Stuttering Issues with GeForce Driver v331.4

Hi guys.:hi2:
My Alienware M17xR3 GTX675M been having stuttering issues quite badly.
Frequent fps drop to as low as 3 fps every 5 minutes while playing games.
Is there any fix for this problem?
I'd reset the bios and disabled multicore rendering by turning off threaded optimization but still same thing happened.

 

Hi

I'd try 314.22. I've has no end of problems like yours with all the later drivers - and not alone, the NV driver forum is groaning with Fermi users having the same issues.

Check out your DPC execution time with Latencymon before and after a driver change as well

 

Are GPU & CPU temperatures OK when you're gaming? Throttling of some description is the first thing that comes to mind when I read your post. Not sure disabling threaded optimization would help, what's wrong with the extra virtual cores.

 

Yeah, as Robbo says, it could be P1 throttling. I also had a lot of issues with this, needing a modded vbios to stop it. Use MSIAfterburner and set it up for the on-screen display with clocks and temps displaying while gaming.

If you see a drop from 620 to 420 clock then this is the power throttle cutting in. If it goes lower than that I would expect it to be overheating and this throttle occurs at 78c if I remember correctly.

Curiously I recently had mine replaced and this one NEVER throttles, but the stuttering on later drivers remains

 

I think it's worth a try on 314.22. Will do it.

 

I think it's worth a try on 314.22. Will do it.

 

I'm not really sure whether my GPU & CPU are OK.

 

I'm not familiar with Modded vbios & Throttling knowledge.

I had installed BIOS A12 unlocked & SATA tweak.

I noticed on the screen display when stuttering the core clock drop from 620 to 74 and gpu temperature 66 degree celcius.

What does it mean?

 

No problem. Do you know where the 675 came from? The R3 top NV GPU was a 580m.

That drop is the full blown throttle to P8 and the only time I have come across it is when overheating - and yours is just fine

I believe mine has a more recent vBIOS since the base voltage drop, when the first throttle cuts in, is no longer there in this vBIOS- so I don't need a modded vBIOS. If you like I can post the files needed but flashing is not without risk so only if you want to try it. Doesn't sound like it will help though

 

I donno where the 675M comes from.
I recently bought the laptop with some forumer.
A new one with ship date on 19/9/2013.
Is mine having throttle problem?
I didn't get the meaning of 'That drop is the full blown throttle to P8 '.
Yes,i would love to. Kindly send me the files.
So i guess the only solution for me this moment is to try v314.22.
If i install an old geforce driver, will i be missing fps? because latest driver seems to have lots of improvement on certain games.

 

Sorry, went 'prop head' on you! The P8 clock speed is normally only seen when in windows. It the lowest power saving setting along with the card's 'I'm having issues' throttle. If you see it while gaming (using the msiafterburner on-screen-display) then something is going wrong. Like I said I've only seen this on mine and others with a temp over 75c

Below contains the files you need and some instructions. The instructions were as supplied to me. I do it slightly different but the result is the same.

The 675mod rom is the original dell one with P0 (full power) clock and voltage matched with P1 (2D clock) - not your problem as far as I can see.

The .92v one has the same mod along with an increased base voltage - not recommended!!!!

The original one is straight from my new card without any mods - give this one a try first.
View attachment 675_noP1.zip

Edit: forgot, re loosing anything; in all my testing the newer drivers did give a small FPS increase in some games, but nothing I would have even noticed if I wasn't' looking for it. The BF4 beta stated on startup a later driver is needed - I call BS! It worked just fine on the old driver (but did not on the new one ) and if I'm missing some feature I cannot see it

 

I couldn't get the process running.


Btw, i had installed the v314.22. But . The same stuttering thing happened.

 

Weird looking command! You seem to be mixing up what is a command that should be typed in and the instructions. You should have pressed 'enter' after the first command that ends 'backup.rom'. I see this is missing from the first paragraph of the instructions.

Then the word 'nvflash.exespace,' should be just nvflash.exe with a space after it. Then the words 'after 6' should not be there, again instructions not to be typed. The -4 -5 -6 should not have " marks around it and a space between each one as I have typed. And finally you are trying to use the .92v one that is not recommended for the inexperienced - it can burn cards out!

Phew, sorry, it is not easier....

 

Place the.rom file inside nvflash folder and copy to desktop.
cmd(admin) type "cd C:\Users\your user name\Desktop\nvflash" without quotes >enter
Type "nvflash.exe -b backup.rom"
Flash type "nvflash.exe -6 vbiosname.rom"
Flash successful>reboot

 

LOL, I had just typed almost exactly the same when you assisted

Just to clarify for the OP, vbiosname.rom is intended to be replaced with the name of the vbios file you want to use

 

Is this consider done?

