out performs my 780M's This is very exciting. Thanks alot.

 

Well, I suspect that nobody should hold their breath on AMD doing anything to take the mobile scene by storm. The R9 M295X should run about on par GTX 780M, which is a massive leap for AMD and actually a very good level of performance. But it's almost 2015 and time for something stronger than the 780M from 2013. It's unfortunate that AMD have fallen into a trend of staying two steps behind NVIDIA where laptop GPU performance is concerned. So, I think it is safe to expect their newest flagship to equal NVIDIA's flagship from 12 to 16 months ago.

Wow! That's great news. Thanks for testing it. This should make DumbDumb and Ethrem really happy, too.

 

Frankly i see AMD fading away to make space for another microchip company for far too many years they have made fundamental mess after another... they just cant see to foxus and instead do everything half arsed half decent card worse drivers in history decent drivers same card 3 years in a row... about time we had a new big name let amd faf around with playstation and xbox sad i used to love amd but after an i7 i could never even look at an amd spec sheet again,now gpus arent much better...
Was Offered an aw18 for 900 with crossfire 290's and said no.. you cant give them away...rather have a hotashell i7 and pair of defective 880ms tha anything they have..Besides dell has a lot of spare cards so it doesnt matter if i blow a card or 2

 

I believe Ethrem is getting a pair of 980Ms for his machine, actually. They're fixing all the stuff that's wrong with his machine finally XD.

 

Well, that's even better news. Finally is right. Took forever for them to sort him out.

But, I think DumbDumb is still struggling with Alienware over his little slice of 880M hell. If this driver works as well for him as it seems to for Brother woodzstack, then perhaps it will make it a less hellish affair.

Wouldn't that be something if all of the horrible 880M problems folks have experienced would turn out to have been mostly attributable to buggered up drivers all this time?

 

Oh that would indeed be something. I hope it works for people. Hope I get it working too; would be nice. Will you be doing one with all the drivers if this works well?

I still want to see what my perfect machine would run like haha. I only need to make a few thousand USD first XD

 

Well us desktop 970/980 users still struggle with the ridiculous SLI voltage discrepancy issue, something which nVidia has refused to fix or even acknowledge is an actual issue, instead resorting to ridiculous comments like "is normal" and "the user gains nothing by trying to fix it". That is 100% bona fide BS, because I know for a fact that people have reported gains when fixing the issue. It's blatant lies like these that push me away from team green.

So yeah for desktop 900 card owners nVidia's drivers are every bit as bad or arguably worse than AMD's drivers.

 

So, what did the people that fixed it and reported gains do to fix it? Does the fix work for everyone, or just randomly works for a lucky few?

I remember the 580M throttling problem like it was yesterday and NVIDIA never fixed it, but end users took care of it by matching P1 to P0. I wonder if a vBIOS mod to match clocks and voltage across the "wavy plane" to force static 3D performance regardless of load state would help on the desktop cards?

You know, the whole concept of having different domains that vary based on loads is just downright stupid. It's a "green" imbecile way of thinking. GPUs only need two domains: full blast balls-to-the-wall mode and 2D low-power desktop mode. Anything in between is worthless and just begging for glitches.

Probably so, for at least one or two more driver updates... then maybe a post showing everyone how to do it on their own. It's really simple to do, and half of the mod consists of intentionally messing up what NVIDIA did on purpose, LOL. Let's see how it goes for more people. I will test it further with 980M after I get finished with the 980M GPU compatibility testing to be done on the M18xR1. I will use this same driver mod for 780M SLI when I move it to the R1.

 

The issue seems to be linked to ASIC quality through circumstantial evidence. What is really odd though is the issue seems to be card agnostic. As in if one card has 75% and the other 65%, for some reason it will always be the master GPU that runs a different voltage, even if you switched the cards around or put them in different PCIe slots. I know this both from personal testing as well as reports from others. In addition, Kepler cards that ran fine on previous drivers, also started to show this voltage discrepancy when updating to the new 344.xx drivers. I take this as strong evidence that this is a driver-based issue.

