djdblaster

Would this combination work? GPU is from AW 17 R5.

So I got this barebone system. My GPU (GTX 670M 1.5GB) experienced a surge and I'm looking for new cards now. I have tried a card from Alienware system (GTX 675M which has a TDP of 100W) and it would always go beyond 90 degrees C in games. In tests, it would go to 105 and rising. I have screwed the heatsink properly, put MX-2 paste on it and still temps were high. GTX 670M has a TPD of 75W, so I believe that was the problem? Like more TPD, more heat. In this case, increase of 33% in both categories, right?

According to that, I should be looking at the cards that have similar TDP? But then again, how do others fit more power hungry cards in? Heatsinks are same for every card, right?

Does anyone know the list of GPU's that can fit my laptop? Also, I'd like to know if there is any good German repair service that could take a look at the my old GTX 670M and possibly repair it?

As you can see, there is a lot of questions. My priorities would be to fix my old GPU and if that doesn't work, to buy the same GPU or some else that will work fine.

Thanks for taking time and effort in reading this!

Hello guys, I would like to refresh this thread with new question.

What other GPU's could I install in MS 16F32? Would it be possible to put in GTX 680M, 770M or 780M? Any info is helpful!

Hello guys! CLEVO W740SU with Win8.1 64-bit.

I've got 3 questions for this model:

1) I've got this laptop under warranty and would like to replace the keyboard (doesn't register all key strokes) by either Clevo Germany or UK.

2) The laptop throttles a lot on battery. Seen people getting BIOS fixed on other brands like Sager 2740, but I'm not sure if it would cause any damage to mine.

3) Installed Clevo Control Center v8.0153 and it doesn't show Bluetooth option to turn on/off like other things. (On Win7, I don't know CCC version, there was BT option for on/off). To have it turned off, I have to disable it via Device Manager. Any fixes for this?

I'd be happy to cooperate to solve these problems as well as I believe other people have them too.

So just to drop vBios here?

I suspect vRam might be off, but just in case, I would like to test modded vBios once again. Okay, so how do I dump it to svl7?

I would and have tried it, but it keeps rolling back to original clocks. Sorry for not mentioning this before. I simply cannot change clocks.

Can we try lower memory clock? Like 1400 mhz?

I got it off eBay with the same chipset as my dead GPU. Why wouldn't it work then?

Ah I see. I think sometimes people solder replacement chips on their boards, but I've got no idea if that's possible with the memory. Google might help of course. You might need to buy a replacement card if increasing the voltage doesn't help. Increasing the voltage is only increasing the voltage for the GPU core, and does not affect the voltage that is supplied to the memory chips. So if it's the RAM, then increasing the voltage is not going to help. Well, good luck, I can't think of what else to suggest. Try the driver cleaning & installation of drivers to rule out software, and maybe try a VBIOS with increased voltage.

EDIT: As an alternative to raising the voltage, you could instead just underclock the card a little to see if that helps. Keep taking 20Mhz chunks off the core clocks to see if that helps eventually. If that doesn't help after taking 100 - 200 Mhz off the clock, then try the same thing, but with the memory VRAM clock. Actually, this is a lot quicker way to test that this is the problem, rather than flashing countless new VBIOS files with increased voltages. (Also you won't be damaging the card any further, because you're not increasing the voltage). Your description of your artifacts you've seen reminds me of artifacts when overclocking too far when I had a Fermi card, and when I had my 8600M GT, so this could indeed be your problem. You might just need to game at slightly lower than stock clocks.

EDIT 2: I see you've been using Furmark to test your card. I wouldn't do that, there's no value to it, and it can damage your card. NVidia throttle their cards in Furmark anyway since Fermi, so your card will be running at reduced clocks to protect itself when running Furmark, so that's probably why it's stable during Furmark - because it's running at lower clocks. That further supports my idea that you can solve this by underclocking your card, either the core or memory depending on which component is causing you all these visual artifacts.

I just saw these 2 edits so I'll check that out later tonight. Hope something works! Thanks for advice!

Seems like the same issue I have... Unfortunately, I believe your GPU might be running out. You're lucky you have warranty.

This vbios with 0,92V.

Thanks! I'm going to try it out right now.

EDIT: Tried it, doesn't work. Would you please make 0.800V, 0.825V, 0.8500V and 0.875V?