 

Yup, that is exactly what I would expect to see. You do have to power down and restart for it to take effect.

You will also find a new file in the folder titled backup.rom. This is your original and to roll back just use this name in place of the new one

 

Thank you so much for helping me..
I will try to run the game after this to see whether stuttering happen or not.
Btw, i got some few questions.
What are the precautions that i need to pay attention for this .92v installation? I'm just a normal user who doesn't do much on altering things up.
and actually, what does .92v stand for?voltan?
If i need to reformat laptop just to make sure nothing goes wrong, do i need to do this .92v installation all over again?

 

No problem but I'm not that hopeful it will address your problem, more of a 'hit and hope' suggestion - fingers crossed.

The base voltage the card runs at internally is 0.87 volts. By increasing this you can get more speed out of it by overclocking further than the stock voltage can manage. You should be able to overclock to around 720Mhz on the stock voltage.

You have to watch temps and look for artifacts in games (and benchmark programs). You can monitor temps and clocks while gaming using MSIAfterburner and when overvolting this is a must and probably more than you want to get involved with - unless you see a new hobby developing . Another reason for avoiding the .92v version is that these cards are already unreliable (I'm on my fourth with two machines and have never used overvolt!) so it's a dice roll.

The update you have just done is permanent and stored on the card its self so no need to worry about a re-format.

Edit: Just rechecked and you used a .87v one this time so no problem there.

 

i have tried playing games but voila! the same stuttering thing happened..
i guess the last thing to do is reformat my laptop,install back the driver and test it.
thanks for being helpful anyway.
btw, how do i rollback to the last 0.87v?

 

Hurmm...I guess reformatting the laptop doesn't fixed the problem either.
Am i doomed to play games with stuttering every 5 minutes for the rest of my life with the new laptop i juz bought?
If i ask Dell to fix this problem, will it void the warranty or is Dell going to take the GTX675M and replace it with GTX560M?
because i juz noticed on dell website the original configuration for this laptop didn't come with GTX675M but instead GTX 560M.

 

until now i replace from GTX 560M to GTX 660M ... no problem at all .. running smooth
+ i just modded with my own .inf file to make Nvidia display driver support for my laptop.. hahaha

 

Are you certain that you have been checking the temperatures of the GPU and the corresponding clocks (Mhz) when gaming, and not just checking the temperature of the card once you finish & exit the game? It still just looks suspiciously like temperature throttling to me. It's the maximum achieved temperature that's important to note.

 

Exactly what I was going to say . Did you get msiafterburner and setup the on-screen display?



When gaming you should see a display in the top left with the numbers.

I think that maybe the person that fitted the 675 did a poor job with the heatsink (or even used the wrong mounting bracket) and it IS overheating. The 675 will not be covered by the warranty as it is a non-standard part. In fact you could have trouble getting waranty help on anything if the tech decides that he wants the machine returned to standard before attempting repair.

If you can prove overheating then the easiest way forward would be to purchase a heatsink and x-bracket from dell for the 675m in the R4 - it will fit with no problem. The only problem is that you need a friendly dell person on the phone that will look up the R4 part when you have an r3 - but I know it is possible for them, they just have to look up the part numbers.

A new HS from dell comes complete with all the pads and paste preapplied, which is the trickiest part if you have never seen or done this before.

Another option is that if you have a local PC repair shop take in the card and heatsink and see how much they will charge you to repaste it?

 

I'm not sure about checking the temperatures of GPU & corresponding clocks (Mhz) while gaming without minimizing the game using msiafterburner as it doesn't have logging start before gaming like HWiNFO software.

Here is my HWiNFO data when logging start to stop on gaming.Open it with microsoft excel.

Testing

The game is shadow warrior on high setting and stuttering every 2 minutes.

 

That should be fine as long as you look at the peak number. I prefer afterburner simply because I can see exactly what is going on on-screen when things go bad when gaming. Also when I'm pushing it it makes sense to always see the temp. No logging (well, not easily accessible) just use it for the real-time data.

Can you post a screenshot (use windows clipping tool). I don't use any DL's that require a downloader

 

Or you can use a little program called GPUZ to track the temperatures. That's how I do it. After finishing a game I quickly exit to the desktop & view GPUZ sensors, then click on GPU Temperature & select Max Temperature. (Obviously you load up GPUZ before you start the game).

 

How to upload an attachments on this site?

 

Click 'go advanced' below the reply. Then scroll down to see manage attachments. Add Files. Navigate to upload the image and then attach it. Hit 'insert inline' to place the pic or attachment at the cursor.

 

FPS When Stuttering


Before Stuttering


After Stuttering


Overall Log for All Hardware
View attachment Overall Log for All Hardware.zip

What can you conclude based on the information above?