And yes the only way to fix it is through a vBIOS mod. I wouldn't mind if it was as simple as disabling boost, but no you basically have to go into the boost and power tables and mess around there. What makes it really aggravating is that this is an issue even on stock clocks for certain games. Because of Maxwell's "efficiency" due to dynamic throttling, sometimes the voltage doesn't ramp up fast enough to catch the boost clocks and so you get a crash. Other times the boost bins "crossover" and the corresponding Vddc becomes screwed up resulting in another FUBAR.

Sums up Maxwell's problem right there.

Why Kepler cards never had (or at least, much fewer reports of) this issue, I can only guess it's because the dynamic throttling wasn't anywhere near as aggressive as Maxwell's. Seriously, have a look at Tom's testing and see how Maxwell throttles all over the godamn place in order to improve efficiency. Let me snip a few quotes:

All good in theory, but we've seen in practice this leads to problems, especially when you put out terribad sub-AMD quality drivers.

And this right there sums it all right up:

So basically Maxwell is NOT inherently more efficient than Kepler, nVidia just implemented a more aggressive throttling algorithm.

I rest my case.

 

Well, Maxwell is still new. Give Johnksss and svl7 time to fix it and it may be just a bad memory before long. If I knew how to do it, I would be working on it, too. Those guys, as well as Prema, have knowledge that is priceless. Without their assistance to correct the mistakes of the "firmware experts" there would not be a laptop anywhere that isn't a pile of crap, including the P570WM3.

Stock Kepler vBIOS for desktop and mobile was pretty screwed up as well, but we tend to forget that after enjoying a sweet vBIOS mod for so long. In fact, I never even personally experienced it with Kepler for months because I replaced the stock vBIOS with an svl7/johnksss mod on day one with 680M and 780M. I tested the stock vBIOS on both GPUs after many months of using only the modded version. And, I tested the stock vBIOS just to see how terrible it was... and terrible it was, for BOTH 680M and 780M. John's Titan became a holy terror when he was done modding the vBIOS. Although we should not need to, we have to just expect that from the Jolly Green Giant. Their vBIOS engineers need to find a new line of work, because they suck as their current jobs.

 

I have no doubt john and svl7 will fix this. My main gripe is with 2 things:

1. nVidia releasing borked drivers and
2. Refusing to acknowledge it's a problem, going so far as to include a paragraph in driver release notes stating it's normal, and lying through their teeth saying "An end-user gains nothing by attempting to raise the voltage of the higher performance GPU because its clocks must not exceed those of the other GPU." What actually happens is the boost clocks are lowered on both cards when this happens, and when fixed both cards boost to a higher clock. For example in SLI my 970s only boost to 1366, but when I run single card the higher ASIC card boosts to 1404, while the lower ASIC one still boosts to 1389.

See I bought nVidia to get away from AMD drivers, generally described using different variations of the word "feces". Then nVidia starts doing the same, and blatantly lies about it rid themselves of any responsibility, while still charging more for their products. I don't know what to say at this point. Blegh.

Also I may have had one too many drinks this evening...

 

Well, in a way they are not lying. But, not honest in a way they would like for us to think of it as being honest. I say this because it is "normal" for NVIDIA vBIOS to be garbage. Their GPUs malfunction by design and they don't ever become amazing until someone else fixes it. Sad, but true. So, yeah... it's "normal" for their GPUs to act stupid.

The only edge NVIDIA has over AMD in the mobile realms is their products have potential to be amazing, and they don't die prematurely and malfunction as frequently as AMD. There's no cure for AMD-titis on a mobile platform.

It would be really nice if one or both of them would just become awesome at everything without third-party expertise always having to intervene.

 

yeah sadly it seems AMD has bailed on the mobile platform, giving Jen-Hsun a free pass to do whatever his royal highness pleases

I understand AMD's current focus on the console/APU market is because that's where they're raking in the (badly needed) money to continue their operations so I don't fault them for it. I'm fine with that as long as it means that money is put to good use in churning out good desktop GPUs. God knows we don't need another Intel.