The card is from AW m15x system. It has 1.5GB vRam. I'll try to flash this vBios that you provided Klem. As for the voltages, I would like to try different voltages. Also, if you guys know any software to test vRam, that would be great.

If you need to modify your voltages to keep your card stable at stock clocks then you have a problem with the hardware & you can send it back for a warranty refund. Be careful with increasing the voltage on Fermi, lots of dead 580M & 675M cards around due to people upping the voltages, and also due to one particular NVidia driver that raised the voltage to I believe 0.9 or 0.925V (something like that), and a lot of people thought this was their reason for their dead cards.

I don't have warranty on this card. What programs can I use to check vRam? Also, if one of them is bad, can I buy replacement chips?

I did make a new thread about the same system and GPU along with its vBios. Haven't had success so far. That is why I came to this thread hoping someone would be able modify my voltages.

The thing is that I actually got it working perfectly and the voltage was 800mV. Installed driver 327.29 I believe, don't hold me on this one and everything was smooth and no errors at all. Updated drivers and issues appeared. Reverted drivers back and issues were still there. GPU is fine. It came 100% working from AW system. Im gonna upload GPU-Z screenshot in next 30 minutes.

EDIT: Here is the GPU-Z screenshot for GTX 670M.

Started Crysis 3 mission and I immediately got CryEngine error, game crashed.

Is there any way someone edits my vBios? I would like to try different voltages.

I've got practice now, but I'll upload it in 3 hours.

Flashed it, fired up Crysis 3, its running great, just like previous vBios, but I still see some blue squares on walls, floors and doors. I fired up Skyrim, starting menu is great, no errors, I press new game and then I see blue squares, but wait, this time it actually started the game. I would get crash on loading screen. Now I see a couple of yellow and a lot of small blue squares. Gameplay is smooth thought. Then I press alt tab to check temps and vddc in GPU-Z When alt tabbing back, there are blue spots on the horse and clothes, few of them and those small blue and yellow squares turned into invisible, but you can still see them if you focus on them. Im running 332.21 drivers. GPU-Z reports the same info just like previous 1.5GB vBios. Temp in Skyrim was 70 C.

6 on top, 6 on bottom.

6 chips x 256 mb ram.

1.5GB for sure. I also checked with GPU-Z to and others like AIDA and similar.

Thank you Klem. I'm going to try it out in next 10 minutes. Hope my problem goes away. If I may ask, where did you get it?

EDIT: Flashed it, did not solve the problem. I still see blue squares when loading games and then they crash along with nVidia popup error. GPU tested with FurMark for like 30 seconds just to see if there are any artifacts, everything works great. Please, we could carry this thread to my private one that I created in MSI section because this thread is for Kepler mods and I don't want to disturb others. Although I still believe some slight voltage changes would solve this problem... svl7?

EDIT #2: I created 2 same threads... SHM... Sorry! LINKS:

http://forum.techinferno.com/msi/5812-ms-16f32-gtx-670m-vbios-help.html

http://forum.techinferno.com/msi/5811-ms-16f32-gtx-670m-vbios-help.html

EDIT #3: When I'm still here... Tried vBios Patcher RC1 and couldn't load any .rom file. Opened the program with Admin rights and nVidia GPU, still nothing.

MS 16F32 with GTX 670M here. Posted threads on MSI section and other forums. Dont wanna be and sound rude. Im completely open and willing to share whatever you guys need.

I would like to solve my issue with GTX 670M, I do know its Fermi card, but like I said, if Im rude please tell me or guide me to Fermi thread (which I havent found). I hope this post gets seen by svl7 as I believe he could solve my issue really quickly. Thanks again! Cheers everyone!

That's MX card. I have just M.

MS 16F32 here. Might be off topic, but is there any GTX 670M 1.5GB vBios? I have downloaded MSI's 3GB version with 620 mhz core clock, my voltage is 0.900Vin games - P3 stage. They crash right away along with nVidia error.

Tried Clevo vBios with 597 mhz core and 0.850V in P3 stage, games report blue squares. I somehow got it to 0.800V and thry ran fine.

Svl7, do you mind looking at my vBios and edit it or you post original vBios for GTX 670M 1.5GB?

Thanks in advance. I hope this doesn't sound rude cause I'm new here.