 

Aha, all I needed to do was look at the first 'FPS When Stuttering' pic you included - it's throttling due to temperature. When it gets to 78degC it's dropping the clocks of the video card to prevent overheating. So, to fix the problem you can improve the cooling efficiency, and there might be a modded VBIOS that you can flash to remove the 78 degC throttling (but not sure about that last point). So, to improve the cooling, make sure you're using the laptop on a hard flat surface, so the intake vents at the bottom of the laptop aren't blocked. And/or use a can of compressed air to blow the dust out from the heatsink fins & fans. And/or raise the back of the laptop a couple of centimetres by placing laptop on a couple of bottle caps OR getting a laptop cooler. And/or repasting the GPU if the cause of the high temperatures is due to a bad paste job. (Or is it like 33 degC in your room or something?). The most significant things that will effect your GPU temperature are: blocked heatsinks with dust, bad paste jobs on GPU (and room temperature, but can't influence that one much I guess!). Those are the first things to tackle.

 

I can clearly see the full overheat throttle cutting in at around 77c, exactly as designed so overheating is the issue. Curiously in the screen shots the temps are low but the log file is clear looking at the GPU temp and gpu clock correlation... ah just seen Robbo has advised

There is no fix for the 78c throttle you just have to fix the problem of the heat not being correctly dissipated. I was part of a mega-thread a while ago all about this throttle - even got Dell HQ on a conference call about it. I'll hunt it down and post for you but basically a clean and repaste fixes. Dell were even offering 580m owners (the 675m is the same card rebadged) at the time a quick way to apply for a replacement heatsink with new pads and paste for those under warranty.

Basically the problem was widespread due to poor factory paste jobs or more specifically a piece of thermal putty along one side of the card was getting missplaced and hindering contact in many cases.

http://forum.notebookreview.com/ali...x-580m-78c-throttling-see-post-191-later.html

 

Yeah, almost one million posts later '580m gpu throttle'..... ... ... This will work with the 675m http://forum.notebookreview.com/ali...ttling-m17x-r3-580m-without-120hz-screen.html
Drop temps, 75C high for gaming, mine 580m (675m rebranded 580m) temps 65C.

 

Interesting, I never saw that one. I guess an earlier bios, blind flashed, removed it. Still just a workaround though and the temps should never get that high.

I don't recommend trying a blind flash with an earlier bios. Dell made some major changes with the later Bios' and some people ended up with totally dead machines when trying a roll back (for a different issue)

Mine also games around 64c on stock clocks

 

M17xR3 so there is no issue with it.

 

Cool, I was mixing it up :thumbsup:

 

Well me too...all of a sudden I got an R4....

 

So I guess I should redo the paste jobs on GPU and get a decent thermal pad. My laptop isn’t covered much with dusts as it just one month old laptop. Just checked it.

Lol. This guy is funny. I believe in 100 years later there is still a guy talking about 680M/780M and that will be me.
I guess I should try to remove bios gpu throttling. Where do I start? Before this I had the 0.92v flashing, will 0.92v flashing gonna mix up if I trying to do a blind flash with an earlier bios? Or should I downgrade the 0.92v flashing to the original first?

 

I don't think you used the .92 vbios when I looked back but you should flash with the one titled '675m_original.rom'. A .92v vBIOS will only make it worse.
http://forum.notebookreview.com/alienware/721878-nvidia-320-18-driver-warning.html

Probably just a poor job by whoever fitted the 675m . It may have even had an AMD GPU originally and the x-bracket from one of those will fit but provide a poor contact. Also, when removing the HS, look for black tape aroung the core area of the HS. If it is there remove it. It matches the size of the original GPU and will hinder contact on a slightly larger core.

Go for the blind flash and roll back ONLY if nothing else works. Who know what temps it will get up to when there isn't a throttle stopping it at 78c - and running hot all the time is going to shorten the life or an already unrelaible card

 

Yep, just do the paste job, don't bother with the VBIOS yet, got to sort those high temperatures out first!

 

SO just to conclude:

First step, i need to flash with the one titled '675m_original.rom'

Second step, replacement of heatsink with new pads and paste. My seller suggest me to call Dell and tell them 'assume u are using GTX560m which u have problem with heat issue. ask them replace a new heatsink for u. dont mention the GTX675m'.
Actually, what exactly is HS?I don't get the meaning of this sentence 'It matches the size of the original GPU and will hinder contact on a slightly larger core'?