 

Here's a video I just uploaded. I just grabbed this game on the Steam sale and played about 15 minutes this afternoon. I am using this driver mod. GPU utilization show 0% because K-Boost is enabled.

It sucks that Konami had to ruin this with a 60 FPS lock-down. But, it plays really nice and looks amazing in spite of being crippled like that.

 

Those drivers work great for me (on par with most modded drivers i have tried). As for Maxwell bios stability it seems to be fine at stock (allows a decent OC), and issues i have experienced with Maxwell are more than likely related to purely bad driver engineering. Since some of us like to use unsupported hardware we can expect additional issues to arise. Since i have tried several methods to remedy a particular throttling issue i have been having after a overnight PC shutdown and some weird randomly occurring OS issues creating a registry fragmentation and hanging when i boot to desktop where only a few of the required windows programs will load leaving my desktop non responsive requiring a safe boot into a registry optimize before it works again. Last ditch resort would be to wipe and install a clean 8.1 OS... sometimes the worst case scenario has to happen, and i really hope it is just a bad OS issue that can be fixed with a clean boot because no one likes installing all their games, programs etc all over again...

 

It just occurred to me that nVidia is starting to turn into Alienware, while AMD is becoming Clevo.

...I've been dying to use the Alienware/Clevo analogy for nVidia and AMD, please don't kill me lol

 

Nah, we won't kill you. Although, I suspect Clevo is just slow to join the BGA garbage party. I anticipate they will contract the cancer soon enough, or maybe they already have and it hasn't fully metastasized yet due to leftover parts that need to be sold. If they refuse to participate in the techno-feces game they will definitely become my new hero. Unless they go back to doing things the right way with sockets and slots in their laptops, I am already done with Alienware. I am just getting all the mileage I can from their awesome products from yesterday. I will not purchase any laptop as a pleasure/gaming/benching high performance system if the CPU or GPU is not upgradable, no matter whose name is plastered on it. I refused to degrade myself or lower my standards with childish toy PC filth like that. Homey don't play dat.

 

Clevo has definitely released a few BGA machines already. However I think they realize the evils of BGA, and their new strategy seems to be offering machines that take socketed desktop chips as a way of getting around Intel, since desktop chips are still safe (for now anyway).

I surely hope the P7x1ZM isn't just a one time fad product but the beginning of a silent but ever-enduring protest against Intel's draconian BGA d!cktatorship

 

Well, they have both. Desktop CPU (mainstream i5 and i7 support) models with single GPU (so far) that's the P7xxZM models and then the thin/light 4710HQ stuff that's the P6xxSx models. To be honest, I don't care as long as there's choice. If people are fine with the locked down stuff, I'm happy for them... but I will never find that to be the best way. And EVEN if I were to ever buy a BGA-type machine, the 4980HQ would be the minimum CPU requirement for me, or whatever its successors were. People really underestimate how much CPU power I can do with XD.

If I lived in the US, there would be zero acceptance. If I was in the states and I had a decent job right now, I'd have likely already had at least a 4910MQ with some CLU and spare heatsinks being lapped in preparation for a constant 4.2GHz daily clock minimum.

Either way, this whole non-race thing is annoying me. Before maxwell came out, everyone jumped on AMD's R9 280x, 290 and 290x cards all over the net. Now, it's all "970 970 970" everywhere you see. No 980, no 290x... if you can buy a 980 save and buy two 970s. It's making me facepalm. It made me ticked that people would instantly ignore any nVidia suggestion three months ago in desktop-centric forums, and now it's annoying me that nobody seems to remember AMD exists. Don't people understand "buy for your needs"? *rants and raves*

Anyway. Regardless of what anyone says, AMD needs to properly trounce nVidia before stuff gets better. I don't care who is an engineering overseer or who's running marketing or what. If AMD brings out their good stuff and gets some decent driver support out the gate, bet your next paycheck nVidia will start improving. Remember what happened early this year when about 5 games came out using mantle and AMD were doing their little chugga chugga dance and mooning green team, then nVidia went *sigh* *cracks knuckles* and then we got that supreme DX11 optimization driver? I don't think mantle being big in the public eye was a coincidence just before that super optimization came out. And anybody who believes it is is fooling himself/herself.

At the very least, I'm fine using modified stuff. I now run throttlestop 24/7 with TimePeriodAC=1 because it fixed binding of isaac stuttering and the random judder I get in USF4 when playing SP and such. I'm also using Dufus' executable for high multipliers so my CPU sits at max boost 24/7. It doesn't negatively affect my temps in any way either, because as long as util is low it's not a problem, and if I ever need to use battery I can use throttlestop or windows to limit my processor multiplier. Not like it matters; I've been checking the watts it draws. There's no change at low usage. Now that my shutdown issues are gone without a trace and I got over my PTSD about a PC turning off or on (I'm DEAD serious I would get a mini-panic attack whenever my sister turned on her Windows 7 netbook due to the sound... and my PC runs Windows 8) I'm much more excited and ready to do things like the foil tape mod for my main GPU's fan and I wanna get the AMD 8970 heatsink for it too, as well as putting some CLU on my CPU and keeping it 3.9GHz stock if I can. Hell, I'm even up for keeping my GPUs on a 24/7 mild OC if I know it won't crash; if any game is JUST that hot I could always load the default clocks macro; it works while i'm mid-game and stuff. It's even easier for me since I use borderless windowed so often.

 

Trust me, you do not want a 4980HQ CPU. It's a low-TDP piece of excrement and won't hold full turbo speed under 100% load no matter what you do to it with XTU. It's a compromiser CPU and not as powerful as your 4800MQ. That CPU is just a gimmick for people dumb enough to believe an Intel Iris Pro iGPU is capable of delivering performance at an amazing level, or people that just want something nice for work or web surfing. It is definitely not an enthusiast grade processor. (The sucker has an insane iGPU core clock, but still runs like a GPU from the 500MHz era.) Iris Pro is competing against other Intel integrated graphics options and the lowest end bottom-feeder integrated solutions from AMD and NVIDIA. It is something fitting for a MacBook, or an affordable mass-production business solution, just like all of the other HQ turds. While it might be better than most HQ crap, 2 times 0 still equals 0.

Yeah, you're totally right about the competition thing. It was so obvious NVIDIA was holding out on us when they effortlessly stomped on the AMD Mantle sideshow. I'm convinced they were holding that card for just the right time to crap all over their parade, which is really dirty pool against AMD and sort of a betrayal of their own customers, too. Imagine what NVIDIA could give us if they were not playing some kind of stupid techno-poker, cat-and-mouse game. If they would just unleash their best product it would dwarf the performance of AMD's best to such an extent that AMD would be compelled to totally fold up shop and stop selling mobile PC graphics cards or put more effort into the game to avoid extinction. Hopefully, they would choose the latter.

I am not a red or green fanboy and actually preferred ATI for a very long time. I am at the point now where I don't like AMD only because I have a low tolerance for mediocrity. So, against my wishes, and totally by default, the win goes to NVIDIA. I don't give my money to AMD because I don't believe in funding an underdog that is content to stay in second place. AMD doesn't get any points from me for trying because it seems to me they aren't.

I would jump off the green train onto the red train in a heartbeat if there was a compelling reason to want to. But there's not and now we are left with NVIDIA playing silly games like only being good enough to be better than the alternative. It's really sad when the business strategy is so obviously to maintain lesser of two evils status. I really do miss the venom that use to spew in forums between the red versus green fanboys and I wish it would make a comeback. Those were some fun times, and it made both brands better.

 

Can't wait to try this in my Clevo P377SM-A with dual 880M's. MrFox.. Is the upgrade from dual 880M to dual 980M's worth it? Wanna read your thoughts about it. Thanks my friend

 

I'm going to have to "yes" it sure is, but my GPUs are supposed to be delivered later today and I am optimistic based on numbers I am seeing floating around. In a day or two I'll show you some numbers to confirm whether or not I am correct. It sure as heck better be awesome at just over $1500 for pair of them.