Third step, only blind flash and roll back ONLY if nothing else works.
Actually, i don't get it. Removing throttle means there isn't a throttle stopping it at 78c?
How can j95 stated that his having 75C high for gaming, 580m (675m rebranded 580m) temps 65C after 'How to remove bios GPU throttling for M17X-R3 580m with or without 120Hz screen' thing.
I thought removing bios GPU throttling means it fixing your gpu temperature by lowering the temperature

 

I don't really understand your ROM issues, so maybe MickyD or someone else can pick up on that one. Regardless, first step is just to re-paste using your existing heatsink unless you're certain that the heatsink is not fitting properly (this is assuming all the pads are located in the correct places). So, once you've done the re-paste, then check the temperatures immediately, even at idle in case you've messed up the process, and then while gaming or benchmarking. If temperatures are good, then you're sorted. If they're not good, then remove heatsink from GPU and check to see if the thermal paste has been distributed evenly over the whole of the GPU chip & heatsink. If it's not evenly distributed, then maybe you've not put on enough paste, or maybe your heatsink is not sitting level on the GPU chip (thermal pad arrangement can affect this). If when you remove the heatsink you see that the paste layer is very thick on the GPU chip, then you could conclude that the X-bracket is the incorrect part (the posts are too high), and you would need to buy the correct X-bracket to fit your GPU.

(You can re-use your old pads as long as they're not damaged)

 

Step one - yes

Step two: I assume that the 560m is the original GPU (?) but you want a 580m one from dell, the card is identical so will be perfect and it will be listed for the R3 (so the salesperson will see it!). The seller may also be trying to help with getting the part under warranty but you don't want one from a 560m. The mounting x-bracket, that attaches the HS (heatsink ) has post heights matched to the thickness of the core that is straddles.

On the HS, the area that meets the core sometimes has a black tape surround, don't know why they do this, it just looks neat and tidy. Problem is that this tape has a thickness. If it does not fall outside of the contact area it will stop the HS from getting proper contact. If you get a new heatsink for a 580m then this will not be a problem. Hope this helps - of course I can 'see' in my mind exactly what I'm talking about

Removing the throttle fixes nothing. It just allows the card to run hoter without throttling. 75c is the maximum acceptable temp (IMO) and will happen when overclocking. At stock 65c is a more realistic number, I hit around 72c when overclocking above 720Mhz (620Mhz stock clock).

 

Are you sure there's a difference between the heatsinks of 580M and 560M? I have the 560M heatsink & it has 3 heat pipes, so wouldn't the cooling efficiency still be the same as the supposed 100W 580M heatsink? I'm thinking the 560M & 580M heatsinks just have different part numbers due to different thermal pad arrangements (and lack of black tape)?

 

Not sure at all! i just try to avoid issues with correct parts (if you're going to purchase them). I know a lot of people get away with heatsinks from different models, but some don't. I think you are right about the part numbers not being important with respect to the heat dissipation properties.

If the 560m is a 'fermi' core, as the 580m is, then the thickness of the core will be the same so it should be fine. It gets confusing (for me!) with the 6xx series as they are a mix of Fermi and Kepler cores

 

Haha, yes, if he has to purchase a heatsink, then always best to go with one with the correct part number! Seeing as he's already got one, I think he should have a go at repasting his existing heat sink. Yes, it gets confusing with all the different permutations of heatsinks & cards; the 560M heatsink & X-bracket seems quite flexible, as it works with GK104 (Kepler) too, my 670MX - working out the thermal padding is the hardest part, but I agree that not all heatsinks work with all GPUs, especially if some of the metal parts of the heatsink are too low (some parts of chips on card fouling on metal of heatsink before heatsink can sit flat on GPU core) or if the heatsink doesn't cover all of the chips on the card.

 

OMG you guys! I think mine just works like a charm

MSI Afterburner Hardware Monitor
View attachment 103596

GPU Temperature
View attachment 103597

GPU Voltage
View attachment 103598

All i did was flashing the vbios to the '675m_original.rom'
View attachment 103599

Thanks a lot u guys MickyD1234 Robbo99999 j95 :thumbsup: . It will be my pleasure to treat u guys to dinner .

 

That's great news! I didn't expect that, it was more of a hope (and a good starting point) with than new vbios.

FYI, the one you used is a dell original from my card that was new about 4 weeks ago. I noticed that it acted very differently with respect to voltages and the power management. I was looking for someone to check it out when I first got it - job done - thank you

Something seems to have gone wrong with the attachments but I'm hoping temps are good and it's not just that it does not throttle at 78c+??

 

MSI Afterburner Hardware Monitor


NVIDIA GeForce GTX 675M Temperature


NVIDIA GeForce GTX 675M Voltage


Flashing the vbios to the '675m_original.rom'

 

Wow...it did save my life. :thumbsup:
So that vbios is a new one from Dell? i guess Dell did some improvement.
The temperature seems to be just fine.