 

This is the score I get with this driver with the same overclock amount in 780M SLI... not close to 10000 points in 3DMark but better than my old scores.

 

Mr. Fox i made you a mirror in case you wanna add it to OP:

https://mega.co.nz/#!IgAk3JQR!gQ3RbiPZHpV6LTudV3qGq0jzZ5Gl4RLrmyBiYT7ugA8

Thanks once again, today i'm gonna try the drivers and will drop some feedback on them

 

How well performs these mod drivers without mod vbios compared with standard laptop drivers?

 

yeah okay, you turned down an AW18 for 900$ cause it had 2x290's, TROLLOLOLOL

Anyways......... Fox, the driver you have for download there, works great until I touch eVga precision or nVidia inspector or anything that could alter somethng with the videocard, then it reverts back to original performance. HOWEVER, it still performs better then any clock Ive used before AS-IS , and for the first friggin time, in years Ive actually HEARD my fans start spinning like an aircraft during a benchmark too. I get nearly 6000 on firestrike on extreme with a single card. was getting like 4,000's before. Never having beat 5300 as a max until now. So, thanks alot, i have no idea what you did, but I wish to know more to understand it further, and maybe contribute to the knowledge in the future.

 

I have noticed that I was able to hold high clocks (1007/1500) all the way through Firestrike and 3DMark 11 without utilization dipping below 96% on either card. Scores are the highest I have ever seen. Great job Mr. Fox!

On a side note, since presumably only three files need to be replaced, I'd suggest that you just put them into a folder called Display.Driver and ask people to merge that folder into their driver folder and overwrite existing files. That makes the download mere kilobytes.

I cheated by using a text comparison tool to look at how the mod was performed. I feel that while it's no rocket science to make such a maneuver, it was pure genius to predict the tremendous performance benefits. Again hats off to Mr. Fox.

 

Fantastic work Mr. Fox. Finally the full power of 880M is unleashed . Couldn't be happier ehehe

 

Would that not cause the updated driver to not install due to breaking the signed package though? Easier to send a full driver package and instruct on installation than to ask people to fiddle otherwise, no? I think I'll give this another try, but overwrite my driver and see if I notice any benefits. And if Valley will FINALLY HOLD MY BLOODY 1100/1500 CLOCKS HNGGG (I even tried +87.5mV and it wouldn't work; doesn't even overdraw it just crashes super easy).

 

Desktop driver mod is working real nice on 980M SLI as well. And, the throttling issue appears to be too high CPU clock speed triggering a GPU downclock. Either NVIDIA, Windows 8, or both trying to force lower power consumption. Might be appropriate for a gaming jokebook that has low TDP BGA pile of crap CPU and a mickey mouse AC adapter..

See this post... http://forum.notebookreview.com/ali...-r2-dual-980m-sli-upgrade-94.html#post9869462

 

Now that you mention it, I do remember having to manually sign a modified driver (for something as trivial as a USB-to-ethernet converter) to make it work on Windows 8. It was very messy business indeed. Thanks for clarifying.

1100/1500 is some serious clocks

 

If Mr. Fox can do 1150/1500, I should be able to do 1100/1500 without throwing my CPU clocks to the wind XD

 

Mr fox using your new driver my 3dmark score jumped a pretty huge number of points!. And currently my CPU overclock is all jacked up so i have my CPU running slow because of issues with throttle stop. Scored a 9894, i know it isn't the best in the world, but your driver made a big difference with my system. I can't thank you enough, your hard work and your advice as always, is appreciated. sorry if this should have gone in his driver post, but I didnt want the thank you to get lost in a long post. SO thank you again. Now if I can get my CPU overclock fixed and throttle stop fixed I can only imagine how much higher i could go.


NVIDIA GeForce GTX 780M video card benchmark result - Intel Core i7-4930MX,Alienware 01W2J2

 

If people are getting nice boosts using this desktop driver mod, it kinda makes you wonder whether nVidia is deliberately crippling mobile GPUs for nefarious reasons.